Medicare Program; Changes to the Hospital Inpatient Prospective Payment Systems and Fiscal Year 2004 Rates

AGENCY:

Centers for Medicare and Medicaid Services (CMS), HHS.

ACTION:

Final rule.

SUMMARY:

We are revising the Medicare hospital inpatient prospective payment systems (IPPS) for operating and capital costs to implement changes arising from our continuing experience with these systems. In addition, in the Addendum to this final rule, we are describing changes to the amounts and factors used to determine the rates for Medicare hospital inpatient services for operating costs and capital-related costs. These changes are applicable to discharges occurring on or after October 1, 2003. We also are setting forth rate-of-increase limits as well as policy changes for hospitals and hospital units excluded from the IPPS that are paid on a cost basis subject to these limits.

Among other changes that we are making are: changes to the classification of cases to the diagnosis-related groups (DRGS); changes to the long-term care (LTC)-DRGs and relative weights; the introduction of updated wage data used to compute the wage index; the approval of new technologies for add-on payments; changes to the policies governing postacute care transfers; payments to hospitals for the direct and indirect costs of graduate medical education; pass-through payments for nursing and allied health education programs; determination of hospital beds and patient days for payment adjustment purposes; and payments to critical access hospitals (CAHs).

EFFECTIVE DATES:

The provisions of this final rule, except the provisions of § 412.230(e)(2)(ii)(A) (because it grants an exemption) and § 412.278(f)(2)(i), are effective on October 1, 2003. The provisions of § 412.230(e)(2)(ii)(A) and § 412.278(f)(2)(i) are effective on August 1, 2003. This rule is a major rule as defined in 5 U.S.C. 804(2). Pursuant to 5 U.S.C. 801(a)(1)(A), we are submitting a report to Congress on this rule on August 1, 2003.

FOR FURTHER INFORMATION CONTACT:

Stephen Phillips, (410) 786-4548, Operating Prospective Payment, Diagnosis-Related Groups (DRGs), Wage Index, New Medical Services and Technology, Patient Transfers, Counting Beds and Patient Days, and Hospital Geographic Reclassifications Issues.

Tzvi Hefter, (410) 786-4487, Capital Prospective Payment, Excluded Hospitals, Nursing and Allied Health Education, Graduate Medical Education, and Critical Access Hospital Issues, and Long-Term Care (LTC)-DRGs.

Sandra Hetrick, (410) 786-4542, RCE Limits.

SUPPLEMENTARY INFORMATION:

Availability of Copies and Electronic Access

Copies: To order copies of the Federal Register containing this document, send your request to: New Orders, Superintendent of Documents, PO Box 371954, Pittsburgh, PA 15250-7954. Specify the date of the issue requested and enclose a check or money order payable to the Superintendent of Documents, or enclose your Visa or Master Card number and expiration date. Credit card orders can also be placed by calling the order desk at (202) 512-1800 or by faxing to (202) 512-2250. The cost for each copy is $10.00. As an alternative, you can view and photocopy the Federal Register document at most libraries designated as Federal Depository Libraries and at many other public and academic libraries throughout the country that receive the Federal Register.

This Federal Register document is also available from the Federal Register online database through GPO Access, a service of the U.S. Government Printing Office. Free public access is available on a Wide Area Information Server (WAIS) through the Internet and via asynchronous dial-in. Internet users can access the database by using the World Wide Web; the Superintendent of Documents Home page address is http://www.access.gpo.gov/nara_docs/ , by using local WAIS client software, or by telnet to swais.access.gpo.gov, then login as guest (no password required). Dial-in users should use communications software and modem to call (202) 512-1661; type swais, then login as guest (no password required).

Acronyms

AHIMA American Health Information Management Association

AHA American Hospital Association

CAH Critical access hospital

CBSAs Core Based Statistical Areas

CC Complication or comorbidity

CMS Centers for Medicare & Medicaid Services

CMSA Consolidated Metropolitan Statistical Areas

COBRA Consolidated Omnibus Reconciliation Act of 1985, Pub. L. 99-272

CPI Consumer Price Index

CRNA Certified registered nurse anesthetist

DRG Diagnosis-related group

DSH Disproportionate share hospital

FDA Food and Drug Administration

FQHC Federally qualified health center

FTE Full-time equivalent

FY Federal fiscal year

GME Graduate medical education

HIPC Health Information Policy Council

HIPAA Health Insurance Portability and Accountability Act, Pub. L. 104-191

HHA Home health agency

ICD-9-CM International Classification of Diseases, Ninth Revision, and Clinical Modification

ICD-10-PCS International Classification of Diseases Tenth Edition, and Procedure Coding System

IME Indirect medical education

IPPS Acute care hospital inpatient prospective payment system

IRF Inpatient Rehabilitation Facility

LDP Labor, delivery, and postpartum

LTC-DRG Long-term care diagnosis-related group

LTCH Long-term care hospital

MCE Medicare Code Editor

MDC Major diagnostic category

MDH Medicare-dependent small rural hospital

MedPAC Medicare Payment Advisory Commission

MedPAR Medicare Provider Analysis and Review File

MEI Medicare Economic Index

MGCRB Medicare Geographic Classification Review Board

MPFS Medicare Physician Fee Schedule

MSA Metropolitan Statistical Area

NECMA New England County Metropolitan Areas

NCHS National Center for Health Statistics

NCVHS National Committee on Vital and Health Statistics

O.R. Operating room

PPS Prospective payment system

PRA Per resident amount

ProPAC Prospective Payment Assessment Commission

PRRB Provider Reimbursement Review Board

RCE Reasonable compensation equivalent

RHC Rural health center

RRC Rural referral center

SCH Sole community hospital

SNF Skilled nursing facility

TEFRA Tax Equity and Fiscal Responsibility Act of 1982, Pub. L. 97-248

UHDDS Uniform Hospital Discharge Data Set

Table of Contents

I. Background

A. Summary

B. Summary of the Provisions of the May 19, 2003 Proposed Rule

C. Public Comments Received to the May 19, 2003 IPPS Proposed Rule

II. Changes to DRG Classifications and Relative Weights

A. Background

B. DRG Reclassification

1. General

2. Review of DRGs for Complications or Comorbidity (CC) Split

3. MDC 1 (Diseases and Disorders of the Nervous System)

a. Revisions of DRGs 1 and 2

b. DRG 23 (Nontraumatic Stupor and Coma)

4. MDC 5 (Diseases and Disorders of the Circulatory System)

a. DRG 478 (Other Vascular Procedures With CC) and DRG 479 (Other Vascular Procedures Without CC)

b. DRGs 514 (Cardiac Defibrillator Implant With Cardiac Catheterization) and 515 (Cardiac Defibrillator Implant Without Cardiac Catheterization)

5. MDC 8 (Diseases and Disorders of the Musculoskeletal System and Connective Tissue)

6. MDC 15 (Newborns and Other Neonates with Conditions Originating in the Perinatal Period)

a. Nonneonate Diagnoses

b. Heart Failure Codes for Newborns and Neonates

7. MDC 17 (Myeloproliferative Diseases and Disorders and Poorly Differentiated Neoplasms)

8. MDC 23 (Factors Influencing Health Status and Other Contracts with Health Services)

a. Implantable Devices

b. Malignancy Codes

9. Medicare Code Editor (MCE) Change

10. Surgical Hierarchies

11. Refinement of CCs

12. Review of Procedure Codes in DRGs 468, 476, and 477

a. Moving Procedure Codes from DRG 468 or DRG 477 to MDCs

b. Reassignment of Procedures among DRGs 468, 476, and 477

c. Adding Diagnosis Codes to MDCs

13. Changes to the ICD-9-CM Coding System

14. Other Issues

a. Cochlear Implants

b. Burn Patients on Mechanical Ventilation

c. Multiple Level Spinal Fusion

d. Heart Assist System Implant

e. Drug-Eluting Stents

f. Artificial Anal Sphincter

C. Recalibration of DRG Weights

D. LTC-DRG Reclassifications and Relative Weights for LTCHs for FY 2004

1. Background

2. Changes in the LTC-DRG Classifications

a. Background

b. Patient Classifications into DRGs

3. Development of the Final FY 2004 LTC-DRG Relative Weights

a. General Overview of Development of the LTC-DRG Relative Weights

b. Data

c. Hospital-Specific Relative Value Methodology

d. Low Volume LTC-DRGs

4. Steps for Determining the Final FY 2004 LTC-DRG Relative Weights

E. Add-On Payments for New Services and Technologies

1. Background

2. FY 2004 Status of Technology Approved for FY 2003 Add-On Payments:

Drotrecogin Alfa (Activated)—Xigris®

3. FY 2004 Applicants for New Technology Add-On Payments

a. Bone Morphogenetic Proteins (BMPs) for Spinal Fusions

b. GLIADEL® Wafer

4. Review of the High-Cost Threshold

5. Technical Changes

III. Changes to the Hospital Wage Index

A. Background

B. FY 2004 Wage Index Update

C. FY 2004 Wage Index Changes

1. Elimination of Wage Costs Associated with Rural Health Clinics and Federally Qualified Health Centers

2. Paid Hours

D. Verification of Wage Data from the Medicare Cost Reports

E. Computation of the FY 2004 Wage Index

F. Revisions to the Wage Index Based on Hospital Redesignation

1. General

2. Effects of Reclassification

G. Requests for Wage Data Corrections

H. Modification of the Process and Timetable for Updating the Wage Index

IV. Other Decisions and Changes to the IPPS for Operating Costs and GME Costs

A. Transfer Payment Policy

1. Transfers to Another Acute Care Hospital

2. Technical Correction

3. Expanding the Postacute Care Transfer Policy to Additional DRGs

B. Rural Referral Centers

1. Case-Mix Index

2. Discharges

C. Indirect Medical Education (IME) Adjustment and Disproportionate Share Hospital (DSH) Adjustment

1. Available Beds and Patient Days: Background

2. Unoccupied Beds

3. Nonacute Care Beds and Days

4. Observation Beds and Swing-Beds

5. Labor, Delivery, and Postpartum Beds and Days

6. Days Associated with Demonstration Projects under Section 1115 of the Act

7. Dual-Eligible Patient Days

8. Medicare+Choice (M+C) Days

D. Medicare Geographic Classification Review Board (MGCRB) Reclassification Process

E. Costs of Approved Nursing and Allied Health Education Activities

1. Background

2. Continuing Education Issue for Nursing and Allied Health Education Activities

3. Programs Operated by Wholly Owned Subsidiary Educational Institutions of Hospitals

F. Payment for Direct Costs of Graduate Medical Education

1. Background

2. Prohibition Against Counting Residents Where Other Entities First Incur the Training Costs

3. Rural Track FTE Limitation for Purposes of Direct GME and IME for Urban Hospitals that Establish Separately Accredited Approved Medical Programs in a Rural Area

a. Change in the Amount of Rural Training Time Required for an Urban Hospital to Qualify for an Increase in the Rural Track FTE Limitation

b. Inclusion of Rural Track FTE Residents in the Rolling Average Calculation

4. Technical Changes Related to Affiliated Groups and Affiliated Agreements

G. Notification of Updates to the Reasonable Compensation Equivalent (RCE) Limits

1. Background

2. Publication of the Updated RCE Limits

3. Application of RCE Limits

4. Exceptions to RCE Limits

5. Geographic Area Classifications for RCE Limits

V. PPS for Capital-Related Costs

VI. Changes for Hospitals and Hospital Units Excluded from the IPPS

A. Payments to Excluded Hospitals and Hospital Units

1. Payments to Existing Excluded Hospitals and Hospital Units

2. Updated Caps for New Excluded Hospitals and Units

6. Implementation of a PPS for IRFs

4. Development of a PPS for Inpatient Psychiatric Facilities

5. Implementation of a PPS for LTCHs

6. Report of Adjustment (Exception) Payments

B. Payment for Services Furnished at Hospitals-Within-Hospitals and Satellite Facilities

C. Clarification of Classification Requirements for LTCHs

D. Criteria for Payment on a Reasonable Cost Basis for Clinical Diagnostic Laboratory Services Performed by CAHs

E. Technical Changes

VII. MedPAC Recommendations

VIII. Other Required Information

A. Requests for Data from the Public

B. Collection of Information Requirements

Regulation Text

Addendum—Schedule of Standardized Amounts Effective with Discharges Occurring On or After October 1, 2003 and Update Factors and Rate-of-Increase Percentages Effective With Cost Reporting Periods Beginning On or After October 1, 2003

Tables

Table 1A—National Adjusted Operating Standardized Amounts, Labor/Nonlabor

Table 1C—Adjusted Operating Standardized Amounts for Puerto Rico, Labor/Nonlabor

Table 1D—Capital Standard Federal Payment Rate

Table 2—Hospital Average Hourly Wage for Federal Fiscal Years 2002 (1998 Wage Data), 2003 (1999 Wage Data), and 2004 (2000 Wage Data) Wage Indexes and 3-Year Average of Hospital Average Hourly Wages

Table 3A—3-Year Average Hourly Wage for Urban Areas

Table 3B—3-Year Average Hourly Wage for Rural Areas

Table 4A—Wage Index and Capital Geographic Adjustment Factor for Urban Areas

Table 4B—Wage Index and Capital Geographic Adjustment Factor for Rural Areas

Table 4C—Wage Index and Capital Geographic Adjustment Factor for Hospitals That Are Reclassified

Table 4F—Puerto Rico Wage Index and Capital Geographic Adjustment Factor

Table 4G—Pre-Reclassified Wage Index for Urban Areas

Table 4H—Pre-Reclassified Wage Index for Rural Areas

Table 5—List of Diagnosis-Related Groups (DRGs), Relative Weighting Factors, and Geometric and Arithmetic Mean Length of Stay (LOS)

Table 6A—New Diagnosis Codes

Table 6B—New Procedure Codes

Table 6C—Invalid Diagnosis Codes

Table 6D—Invalid Procedure Codes

Table 6E—Revised Diagnosis Code Titles

Table 6F—Revised Procedure Code Titles

Table 6G—Additions to the CC Exclusions List

Table 6H—Deletions from the CC Exclusions List

Table 7A—Medicare Prospective Payment System Selected Percentile Lengths of Stay FY 2002: MedPAR Update March 2003 GROUPER V20.0

Table 7B—Medicare Prospective Payment System Selected Percentile Lengths of Stay: FY 2002 MedPAR Update March 2003 GROUPER V21.0

Table 8A—Statewide Average Operating Cost-to-Charge Ratios for Urban and Rural Hospitals (Case-Weighted)

Table 8B—Statewide Average Capital Cost-to-Charge Ratios (Case-Weighted)

Table 9—Hospital Reclassifications and Redesignations by Hospital—FY 2004

Table 10—Thresholds to Qualify for New Technology Add-On Payments: FY 2004

Table 11—LTC-DRGs Relative Weights and Geometric and Five-Sixths of the Average Length of Stay—FY 2004

Appendix A—Regulatory Impact Analysis

Appendix B—Recommendation of Update Factors for Operating Cost Rates of Payment for Inpatient Hospital Services

I. Background

A. Summary

1. Acute Care Hospital Inpatient Prospective Payment System (IPPS)

Section 1886(d) of the Social Security Act (the Act) sets forth a system of payment for the operating costs of acute care hospital inpatient stays under Medicare Part A (Hospital Insurance) based on prospectively set rates. Section 1886(g) of the Act requires the Secretary to pay for the capital-related costs of hospital inpatient stays under a prospective payment system (PPS). Under these PPSs, Medicare payment for hospital inpatient operating and capital-related costs is made at predetermined, specific rates for each hospital discharge. Discharges are classified according to a list of diagnosis-related groups (DRGs).

The base payment rate is comprised of a standardized amount that is divided into a labor-related share and a nonlabor-related share. The labor-related share is adjusted by the wage index applicable to the area where the hospital is located; and if the hospital is located in Alaska or Hawaii, the nonlabor-related share is adjusted by a cost-of-living adjustment factor. This base payment rate is multiplied by the DRG relative weight.

If the hospital treats a high percentage of low-income patients, it receives a percentage add-on payment applied to the DRG-adjusted base payment rate. This add-on payment, known as the disproportionate share hospital (DSH) adjustment, provides for a percentage increase in Medicare payments to hospitals that qualify under either of two statutory formulas designed to identify hospitals that serve a disproportionate share of low-income patients. For qualifying hospitals, the amount of this adjustment may vary based on the outcome of the statutory calculations.

If the hospital is an approved teaching hospital, it receives a percentage add-on payment for each case paid under the IPPS (known as the indirect medical education (IME) adjustment). This percentage varies, depending on the ratio of residents to beds.

Additional payments may be made for cases that involve new technologies that have been approved for special add-on payments. To qualify, a new technology must demonstrate that it is a substantial clinical improvement over technologies otherwise available, and that, absent an add-on payment, it would be inadequately paid under the regular DRG payment.

The costs incurred by the hospital for a case are evaluated to determine whether the hospital is eligible for an additional payment as an outlier case. This additional payment is designed to protect the hospital from large financial losses due to unusually expensive cases. Any outlier payment due is added to the DRG-adjusted base payment rate, plus any DSH, IME, and new technology add-on adjustments.

Although payments to most hospitals under the IPPS are made on the basis of the standardized amounts, some categories of hospitals are paid the higher of a hospital-specific rate based on their costs in a base year (the higher of FY 1982, FY 1987, or FY 1996) or the IPPS rate based on the standardized amount. For example, sole community hospitals (SCHs) are the sole source of care in their areas, and Medicare-dependent, small rural hospitals (MDHs) are a major source of care for Medicare beneficiaries in their areas. Both of these categories of hospitals are afforded this special payment protection in order to maintain access to services for beneficiaries (although MDHs receive only 50 percent of the difference between the IPPS rate and their hospital-specific rates if the hospital-specific rate is higher than the IPPS rate).

Section 1886(g) of the Act requires the Secretary to pay for the capital-related costs of inpatient hospital services “in accordance with a prospective payment system established by the Secretary.” The basic methodology for determining capital prospective payments is set forth in our regulations at 42 CFR 412.308 and 412.312. Under the capital PPS, payments are adjusted by the same DRG for the case as they are under the operating IPPS. Similar adjustments are also made for IME and DSH as under the operating IPPS. In addition, hospitals may receive an outlier payment for those cases that have unusually high costs.

The existing regulations governing payments to hospitals under the IPPS are located in 42 CFR part 412, Subparts A through M.

2. Hospitals and Hospital Units Excluded From the IPPS

Under section 1886(d)(1)(B) of the Act, as amended, certain specialty hospitals and hospital units are excluded from the IPPS. These hospitals and units are: psychiatric hospitals and units, rehabilitation hospitals and units; long-term care hospitals (LTCHs); children's hospitals; and cancer hospitals. Various sections of the Balanced Budget Act of 1997 (Pub. L. 105-33), the Medicare, Medicaid and SCHIP [State Children's Health Insurance Program] Balanced Budget Refinement Act of 1999 (Pub. L. 106-113), and the Medicare, Medicaid, and SCHIP Benefits Improvement and Protection Act of 2000 (Pub. L. 106-554) provide for the implementation of PPSs for rehabilitation hospitals and units (referred to as inpatient rehabilitation facilities (IRFs)), psychiatric hospitals and units, and LTCHs, as discussed below. Children's hospitals and cancer hospitals continue to be paid under reasonable cost-based reimbursement.

The existing regulations governing payments to excluded hospitals and hospital units are located in 42 CFR parts 412 and 413.

a. Inpatient Rehabilitation Facilities

Under section 1886(j) of the Act, as amended, rehabilitation hospitals and units (IRFs) have been transitioned from payment based on a blend of reasonable cost reimbursement subject to a hospital-specific annual limit under section 1886(b) of the Act and prospective payments for cost reporting periods beginning January 1, 2002 through September 30, 2002, to payment on a full prospective payment system basis effective for cost reporting periods beginning on or after October 1, 2002 (66 FR 41316, August 7, 2001 and 67 FR 49982, August 1, 2002). The existing regulations governing payments under the IRF PPS are located in 42 CFR part 412, subpart P.

b. LTCHs

Under the authority of sections 123(a) and (c) of Public Law 106-113 and section 307(b)(1) of Public Law 106-554, LTCHs are being transitioned from being paid for inpatient hospital services based on a blend of reasonable cost-based reimbursement under section 1886(b) of the Act to fully Federal prospective rates during a 5-year period, beginning with cost reporting periods that start on or after October 1, 2002. For cost reporting periods beginning on or after October 1, 2006, LTCHs will be paid under the fully Federal prospective payment rate (the June 6, 2003 LTCH PPS final rule (68 FR 34122)). LTCHs may elect to be paid based on full PPS payments instead of a blended payment in any year during the 5-year transition period. The existing regulations governing payment under the LTCH PPS are located in 42 CFR part 412, subpart O.

c. Psychiatric Hospitals and Units

Sections 124(a) and (c) of Public Law 106-113 provide for the development of a per diem PPS for payment for inpatient hospital services furnished in psychiatric hospitals and units under the Medicare program, effective for cost reporting periods beginning on or after October 1, 2002. This system must include an adequate patient classification system that reflects the differences in patient resource use and costs among these hospitals and maintain budget neutrality. We are in the process of developing a proposed rule, to be followed by a final rule, to implement the PPS for psychiatric hospitals and units (referred to as inpatient psychiatric facilities (IPFs).

3. Critical Access Hospitals

Under sections 1814, 1820, and 1834(g) of the Act, payments are made to critical access hospitals (CAHs) (that is, rural hospitals or facilities that meet certain statutory requirements) for inpatient and outpatient services on a reasonable cost basis. Reasonable cost is determined under the provisions of section 1861(v)(1)(A) of the Act and existing regulations under 42 CFR parts 413 and 415.

4. Payments for Graduate Medical Education

Under section 1886(a)(4) of the Act, costs of approved educational activities are excluded from the operating costs of inpatient hospital services. Hospitals with approved graduate medical education (GME) programs are paid for the direct costs of GME in accordance with section 1886(h) of the Act; the amount of payment for direct GME costs for a cost reporting period is based on the hospital's number of residents in that period and the hospital's costs per resident in a base year. The existing regulations governing payments to the various types of hospitals are located in 42 CFR part 413.

B. Summary of the Provisions of the May 19, 2003 Proposed Rule

On May 19, 2003, we published a proposed rule in the Federal Register (68 FR 27154) that set forth proposed changes to the Medicare IPPS for operating costs and for capital-related costs in FY 2004. We also set forth proposed changes relating to payments for GME costs, payments to CAHs, and payments to providers classified as psychiatric hospitals and units that continue to be excluded from the IPPS and paid on a reasonable cost basis. These changes were proposed to be effective for discharges occurring on or after October 1, 2003.

The following is a summary of the major changes that we proposed and the issues we addressed in the May 19, 2003 proposed rule:

1. Changes to the DRG Reclassifications and Recalibrations of Relative Weights

As required by section 1886(d)(4)(C) of the Act, we proposed annual adjustments to the DRG classifications and relative weights. Based on analyses of Medicare claims data, we proposed to establish a number of new DRGs and make changes to the designation of diagnosis and procedure codes under other existing DRGs.

Among the proposed changes discussed were:

• Expansion of the number of DRGs that are split on the basis of the presence or absence of complications or comorbidities (CCs). The DRGs we proposed to split were: DRG 4 (Spinal Procedures) into proposed new DRGs 531 and 532 (Spinal Procedures With and Without CC, respectively); DRG 5 (Extracranial Vascular Procedures) into proposed new DRGs 533 and 534 (Extracranial Vascular Procedures With and Without CC, respectively); DRG 231 (Local Excision and Removal of Internal Fixation Devices Except Hip and Femur) into proposed new DRGs 537 and 538 (Local Excision and Removal of Internal Fixation Devices Except Hip and Femur With and Without CC, respectively); and DRG 400 (Lymphoma and Leukemia With Major O.R. Procedure) into proposed new DRGs 539 and 540 (Lymphoma and Leukemia With Major O.R. Procedure With and Without CC, respectively).
• Creation of a new DRG for patients with an intracranial vascular procedure and an intracranial hemorrhage. The DRG we proposed to create was DRG 528 (Intracranial Vascular Procedure With a Principal Diagnosis of Hemorrhage).
• Creation of two new DRGs, differentiated on the basis of the presence or absence of a CC, for craniotomy patients with only a vascular shunt procedure. The DRGs we proposed to create were DRGs 529 and 530 (Ventricular Shunt Procedure With CC and Without CC, respectively).
• Creation of two new DRGs to differentiate current DRG 514 (Cardiac Defibrillator Implant With Cardiac Catheterization) on the basis of whether the patient does or does not experience any of the following symptoms: acute myocardial infarction, heart failure, or shock. The new DRGs we proposed were DRG 535 (Cardiac Defibrillator Implant With Cardiac Catheterization and With Acute Myocardial Infarction, Heart Failure, or Shock) and DRG 536 (Cardiac Defibrillator Implant With Cardiac Catheterization and Without Acute Myocardial Infarction, Heart Failure, or Shock)
• Changes in the DRG assignment of certain congenital anomalies that currently result in patients being assigned to newborn DRGs even when the patient is actually an adult. We also proposed adding to the list of major problems in newborns that affect DRG assignment.
• Modification of DRG 492 (Chemotherapy With Acute Leukemia as Secondary Diagnosis) to include in this DRG cases receiving high-dose Interleukin-2 (IL-2) chemotherapy for patients with advanced renal cell cancer and advanced melanoma.

We also presented our analysis of applicants for add-on payments for high-cost new medical technologies and proposed a revision to the high-cost threshold for a new technology or medical service to qualify for add-on payments.

• We proposed to continue to make add-on payments for Xigris.
• We discussed new applications for add-on payments for FY 2004.
• We proposed to reduce the high-cost threshold for a new technology or medical service to qualify for add-on payments from 1 standard deviation above the geometric mean standardized charge for cases in the DRGs to which the new technology is assigned to 75 percent of 1 standard deviation.

2. Changes to the Hospital Wage Index

We proposed revisions to the wage index and the annual update of the wage data. Specific issues addressed in this section included the following:

• The FY 2004 wage index update, using wage data from cost reporting periods that began during FY 2000.
• Exclusion of the wage data for rural health centers (RHCs) and Federally qualified health centers (FQHCs) from the calculation of the FY 2004 wage index.
• Exclusion of paid hours associated with military and jury duty leave from the wage index calculation, and request for comments on possible exclusion of paid lunch or meal break hours.
• Revisions to the wage index based on hospital redesignations and reclassifications.
• Amendments to the timetable for reviewing and verifying the wage data that will be in effect for the FY 2005 wage index.

3. Other Decisions and Changes to the PPS for Inpatient Operating and GME Costs

In the proposed rule, we discussed several provisions of the regulations in 42 CFR Parts 412 and 413 and set forth certain proposed changes concerning the following:

• Expansion of the current postacute transfer policy to 19 additional DRGs.
• Clarification of our policies that would be applied to counting hospital beds and patient days, in particular with regard to the treatment of swing-beds and observation beds, for purposes of the IME and DSH adjustments.
• Changes in our policy relating to nursing and allied health education payments to wholly owned subsidiary educational institutions of hospitals.
• Clarification of our policy relating to application of redistribution of costs and community support funds in determining a hospital's resident training costs.
• A change in the amount of rural training time required for an urban hospital to qualify for an increase in the rural track FTE limitation.
• Inclusion of FTE residents training in rural tracks in a hospital's rolling average calculation.

4. PPS for Capital-Related Costs

We discussed the payment requirements for capital-related costs. We did not propose any changes to the policies on payments to hospitals for capital-related costs.

5. Changes for Hospitals and Hospital Units Excluded From the IPPS

We discussed the following proposed revisions and clarifications concerning excluded hospitals and hospital units and CAHs:

• Revisions to the operation of excluded grandfathered hospitals-within-hospitals in effect on September 30, 1999.
• Clarification of the classification criteria for LTCHs.
• Clarification of the policy on payments for laboratory services provided by a CAH to patients outside a CAH.

6. Determining Prospective Payment Operating and Capital Rates and Rate-of-Increase Limits

In the Addendum to the May 19, 2003 proposed rule, we proposed changes to the amounts and factors for determining the FY 2004 prospective payment rates for operating costs and capital-related costs. We also established the proposed threshold amounts for outlier cases. In addition, we addressed update factors for determining the rate-of-increase limits for cost reporting periods beginning in FY 2004 for hospitals and hospital units excluded from the PPS.

7. Impact Analysis

In Appendix A of the proposed rule, we set forth an analysis of the impact that the proposed changes would have on affected hospitals.

8. Recommendation of Update Factor for Hospital Inpatient Operating Costs

In Appendix B of the proposed rule, as required by sections 1886(e)(4) and (e)(5) of the Act, we provided our recommendations of the appropriate percentage changes for FY 2004 for the following:

• Large urban area and other area average standardized amounts (and hospital-specific rates applicable to SCHs and MDHs) for hospital inpatient services paid under the IPPS for operating costs.
• Target rate-of-increase limits to the allowable operating costs of hospital inpatient services furnished by hospitals and hospital units excluded from the IPPS.

9. Discussion of Medicare Payment Advisory Commission Recommendations

Under section 1805(b) of the Act, the Medicare Payment Advisory Commission (MedPAC) is required to submit a report to Congress, no later than March 1 of each year, that reviews and makes recommendations on Medicare payment policies. In the proposed rule, we discussed the MedPAC recommendations concerning hospital inpatient payment policies and presented our response to those recommendations. For further information relating specifically to the MedPAC March 1 report or to obtain a copy of the report, contact MedPAC at (202) 220-3700 or visit MedPAC's Web site at: http://www.medpac.gov.

C. Public Comments Received in Response to the May 19, 2003 IPPS Proposed Rule

We received approximately 4,200 timely items of correspondence containing multiple comments on the May 19, 2003 proposed rule. Summaries of the public comments and our responses to those comments are set forth below under the appropriate heading.

II. Changes to DRG Classifications and Relative Weights

A. Background

Section 1886(d) of the Act specifies that the Secretary shall establish a classification system (referred to as DRGs) for inpatient discharges and adjust payments under the IPPS based on appropriate weighting factors assigned to each DRG. Therefore, under the IPPS, we pay for inpatient hospital services on a rate per discharge basis that varies according to the DRG to which a beneficiary's stay is assigned. The formula used to calculate payment for a specific case multiplies an individual hospital's payment rate per case by the weight of the DRG to which the case is assigned. Each DRG weight represents the average resources required to care for cases in that particular DRG, relative to the average resources used to treat cases in all DRGS.

Congress recognized that it would be necessary to recalculate the DRG relative weights periodically to account for changes in resource consumption. Accordingly, section 1886(d)(4)(C) of the Act requires that the Secretary adjust the DRG classifications and relative weights at least annually. These adjustments are made to reflect changes in treatment patterns, technology, and any other factors that may change the relative use of hospital resources. Changes to the DRG classification system and the recalibration of the DRG weights for discharges occurring on or after October 1, 2003 are discussed below.

B. DRG Reclassification

1. General

Cases are classified into DRGs for payment under the IPPS based on the principal diagnosis, up to eight additional diagnoses, and up to six procedures performed during the stay. In a small number of DRGs, classification is also based on the age, sex, and discharge status of the patient. The diagnosis and procedure information is reported by the hospital using codes from the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM).

For FY 2003, cases are assigned to one of 510 DRGs in 25 major diagnostic categories (MDCs). Most MDCs are based on a particular organ system of the body. For example, MDC 6 is Diseases and Disorders of the Digestive System. This approach is used because clinical care is generally organized in accordance with the organ system affected. However, some MDCs are not constructed on this basis because they involve multiple organ systems (for example, MDC 22 (Burns)). The table below lists the 25 MDCs.

In general, cases are assigned to an MDC based on the patient's principal diagnosis before assignment to a DRG. However, for FY 2003, there are eight DRGs to which cases are directly assigned on the basis of ICD-9-CM procedure codes. These DRGs are for heart, liver, bone marrow, lung, simultaneous pancreas/kidney, and pancreas transplants (DRGs 103, 480, 481, 495, 512, and 513, respectively) and for tracheostomies (DRGs 482 and 483). Cases are assigned to these DRGs before they are classified to an MDC.

Within most MDCs, cases are then divided into surgical DRGs and medical DRGs. Surgical DRGs are based on a hierarchy that orders operating room (O.R.) procedures or groups of O.R. procedures by resource intensity. Medical DRGs generally are differentiated on the basis of diagnosis and age (less than or greater than 17 years of age). Some surgical and medical DRGs are further differentiated based on the presence or absence of a complication or a comorbidity (CC).

Generally, nonsurgical procedures and minor surgical procedures that are not usually performed in an operating room are not treated as O.R. procedures. However, there are a few non-O.R. procedures that do affect DRG assignment for certain principal diagnoses, for example, extracorporeal shock wave lithotripsy for patients with a principal diagnosis of having urinary stones.

Patient's diagnosis, procedure, discharge status, and demographic information is fed into the Medicare claims processing systems and subjected to a series of automated screens called the Medicare Code Editor (MCE). The MCE screens are designed to identify cases that require further review before classification into a DRG.

After patient information is screened through the MCE and any further development of the claim is conducted, cases are classified into the appropriate DRG by the Medicare GROUPER software program. The GROUPER program was developed as a means of classifying each case into a DRG on the basis of the diagnosis and procedure codes and, for a limited number of DRGs, demographic information (that is, sex, age, and discharge status).

After cases are screened through the MCE and assigned to a DRG by the GROUPER, a base DRG payment is calculated by the PRICER software. The PRICER calculates the payments for each case covered by the IPPS based on the DRG relative weight and additional factors associated with each hospital, such as IME and DSH adjustments. These additional factors increase the payment amount to hospitals above the base DRG payment.

The records for all Medicare hospital inpatient discharges are maintained in the Medicare Provider Analysis and Review (MedPAR) file. The data in this file are used to evaluate possible DRG classification changes and to recalibrate the DRG weights. However, in the July 30, 1999 IPPS final rule (64 FR 41500), we discussed a process for considering non-MedPAR data in the recalibration process. In order for us to consider the feasibility of using particular non-MedPAR data, we must have sufficient time to evaluate and test the data. The time necessary to do so depends upon the nature and quality of the non-MedPAR data submitted. Generally, however, a significant sample of the non-MedPAR data should be submitted by mid-October for consideration in conjunction with the next year's proposed rule. This allows us time to test the data and make a preliminary assessment as to the feasibility of using the data. Subsequently, a complete database should be submitted by early December for consideration in conjunction with the next year's proposed rule.

Many of the changes to the DRG classifications are the result of specific issues brought to our attention by interested parties. We encourage individuals with concerns about DRG classifications to bring those concerns to our attention in a timely manner so they can be carefully considered for possible inclusion in the next proposed rule and so any proposed changes may be subjected to public review and comment. Therefore, similar to the timetable for interested parties to submit non-MedPAR data for consideration in the DRG recalibration process, concerns about DRG classification issues should be brought to our attention no later than early December in order to be considered and possibly included in the next annual proposed rule updating the IPPS.

In the May 19, 2003 proposed rule, we proposed numerous changes to the DRG classification system for FY 2004. The changes we proposed to the DRG classification system for FY 2004, the public comments we received concerning the proposed changes, the final DRG changes, and the methodology used to recalibrate the DRG weights are set forth below. The changes we are implementing in this final rule will be reflected in the revised FY 2004 GROUPER version 21.0 and effective for discharges occurring on or after October 1, 2003. Unless otherwise noted in this final rule, our DRG analysis is based on data from the March 2002 update of the FY 2002 MedPAR file, which contains hospital bills received through March 31, 2002, for discharges in FY 2002.

2. Review of DRGs for a Split Based on Presence or Absence of a CC

In an effort to improve the clinical and cost cohesiveness of the DRG classification system, we have evaluated whether additional DRGs should be split based on the presence or absence of a CC. There are currently 116-paired DRGs that reflect a split based on the presence or absence of a CC. We last performed a systematic evaluation and considered changes to the DRGs to recognize the within-DRG cost differences based on the presence or absence of CCs in 1994 (May 27, 1994 IPPS proposed rule, 59 FR 27715). In the May 27, 1994 IPPS proposed rule, we described a refined DRG system based on a list of secondary diagnoses that have a major effect on the resources that hospitals use to treat patients across DRGs. We analyzed how the presence of the secondary diagnosis affected resource use compared to other secondary diagnoses, and classified these secondary diagnoses as non-CC, CC, or major CC. After finalizing the classification of secondary diagnoses, we evaluated which collapsed DRGs should be split based on the presence of a major CC, other CC, or both. However, we did not implement this refined system because we did not believe it would be prudent policy to make changes for which we could not predict the effect on the case-mix (the average DRG relative weight for all cases) and, thus, payments (60 FR 29209). We were concerned that we would be unable to fulfill the requirement of section 1886(d)(4)(C)(iii) of the Act that aggregate payments may not be affected by DRG reclassification and recalibration of weighting factors. That is, our experience has been that hospitals respond to major changes to the DRGs by changing their coding practices in ways that increase total payments (for example, by beginning to include ICD-9-CM codes that previously did not affect payment for a case). Because changes in coding behavior do not represent a real increase in the severity of the overall mix of cases, total payments should not increase. We believe that the only way to ensure this behavioral response does not lead to higher total payments is to make an offsetting adjustment to the system in advance of the fiscal year for which the changes are effective.

Section 301(e) of the Medicare, Medicaid, and SCHIP Benefits Improvement and Protection Act of 2000 (Pub. L. 106-554) authorized the Secretary to make such a prospective adjustment to the average standardized amounts for discharges occurring on or after October 1, 2001, to ensure the total payment impacts of changes to the DRGs do not result in any more or less total spending than would otherwise occur without the changes (budget neutrality).

We are not proceeding with implementing a refined DRG system at this time, pending a decision whether to replace the ICD-9-CM coding system with another classification system. The refined DRG system discussed in the May 1994 IPPS proposed rule involved a complete and thorough assessment of all of the ICD-9-CM diagnosis codes in order to establish an illness severity level associated with each code. Rather than undertaking the time-consuming process of establishing illness severity levels for all ICD-9-CM codes at this time, we believe the more prudent course would be to delay this evaluation pending the potential replacement of ICD-9-CM. For example, the National Committee on Health and Vital Statistics (NCHVS) is considering making a recommendation to the Secretary on whether to recommend the adoption of the ICD-10-CM and the ICD-10—Procedure Coding System (PCS) as the national uniform standard coding system for inpatient reporting.

In the meantime, we have undertaken an effort to identify additional DRGs where a CC split appears most justified. Our analysis identified existing DRGs that meet the following criteria: a reduction in variance in charges within the DRG of at least 4 percent; fewer than 75 percent of all patients in the current DRG would be assigned to the with-CC DRG; and the overall payment impact (higher payments for cases in the with-CC DRG offset by lower payments for cases in the without-CC DRG) is at least $40 million.

The following four DRGs meet these criteria: DRG 4 (Spinal Procedures) and DRG 5 (Extracranial Vascular Procedures) in MDC 1 (Diseases and Disorders of the Nervous System); DRG 231 (Local Excision and Removal of Internal Fixation Devices Except Hip and Femur) in MDC 8 (Diseases and Disorders of the Musculoskeletal and Connective Tissue); and DRG 400 (Lymphoma and Leukemia with Major O.R. Procedure) in MDC 17 (Myeloproliferative Diseases and Disorders and Poorly Differentiated Neoplasms).

The following data indicate that the presence or absence of a CC was found to have a significant impact on patient charges and on average lengths of stay in these four DRGs.

Therefore, we proposed to establish the following new DRGs: proposed DRG 531 (Spinal Procedures With CC) and proposed DRG 532 (Spinal Procedures Without CC) in MDC 1; proposed DRG 533 (Extracranial Procedures With CC) (the proposed rule incorrectly included “Vascular” in the title) and proposed DRG 534 (Extracranial Procedures Without CC) (the proposed rule incorrectly included “Vascular” in the title) in MDC 1; proposed DRG 537 (Local Excision and Removal of Internal Fixation Devices Except Hip and Femur With CC) and proposed DRG 538 (Local Excision and Removal of Internal Fixation Devices Except Hip and Femur Without CC) in MDC 8; and proposed DRG 539 (Lymphoma and Leukemia With Major O.R. Procedure With CC) and DRG 540 (Lymphoma and Leukemia With Major O.R. Procedure Without CC) in MDC 17. We proposed that DRGs 4, 5, 231, and 400 would become invalid.

Comment: Seven commenters supported the proposed expansion of the number of DRGs related to spinal procedures and extracranial vascular procedures and the removal of internal fixation devices. One commenter commended CMS for the proposed change to payments for implanting spinal code stimulation devices. Referring to proposed new DRGs 531 and 532, the commenter stated that most inpatients receiving a spinal cord stimulator implant have a comorbid condition, which adds significantly to the cost of care and can serve as a barrier to patient access. Another commenter specifically supported the new DRGs 533 and 534 for extracranial vascular procedures.

One commenter expressed support for CMS' recognition of cost differences within a given DRG based on the presence or absence of a CC and encouraged CMS to continue to consider secondary diagnoses that can have a substantial effect on hospital resources when restructuring DRGs based on cost considerations.

Response: We appreciate the support for these proposals and are adopting them as final without further modification.

We are establishing new DRGs 531, 532, 533, 534, 537, 538, 539, and 540, effective for discharges occurring on or after October 1, 2003. As a result of establishing these new DRGS, DRGs 4, 5, 231, and 400 are invalid, effective October 1, 2003. We will continue to monitor whether additional DRGs should be split based on the presence or absence of a CC.

3. MDC 1 (Diseases and Disorders of the Nervous System)

a. Revisions of DRGs 1 and 2

In the FY 2003 IPPS final rule, we split DRGs 1 and 2 (Craniotomy Age > 17 With and Without CC, respectively) based on the presence or absence of a CC (67 FR 49986). We have received several proposals related to devices or procedures that are used in a small subset of cases from these DRGS. These proposals argue that the current payment for these devices or procedures under DRGs 1 and 2 is inadequate.

Therefore, we conducted an analysis of the charges for various procedures and diagnoses within DRGs 1 and 2 to assess whether further changes to these DRGs may be warranted. Currently, the average charges for cases assigned to DRGs 1 and 2 are approximately $55,000 and $30,000, respectively. In the May 19, 2003 proposed rule, we proposed to create two separate new DRGs for: (1) cases with an intracranial vascular procedure and a principal diagnosis of an intracranial hemorrhage; and (2) craniotomy cases with a ventricular shunt procedure (absent another procedure). The former set of cases are much more expensive than those presently in DRGs 1 and 2; the latter set of cases are much less expensive.

(1) Intracranial Vascular Procedures

Our analysis indicated that patients with an intracranial vascular procedure and a principal diagnosis of an intracranial hemorrhage were significantly more costly than other cases in DRGS 1 and 2. These patients have an acute condition with a high severity of illness and risk of mortality. There were 917 cases in DRGs 1 and 2 with an intracranial vascular procedure and a principal diagnosis of hemorrhage with average charges of approximately $113,884, which are much higher than the average charges of DRGS 1 and 2 noted above.

We also found 890 cases that had an intracranial vascular procedure without a principal diagnosis of hemorrhage (for example, nonruptured aneurysms). These cases are generally less acutely ill than those involving ruptured aneurysms, and have a lower risk of mortality. Among these 890 cases, the average charges were approximately $52,756, which are much more similar to the average charges for all cases in DRGs 1 and 2.

Based on this analysis, we proposed to create new DRG 528 (Intracranial Vascular Procedure With a Principal Diagnosis of Hemorrhage) for patients with an intracranial vascular procedure and an intracranial hemorrhage. We proposed that cases involving intracranial vascular procedures without a principal diagnosis of hemorrhage would remain in DRGs 1 and 2.

We indicated that proposed new DRG 528 would have the following principal diagnoses:

• 094.87, Syphilitic ruptured cerebral aneurysm
• 430, Subarachnoid hemorrhage
• 431, Intracerebral hemorrhage
• 432.0, Nontraumatic extradural hemorrhage
• 432.1, Subdural hemorrhage
• 432.9, Unspecified intracranial hemorrhage

And operating room procedures:

• 02.13, Ligation of meningeal vessel
• 38.01, Incision of vessel, intracranial vessels
• 38.11, Endarterectomy, intracranial vessels
• 38.31, Resection of vessel with anastomosis, intracranial vessels
• 38.41, Resection of vessel with replacement, intracranial vessels
• 38.51, Ligation and stripping of varicose veins, intracranial vessels
• 38.61, Other excision of vessels, intracranial vessels
• 38.81, Other surgical occlusion of vessels, intracranial vessels
• 39.28, Extracranial-intracranial (EC-IC) vascular bypass
• 39.51, Clipping of aneurysm
• 39.52, Other repair of aneurysm
• 39.53, Repair of arteriovenous fistula
• 39.72, Endovascular repair or occlusion of head and neck vessels
• 39.79, Other endovascular repair of aneurysm of other vessels

(2) Ventricular Shunt Procedures

We also found that craniotomy patients who had a ventricular shunt procedure (absent another procedure) were significantly less costly than other craniotomy patients in DRGs 1 and 2. Ventricular shunts are normally performed for draining intracranial fluid. A ventricular shunt is a less extensive procedure than the other intracranial procedures in DRGs 1 and 2. As a result, if a ventricular shunt is the only intracranial procedure performed, these cases will typically be less costly.

There were 4,373 cases in which only ventricular shunt procedures were performed. These cases had average charges of approximately $27,188. However, the presence or absence of a CC had a significant impact on patient charges and lengths of stay. There were 2,533 cases with CC, with average charges of approximately $33,907 and an average length of stay of 8.2 days. In contrast, there were 1,840 cases without CC, with average charges of approximately $17,939 and an average length of stay of 3.7 days.

Therefore, we proposed to create two new DRGs, splitting with CC and without CC, for patients with only a vascular shunt procedure: proposed new DRG 529 (Ventricular Shunt Procedures With CC) and proposed new DRG 530 (Ventricular Shunt Procedures Without CC).

We indicated that proposed new DRG 529 would consist of any principal diagnosis in MDC 1 (erroneously cited as MDC 5 in the proposed rule), with the presence of a CC and one of the following operating room procedures:

• 02.31, Ventricular shunt to structure in head and neck
• 02.32, Ventricular shunt to circulatory system
• 02.33, Ventricular shunt to thoracic cavity
• 02.34, Ventricular shunt to abdominal cavity and organs
• 02.35, Ventricular shunt to urinary system
• 02.39, Other operations to establish drainage of ventricle
• 02.42, Replacement of ventricular shunt
• 02.43, Removal of ventricular shunt

We proposed that the proposed new DRG 530 would consist of any principal diagnosis in MDC 1 (erroneously cited as MDC 5 in the proposed rule) with one of the operating room procedures listed above for the proposed new DRG 529, but without the presence of a CC.

Comment: Four commenters supported the proposed creation of two DRGs to capture ventricular shunt procedures. Ten commenters supported the proposed creation of new DRG 528 for an intracranial vascular procedure with a principal diagnosis of hemorrhage.

Two commenters requested that CMS verify its GROUPER analysis and clarify in the final rule the estimated number of cases that will be assigned to DRG 528. One commenter also believed that CMS is underestimating the volume of hemorrhagic cases that would be assigned to this new DRG. The commenter indicated that its analysis of MedPAR 2001 data demonstrated 1,550 cases.

Response: We conducted an analysis based on later available MedPAR data and found 1,596 cases that would be assigned to DRG 528 (based on a full year of MedPAR data). This volume is consistent with the commenter's analysis, although different MedPAR files were used in the analysis. In the proposed rule (68 FR 27161), we reported 917 cases based on preliminary data (6 months' worth of cases) that we analyzed when we considered the proposed change in the DRG classification. There were actually 1,354 cases grouped to the proposed new DRG 528 for the proposed rule.

Comment: One commenter suggested the creation of a new companion DRG to DRG 528 for intracranial vascular procedures for unruptured cerebral aneurysms. The commenter was concerned that the charges for endovascular repair of unruptured aneurysms is higher than other procedures currently assigned to DRG 2.

Response: The average charges for unruptured aneurysm cases varied according to the DRG to which the cases were assigned. The average charges for these cases in DRG 1 were slightly higher than the overall charges for that DRG, of approximately $69,682 and $54,900, respectively. However, we found that these charges are consistent with the variation of charges within this DRG and, therefore, did not propose a change in the DRG reclassification. Similarly, for cases assigned to DRG 2, we found the average charges of approximately $36,077 are consistent with the overall average charges of that DRG of approximately $32,000. We will continue to monitor these cases.

Comment: Three commenters requested a change to the DRG assignment of cases involving implantation of GLIADEL® chemotherapy wafers to treat brain tumors. One of the commenters offered two options: create a new DRG or reassign these cases to DRG 484 (Craniotomy for Multiple Significant Trauma). The commenter cited an example in which CMS has in the past grouped together in the same DRG cases that are clinically dissimilar but similar in resource intensity when there were no other options available. For FY 1998 (62 FR 45974), coronary stent cases were moved from DRG 112 (Percutaneous Cardiovascular Procedures) to DRG 116 (Other Permanent Cardiac Pacemaker Implant or PTCA with Coronary Artery Stent Implant). In that instance, CMS concluded that, although coronary artery stent cases are not clinically similar to the pacemaker cases in DRG 116, the resource consumption of these cases is very similar. The commenter contended that, absent another appropriate craniotomy DRG, the same argument could be applied to assigning cases with GLIADEL® wafer to DRG 484.

In a comment on the proposed rule, the manufacturer of this implant provided estimated FY 2003 average costs and charges for these cases. Its report indicated that the costs of the cases of $24,280 would be the same for cases assigned to DRG 1 and DRG 2, and the charges of the cases of $50,394 would be the same for both DRGs. The manufacturer requested that we analyze the available data in the FY 2003 MedPAR file to identify GLIADEL® cases. The manufacturer believed these data support the need for a DRG change.

One commenter agreed with our determination that this technology is currently reflected within the DRG weights and does not meet the definition of a new technology.

Response: In our analysis of the data from the March 2003 update of the FY 2003 MedPAR file, we found a total of 61 cases in which the ICD-9-CM procedure code 00.10 (Implantation of a chemotherapeutic agent) was reported for cases assigned to DRGs 1 and 2. There were 38 cases assigned to DRG 1 and 23 cases assigned to DRG 2. Consistent with the GROUPER logic for these DRGs that splits cases based on the presence or absence of CCs, we found that the average standardized charges in DRGs 1 and 2 were approximately $64,864 and $42,624, respectively. We believe that while the charges for GLIADEL® wafer cases may be higher than the average standardized charges for DRG 2, they are within the normal variation of the overall charges within each DRG.

We note that the DRGs are a system of averages, and there is expected to be variation in the average charges for different procedures and services across all DRGs. Hospitals are expected to be able to finance some higher cost procedures with lower cost procedures within the same DRG as well as across DRGs. Although the average charges of the cases we identified in our analysis are somewhat higher than the average charges of all cases in these DRGs, they are within the range of other procedures included in these DRGs. By way of comparison, we are creating a new DRG for cases with an intracranial vascular procedure and a principal diagnosis of an intracranial hemorrhage on the basis of our analysis that showed the average charges for these cases were $113,884. This is approximately $59,000 more than the average charges in DRG 1 (more than the total charges for the GLIADEL® cases reported by the commenter) and approximately $84,000 more than the average charges in DRG 2.

We also are concerned that there may be insufficient volume of cases to warrant the establishment of a new DRG for this technology. Thus, before considering the creation of a new DRG for these cases, we would like to review a full year of data, as well as consider alternative options if they appear warranted. It would also be necessary to provide opportunity for public comment on any potential changes to the DRG assignment of these cases before proceeding with a final change.

Currently, DRG 484 includes complex, multiple significant trauma cases; that is, patients with a principal diagnosis of trauma and at least two significant trauma diagnosis codes (either as principal or secondary diagnosis) from different body site categories. While this DRG includes craniotomy, it is assigned to MDC 24 (Multiple Significant Trauma). While the treatment for glioblastoma multiforme is significant, we do not believe these cases are clinically similar to other cases currently assigned to DRG 484.

We also are concerned that there may be insufficient volume to warrant the establishment of a new DRG for this technology, and we would like to review a full year of data, as well as consider alternative options if they appear warranted. It also would be necessary to provide opportunity for public comment on any potential changes before proceeding with a final change.

Comment: Two commenters pointed out a typographical error in our proposal. The commenters indicated that we proposed new DRGs 529 and 530 for placement in MDC 5; the correct MDC should have been MDC 1.

Response: We agree with the commenters and have corrected this placement, as indicated in the discussion above.

After consideration of the comments received, we are adopting as final the three new proposed DRGs 528, 529, and 530. These DRGS will be effective for discharges occurring on or after October 1, 2003.

b. DRG 23 (Nontraumatic Stupor and Coma)

In DRG 23 (Nontraumatic Stupor and Coma), there are currently six principal diagnoses identified by the following ICD-9-CM diagnosis codes: 348.4, Compression of the brain; 348.5, Cerebral edema; 780.01, Coma; 780.02, Transient alteration of awareness; 780.03, Persistent vegetative state; and 780.09, Other alteration of consciousness. Code 780.02 is often used to describe the diagnosis of psychiatric patients rather than the diagnosis of patients with severe neurological disorders. The treatment plan for a patient with “transient alteration of awareness” is clinically very different from the treatment plan for a coma patient. Furthermore, many patients with this diagnosis are treated in psychiatric facilities rather than in acute care hospitals.

Although there are neurological patients who present with the complaint of “transient alteration of awareness,” the cause of this alteration of consciousness is commonly identified, and the principal diagnosis for the hospital admission is the etiology of the alteration of consciousness rather than the symptom itself. For the few remaining neurological patients for whom the cause is not identified and for whom code 780.02 is assigned as the principal diagnosis, we believe that the care of these patients is different than the care of patients with coma or cerebral edema.

Because we believe the patients with a principal diagnosis of “transient alteration of consciousness” are more clinically related to the patients in DRG 429 (Organic Disturbances and Mental Retardation) in MDC 19 (Mental Diseases and Disorders), we proposed that patients who are assigned a principal diagnosis of code 780.02 would be assigned to DRG 429 instead of DRG 23. DRG 429 also contains similar diagnoses, such as code 293.81, Organic delusional syndrome and code 293.82, Organic hallucinosis syndrome. (We note that the charges for the patient cases in DRGs 23 and 429 are very similar ($11,559 and $11,713, respectively), so the proposed movement of code 780.02 from DRG 23 to DRG 429 would have minimal payment impact.) Moving this diagnosis code as proposed would also consolidate diagnoses treated frequently in psychiatric hospitals in those DRGs that are likely to be a part of the upcoming proposed Medicare psychiatric facility PPS.

Comment: An organization representing hospitals supported our proposed change, while other commenters opposed the change. The commenters who opposed the change stated that code 780.02 is included in the ICD-9-CM chapter for signs and symptoms of ill-defined conditions. The commenters believed that since this code is included in a chapter with ill-defined conditions, it would be inappropriate to move the code to DRG 429. The commenters stated that this code does not describe a mental disorder; and disagreed with our statement in the proposed rule that code 780.02 was similar to codes 293.81 and 293.82. The commenters further stated that they disagreed with our assertion that many patients with a diagnosis of transient alteration of awareness are treated in psychiatric facilities.

Response: Our review of claims data indicates that code 780.02 is a frequent diagnosis for patients admitted to psychiatric hospitals. Many patients are likely to present with transient alteration of awareness at the time of admission to a psychiatric hospital. The cause of this transient alteration is likely to be diagnosed during the stay, leading to the assignment of another, more specific principal diagnosis.

However, in many patients, this is not the case, and no underlying cause for the transient alteration of awareness is determined. When a more definitive diagnosis cannot be made, the patient is left with the diagnosis of alteration of awareness. We recognize the difficulty in assigning symptoms such as these to the most appropriate DRG. However, we will note that the average charges for DRG 23 (where the code is currently assigned) and DRG 429 are similar.

Therefore, we are proceeding with the assignment of code 780.02 to DRG 429 based on a review of psychiatric hospital data as well as a clinical comparison of cases already assigned to DRG 429.

4. MDC 5 (Diseases and Disorders of the Circulatory System)

a. DRG 478 (Other Vascular Procedures With CC) and DRG 479 (Other Vascular Procedures Without CC)

Code 37.64 (Removal of heart assist system) in DRGs 478 and 479 describes the operative, as opposed to bedside, removal of a heart assist system. Based on comments we received suggesting that code 37.64 was inappropriately assigned to DRGs 478 and 479, we reviewed the MedPAR data for both DRGs 478 and 479 and DRG 110 (Major Cardiovascular Procedures With CC) and DRG 111 (Major Cardiovascular Procedures Without CC) to assess the appropriate assignment of code 37.64.

We found that there were only 17 cases of code 37.64 in DRGs 478 and 479, with an average length of stay of 14.1 days and average charges of $105,153. There were a total of 90,591 cases in DRGs 478 and 479 that did not contain code 37.64. These cases had an average length of stay of 6.6 days and average charges of $31,879. In DRGs 110 and 111, we found an average length of stay of 8.1 days, with average charges of $54,653.

We proposed to remove code 37.64 from DRGs 478 and 479 and reassign it to DRGs 110 and 111. The surgical removal of a heart assist system is a major cardiovascular procedure and, therefore, more appropriately assigned to DRGs 110 and 111. Accordingly, we believe this DRG assignment for this procedure is more clinically and financially appropriate.

We received two comments in support of this change. Therefore, we are adopting as final our proposal to remove code 37.64 from DRGs 478 and 479 and assign it to DRGs 110 and 111.

b. DRGs 514 (Cardiac Defibrillator Implant With Cardiac Catheterization) and 515 (Cardiac Defibrillator Implant Without Cardiac Catheterization)

(1) Cardiac Defibrillator Implant With Cardiac Catheterization With Acute Myocardial Infarction

Prior to the publication of the proposed rule, we received a recommendation to modify DRG 514 (Cardiac Defibrillator Implant With Cardiac Catheterization) and DRG 515 (Cardiac Defibrillator Implant Without Cardiac Catheterization) so that these DRGs are split based on the presence or absence of acute myocardial infarction, heart failure, or shock as a principal diagnosis. We note that the increased cost of treating cardiac patients with acute myocardial infarction, heart failure, or shock is recognized in the payment logic for pacemaker implants (DRG 115 (Permanent Cardiac Pacemaker Implant With Acute Myocardial Infarction, Heart Failure or Shock, or AICD Lead or Generator) and DRG 116 (Other Permanent Cardiac Pacemaker Implant)).

We examined FY 2002 MedPAR data regarding the number of cases and the average charges for DRGs 514 and 515. The results of our examination are summarized in the following table.

A cardiac catheterization is generally performed to establish the nature of the patient's cardiac problem and determine if implantation of a cardiac defibrillator is appropriate. Generally, the cardiac catheterization can be done on an outpatient basis. Patients who are admitted with acute myocardial infarction, heart failure, or shock and have a cardiac catheterization are generally acute patients who require emergency implantation of the defibrillator. Thus, there are very high costs associated with these patients.

We found that the average charges for patients with cardiac catheterizations who also were admitted with acute myocardial infarction, heart failure, or shock were $120,852, compared to the average charges for all DRG 514 cases of $97,133. Therefore, we proposed to split DRG 514 and create a new DRG for patients receiving a cardiac defibrillator implant with cardiac catheterization and with a principal diagnosis of acute myocardial infarction, heart failure, or shock.

Patients without cardiac catheterization generally have had the need for the defibrillator established on an outpatient basis prior to admission. We found 935 cases with acute myocardial infarction, heart failure, or shock, with average charges of $84,140. The average charges for all cases in DRG 515 were $76,537. Because of the relatively small number of patients and the less-than-10-percent charge difference for patients in DRG 515 who have acute myocardial infarction, heart failure, or shock, we did not propose to create a separate DRG for patients with a cardiac defibrillator implant without cardiac catheterization with acute myocardial infarction, heart failure, or shock.

Specifically, we proposed to create two new DRGs that would replace the current DRG 514. We indicated that the two proposed new DRGs would have the same procedures currently listed for DRG 514, but would be split based on the presence or absence of acute myocardial infarction, heart failure, or shock as a principal diagnosis. We proposed to establish new DRG 535 (Cardiac Defibrillator Implant With Cardiac Catheterization and With Acute Myocardial Infarction, Heart Failure, or Shock) and new DRG 536 (Cardiac Defibrillator Implant With Cardiac Catheterization and Without Acute Myocardial Infarction, Heart Failure, or Shock). Proposed new DRG 536 would exclude the following principal diagnosis codes from MDC 5 associated with acute myocardial infarction, heart failure, or shock.

• 398.91, Rheumatic heart failure
• 402.01, Malignant hypertensive heart disease with heart failure
• 402.11, Benign hypertensive heart disease with heart failure
• 402.91, Hypertensive heart disease not otherwise specified with heart failure
• 404.01, Malignant hypertensive heart and renal disease with heart failure
• 404.03, Malignant hypertensive heart and renal disease with heart failure and renal failure
• 404.11, Benign hypertensive heart and renal disease with heart failure
• 404.13, Benign hypertensive heart and renal disease with heart failure and renal failure
• 404.91, Hypertensive heart and renal disease not otherwise specified with heart failure
• 404.93, Hypertensive heart and renal disease not otherwise specified with heart failure and renal failure
• 410.01, AMI anterolateral, initial
• 410.11, AMI anterior wall, initial
• 410.21, AMI inferolateral, initial
• 410.31, AMI inferopost, initial
• 410.41, AMI inferior wall, initial
• 410.51, AMI lateral not elsewhere classified, initial
• 410.61, True posterior infarction, initial
• 410.71, Subendocardial infarction, initial
• 410.81, AMI not elsewhere classified, initial
• 410.91, AMI not otherwise specified, initial
• 428.0, Congestive heart failure, not otherwise specified
• 428.1, Left heart failure
• 428.20, Systolic heart failure, not otherwise specified
• 428.21, Acute systolic heart failure
• 428.22, Chronic systolic heart failure
• 428.23, Acute on chronic systolic heart failure
• 428.30, Diastolic heart failure, not otherwise specified
• 428.31, Acute diastolic heart failure
• 428.32, Chronic diastolic heart failure
• 428.33, Acute on chronic diastolic heart failure
• 428.40, Combined systolic and diastolic heart failure not otherwise specified
• 428.41, Acquired combined systolic and diastolic heart failure
• 428.42, Chronic combined systolic and diastolic heart failure
• 428.43, Acute on chronic combined systolic and diastolic heart failure
• 428.9, Heart failure, not otherwise specified
• 785.50, Shock, not otherwise specified
• 785.51, Cardiogenic shock

(2) Cardiac Resynchronization Therapy (CRT)

Prior to the publication of the proposed rule, we received a comment from a provider who pointed out that we did not include the following combination of codes under the list of procedure combinations that would lead to an assignment of DRG 514 or DRG 515:

• 37.95, Implantation of automatic cardioverter/defibrillator lead(s) only
• 00.54, Implantation or replacement of cardiac resynchronization defibrillator, pulse generator device only [CRT-D]

The commenter pointed out that cases are assigned to DRGs 514 and 515 when a total cardiodefibrillator or CRT-D system is implanted. In addition, cases are assigned to DRGs 514 and 515 when implantation of a variety of combinations of defibrillator leads and device combinations is reported. The commenter indicated that a total defibrillator and CRT-D system may be replaced with a completely new system or all new devices and leads, and added that it is also possible to replace a generator, a lead, or a combination of generators and up to three leads.

When the CRT-D generator (code 00.54) and one of the cardioverter/defibrillator leads are replaced, the case currently is assigned to DRG 115 (Permanent Cardiac Pacemaker Implant with AMI, Heart Failure, or Shock or AICD Lead or Generator Procedure). The commenter recommended that we include the combination of codes 37.95 and 00.54 as a combination that would result in assignment to DRG 514 or DRG 515, as do other combinations of generators and leads. Our medical advisors agree with this recommendation. As discussed previously, we proposed to delete DRG 514 and replace it with proposed new DRGs 535 and 536. Therefore, we proposed to add codes 37.95 and 00.54 to the list of procedure combinations that would result in assignment to DRG 515 or new proposed DRGs 535 and 536.

Comment: Several commenters supported our proposed revision to DRG 514 so that it would be split based on the presence or absence of a principal diagnosis of acute myocardial infarction, heart failure, or shock.

One commenter pointed out a typographical error in the proposed rule in the code number cited for the procedure, Implantation of automatic cardioverter/defibrillator lead(s) only. The code number should have been 37.95 instead of 39.75.

Response: We appreciate the support for our proposed revision of DRG 514. We have corrected the code number for Implantation of automatic cadioverter/defibrillator lead(s) only to 37.95 in the description of this issue above.

Comment: Several commenters supported the addition of codes 37.95 and 00.54 to the list of procedure combinations that would lead to an assignment of DRG 515 and new DRGs 535 and 536. However, one commenter suggested that, in addition to this combination, codes 37.97 (Replacement of automatic cardioverter/defibrillator lead(s) only and 00.54 also should be added to the procedure combination list under DRG 515 and new DRGs 535 and 536. The commenter pointed out that both procedures would involve the insertion of a pulse generator and a lead so that resources required are equivalent to those for a total system implant.

Response: We agree with the commenter that the combination of codes 37.97 and 00.54 also would involve the implantation of a pulse generator and a lead. Therefore, in this final rule, we are adding the combination of procedure codes 37.97 and 00.54 to the list of procedure combinations that will lead to assignment to DRG 515 and new DRGs 535 and 536.

Comment: One commenter recommended that CMS also consider modifying DRGs 115 and 116 to recognize more combination groups of devices and leads. Specifically, the commenter recommended adding the following combination of codes to the list of procedure combinations under DRGs 115 and 116:

• 00.53, Implantation or replacement of CRT-P pulse generator only
• 37.74, Implantation or replacement of epicardial pacemaker lead.

Response: DRGs 115 and 116 have one of the most complex assignment structures of all the DRGs. The DRG logic for DRGs 115 and 116 involves three separate combinations of code groups that can possibly lead to these DRG assignments. Before making a modification to one of the combination groups (particularly the procedure combinations), we believe we should analyze the impact of a modification to the currently existing types of device, lead, and diagnosis combinations. In the future, we will undertake a close review of DRGs 115 and 116 to determine if additional modifications, such as the one suggested, are needed.

Comment: Two commenters supported the proposal to restructure DRG 514 through the creation of new DRGs 535 and 536. One of the commenters supported the division of these new DRGs based on the presence or absence of acute myocardial infarction, heart failure, or shock. However, the commenter believed that this new structure would lead to significant confusion among hospital coders with respect to the coding of CRT-Ds. The commenter stated that hospital coders may be confused when a patient is admitted with one diagnosis, but then develops an acute myocardial infarction, heart failure, or shock after the admission but prior to discharge. In these cases, the acute myocardial infarction, heart failure, or shock would be a secondary diagnosis. The split of DRGs 535 and 536 is based on these conditions when they are the principal diagnosis (reason for the hospital admission). To eliminate the potential for misunderstanding, the commenter requested that the definition of DRG 535 be modified so that patients who receive CRT-D devices are assigned to DRG 535 when an ICD-9-CM diagnosis code for heart failure is present as either a principal or secondary diagnosis.

Response: We appreciate the support from the commenters for our proposal to modify DRG 514 through the creation of new DRGs 535 and 536. We note that the issue of coding the implantation of CRT-Ds has been covered through extensive articles in the American Hospital Association's Coding Clinic for ICD-9-CM. In the past, the coding of cases with acute myocardial infarction, heart failure, or shock has not been problematic for hospital coding specialists. However, should the DRG modifications lead to coding questions on CRT-D cases, we will ask the American Hospital Association to provide additional guidance in its Coding Clinic for ICD-9-CM. Furthermore, the DRG splits for an acute myocardial infarction, heart failure, or shock, which currently are included in DRGs 115 and 116, are based on these conditions being the principal diagnosis. As a result, this is a longstanding DRG logic precedent. We do not believe that replicating the logic used for splitting DRGs 115 and 116 and using it for DRGs 535 and 536 would create confusion for hospital coders. Rather, we believe hospital coders would easily recognize this type of longstanding DRG logic.

Comment: Another commenter supported the proposal to split DRG 514 into DRGs 535 and 536 based on the presence or absence of acute myocardial infarction, heart failure, or shock. The commenter stated that this split would ensure greater consistency within the DRG system and ensure adequate payment to hospitals for the higher costs patients receiving implantable cardioverter-defibrillator implants. However, the commenter recommended that DRG 515 undergo a similar split based on the presence or absence of acute myocardial infarction, heart failure, or shock. The commenter stated that the creation of these additional new DRGs would fully align payment logic across all pacemaker and implantable cardioverter-defibrillator implant devices. The manufacturer also believed that differences between average charges and average length of stay for these cases within DRG 515 would warrant this additional splitting of the DRG.

Response: We appreciate the support for the revisions involving DRGs 514, 535, and 536. However, when we examined the data for DRGs 514 and 515, we found that there were almost three times as many cases with an acute myocardial infarction, heart failure, or shock cases in DRG 515 as in DRG 514. Those cases in DRG 514 with a principal diagnosis of an acute myocardial infarction, heart failure, or shock, had average charges approximately 20 percent greater than the average charges for all cases in DRG 514. However, cases with a principal diagnosis of an acute myocardial infarction, heart failure, or shock in DRG 515 had average charges that were only about 10 percent greater than all cases in this DRG. Therefore, there is a significantly greater need for the DRG split for DRG 514. We will continue to examine cases within this area, and specifically DRG 515, to determine if additional DRG refinements are needed in the future.

Comment: One commenter, who supported the revisions to DRG 514 through the new DRGs 535 and 536, expressed concern about our coverage decisions on automatic implantable cardioverter-defibrillators. The commenter believed the coverage was extremely restricted.

Response: We appreciate the support of the commenter for new DRGs 535 and 536. We will share the concerns relating to coverage decisions on automatic implantable cardioverter-defibrillators with our coverage staff.

5. MDC 8 (Diseases and Disorders of the Musculoskeletal System and Connective Tissue)

Prior to the issuance of the proposed rule, we received a comment that two codes for cervical fusion of the spine are not included within DRG 519 (Cervical Spinal Fusion With CC) and DRG 520 (Cervical Spinal Fusion Without CC). The two cervical fusion codes are:

• 81.01, Atlas-axis spinal fusion
• 81.31, Refusion of atlas-axis

The atlas-axis includes the first two vertebrae of the cervical spine (C1 and C2). These two cervical fusion codes are currently assigned to DRG 497 (Spinal Fusion Except Cervical With CC) and DRG 498 (Spinal Fusion Except Cervical Without CC). Because codes 81.01 and 81.31 involve the cervical spine, we proposed to remove these codes from DRGs 497 and 498 and reassign them to DRGs 519 and 520.

We did not receive any comments on this proposal. Therefore, we are adopting as final our proposal to remove codes 81.01 and 81.31 from DRGs 497 and 498 and reassign them to DRGs 519 and 520, effective for FY 2004.

6. MDC 15 (Newborns and Other Neonates With Conditions Originating in the Perinatal Period)

a. Nonneonate Diagnoses

As indicated earlier, ICD-9-CM diagnosis codes are assigned to MDCs based on 25 groupings corresponding to a single organ system or etiology and, in general, are associated with a particular medical specialty. MDC 15 is comprised of diagnoses that relate to newborns and other neonates with conditions originating in the perinatal period. Some of the codes included in MDC 15 consist of conditions that originate in the neonatal period but can persist throughout life. These conditions are referred to as congenital anomalies. When an older (not neonate) population is treated for a congenital anomaly, DRG assignment problems can arise. For instance, if a patient is over 65 years old and is admitted with a congenital anomaly, it is not appropriate to assign the patient to a newborn DRG. This situation occurs when a congenital anomaly code is classified within MDC 15.

Prior to the publication of the proposed rule, we received a recommendation to move the following congenital anomaly codes from MDC 15 and reassign them to other appropriate MDCs based on the body system being treated:

• 758.9, Chromosome anomaly, not otherwise specified
• 759.4, Conjoined twins
• 759.7, Multiple congenital anomalies, not elsewhere classified
• 759.81, Prader-Willi syndrome
• 759.83, Fragile X syndrome
• 759.89, Specified congenital anomalies, not elsewhere classified
• 759.9, Congenital anomaly, not otherwise specified
• 779.7, Periventricular leukomalacia
• 795.2, Abnormal chromosomal analysis

Each of the congenital anomaly diagnosis codes recommended for reassignment represents a condition that is frequently addressed beyond the neonatal period. In addition, the assignment of these congenital anomaly codes as principal diagnosis currently results in assignment to MDC 15.

We evaluated the recommendation and agreed that each of the identified codes represents a condition that is frequently addressed beyond the neonate period and should therefore be removed from the list of principal diagnoses that result in assignment to MDC 15. Therefore, we proposed to change the MDC and DRG assignments of the congenital anomaly codes as specified in the following table. The table shows the principal diagnosis code for the congenital anomaly and the proposed MDC and DRG to which the code would be assigned.

Comment: Several commenters supported all of the proposed changes relating to congenital anomalies. One commenter supported the changes in general, but mentioned several concerns. While this commenter agreed that it was feasible to move these congenital conditions out of MDC 15, the commenter suggested that those patients who are still in the neonatal period (first 28 days of life) when admitted should continue to be classified to MDC 15.

In addition, this commenter questioned whether the proposed DRG assignments were correct for codes 759.4 (Conjoined twins), code 759.7 (Multiple congenital anomalies, not elsewhere classified), and 759.89 (Specified congenital anomalies, not elsewhere classified). The commenter stated that although the proposed DRG assignments for these three DRGs may be appropriate based on the body system being treated for most cases, these DRGs do not necessarily reflect the body system affected or being treated. The commenter did not suggest alternative DRG assignments.

Response: We acknowledge the commenter's point that, for a minority of cases, the admission will, in fact, be in the neonatal period. However, the majority of cases will continue to be patients well beyond the neonatal period. The proposed DRG modifications will correct the majority of inappropriate DRG assignments that occur when adults are assigned to MDC 15 (Newborns and Other Neonates with Conditions Originating in the Perinatal Period). In the future, we will examine other means to further refine this area, such as making new DRG assignments for congenital anomalies based on the age of the patient. However, at this point, we are attempting to resolve the problems created for the majority of patients.

Regarding the commenter's concern that codes 759.4, 759.7, and 759.89 may not always be appropriately assigned according to our proposal, the commenter did not suggest an alternative. The commenter agreed that many cases with these three codes will be assigned to the appropriate body system by using our proposed DRG assignments. We recognize that reassignment of these codes will not resolve all problems, and some cases may be assigned to the wrong body system based on the patient's actual condition. However, we note that these three codes are vague and do not specify a precise congenital anomaly by body system. Therefore, we had to rely on our medical advisors to determine the most appropriate DRG for the majority of cases. Our main concern was to correct the DRG assignment that resulted in adults being assigned to a neonatal DRG when they had a congenital anomaly. We will continue to examine the data for these cases to determine if additional modifications are needed in the future.

Therefore, we are adopting the proposed revisions as final without modification.

b. Heart Failure Codes for Newborns and Neonates

Under MDC 15, cases of newborns and neonates with major problems may be assigned to DRG 387 (Prematurity With Major Problems) or DRG 389 (Full-Term Neonate With Major Problems). Existing DRG 387 has three components: (1) Principal or secondary diagnosis of prematurity; (2) principal or secondary diagnosis of major problem (these are the diagnoses that define MDC 15); or (3) secondary diagnosis of major problem (these are diagnoses that do not define MDC 15, so they will only be secondary diagnosis codes for patients assigned to MDC 15). To be assigned to DRG 389, the neonate must have one of the principal or secondary diagnoses listed under the DRG.

Prior to the publication of the proposed rule, we received correspondence suggesting that the following diagnosis codes for heart failure, which are currently in MDC 5, be added to the list of secondary diagnosis of major problems for neonates under MDC 15.

These heart failure-related diagnosis codes were new codes as of October 1, 2002. They were an expansion of the previous 4-digit codes for heart failure and provided additional detail about the specific type of heart failure. The codes for heart failure that existed prior to October 1, 2002, are classified as secondary diagnoses of major problems within MDC 15 and are currently assigned to DRGs 387 and DRG 389. We stated in the proposed rule that these other heart failure diagnosis codes should be included as principal diagnosis of major problem codes within MDC 15. However, these heart failure codes are currently listed in the secondary, not principal, diagnoses of major problems within MDC 15.

We agree that diagnosis codes 428.20 through 428.43 listed in the chart above should be included as secondary diagnosis of major problem codes within MDC 15, as are the other heart failure codes. Therefore, we proposed to add them to DRG 387 and 389.

Comment: Several commenters supported the proposal to add codes 428.20 through 428.43 (codes for heart failure that became effective October 1, 2002, listed in the chart above) to DRGs 387 and 389. The commenters agreed that the heart failure codes created on October 1, 2002, should be assigned to DRGs 387 and 389 in the same fashion as were those heart failure codes created prior to October 1, 2002.

One commenter indicated that we incorrectly described the addition of diagnosis codes 428.20 through 428.43 listed in the chart to the list of “principal” diagnosis of major problem codes. The commenter stated that we should have indicated that these codes would be added to the list of “secondary” diagnoses of major problem codes because this category is where the other heart failure codes are currently assigned.

Response: We agree that the codes should have been described as an addition to the list of secondary diagnoses of major problem codes within DRGs 387 and 389. We have clarified this point in the description above.

Comment: One commenter who supported the addition of the heart failure-related diagnosis codes (428.20 through 428.43) to DRGs 387 and 389, asked for clarification of how diagnoses for combined codes that include congestive heart failure will be handled. The commenter mentioned code 402.91 (Hypertensive heart disease with heart failure, unspecified benign or malignant) as an example.

Response: We will conduct an additional review of DRGs 387 and 389 to determine if additional codes should be added to the list of secondary diagnoses of major problems for FY 2005. We encourage commenters to send their recommendations to us to assist in this review.

We are adopting our proposal as final, with the clarification that the major problem codes are secondary, not principal, codes. Accordingly, we are adding codes 428.20 through 428.43 listed above to the list of secondary diagnoses of major problem codes within DRGs 387 and 389.

7. MDC 17 (Myeloproliferative Diseases and Disorders and Poorly Differentiated Neoplasms)

High-dose Interleukin-2 (IL-2) Chemotherapy is a hospital inpatient-based regimen requiring administration by experienced oncology professionals. It is used for the treatment of patients with advanced renal cell cancer and advanced melanoma. Unlike traditional cytotoxic chemotherapies that attack cancer cells themselves, Interleukin-2 is designed to enhance the body's defenses by mimicking the way natural IL-2 activates the immune system and stimulates the growth and activity of cancer-killing cells. The Food and Drug Administration (FDA) approved the IL-2 product on the market for use in 1992.

High-dose IL-2 therapy is performed only in very specialized treatment settings, such as an intensive care unit or a bone marrow transplant unit. This therapy requires oversight by oncology health care professionals experienced in the administration and management of patients undergoing this intensive treatment because of the severity of the side effects. Unlike most cancer therapies, high-dose IL-2 therapy is associated with predictable toxicities that require extensive monitoring. Often patients require one-on-one nursing or physician care for extended portions of their stay.

High-dose IL-2 therapy is significantly different from conventional chemotherapy in terms of the resources required to administer it. Conventional chemotherapy may be given to patients either on an outpatient basis or through a series of short (that is, 1 to 3 day) inpatient stays.

High-dose IL-2 therapy is given during two separate hospital admissions. For the first cycle, the IL-2 is administered every 8 hours over 5 days. Patients are then discharged to rest at home for several days and are admitted for the second cycle of therapy during which the same regimen and dosing is repeated. The two cycles complete the first course of high-dose IL-2 therapy. This regimen may be repeated at 8 to 12 weeks if the patient is responding. The maximum number of courses for any one patient is predicted to be five courses.

Not all patients with end-stage renal cell carcinoma or end-stage melanoma are appropriate candidates for high-dose IL-2 chemotherapy. It is estimated that there are between 15,000 and 20,000 patients in the United States who have one of these two types of cancer. However, only 20 percent of those patients will be appropriate candidates for the rigors of the treatment regimen. It is further estimated that, annually, approximately 1,300 of these patients will be Medicare beneficiaries. However, we have been informed by industry sources that, allegedly due to the level of payment for the DRGs to which these cases are currently assigned, only 100 to 200 Medicare patients receive the treatment each year. According to these industry sources, several treatment centers have had to discontinue their high-dose IL-2 therapy programs for end-stage renal cell carcinoma or end-stage melanoma because of the low Medicare payment.

According to industry sources, the wholesale cost of IL-2 is approximately $700 per vial. Dosages range between 15 and 20 vials per treatment, or between $10,500 and $14,000 per patient, per cycle, for the cost of the IL-2 drug alone. There is no ICD-9-CM procedure code that currently identifies patients receiving this therapy. Therefore, it is not possible to identify directly these cases in the MedPAR data. Currently, this therapy is coded using the more general ICD-9-CM code 99.28 (Injection or infusion of biologic response modifier). When we addressed this issue previously in the August 1, 2000 IPPS final rule (65 FR 47067) by examining cases for which procedure code 99.28 was present, our analysis was inconclusive due to the wide range of cases identified (1,179 cases across in 136 DRGs). However, recent data collected by the industry on 30 Medicare beneficiaries who received high-dose IL-2 therapy during FY 2002 show average charges for these cases of approximately $54,000.

Depending on the principal diagnosis reported, patients receiving high-dose IL-2 therapy may be assigned to one of the following five DRGs: DRG 272 (Major Skin Disorder With CC) and DRG 273 (Major Skin Disorder Without CC) in MDC 9; DRG 318 (Kidney and Urinary Tract Neoplasms With CC) and DRG 319 (Kidney and Urinary Tract Neoplasms Without CC) in MDC 11; and DRG 410 (Chemotherapy Without Leukemia as Secondary Diagnosis) in MDC 17. The following table illustrates the average charges for patients in these DRGs.

Because of the need to identify the subset of patients receiving this type of treatment, the ICD-9-CM Coordination and Maintenance Committee determined, based on its consideration at the December 6, 2002 public meeting, that a new code for high-dose IL-2 therapy was warranted. Therefore, a new code has been created in the 00 Chapter of ICD-9-CM (Procedures and Interventions, Not Elsewhere Classified), in category 00.1 (Pharmaceuticals) at 00.15 (High-dose infusion Interleukin-2 (IL-2)). The code is effective for cases discharged on or after October 1, 2003.

We believe patients receiving high-dose IL-2 therapy are clinically similar to other cases currently assigned to DRG 492 (Chemotherapy With Acute Leukemia as Secondary Diagnosis) in MDC 17. The average charge for patients currently assigned to DRG 492 is $55,581. Currently, DRG 492 requires one of the following two principal diagnoses:

• V58.1, Encounter for chemotherapy
• V67.2, Followup examination following chemotherapy

And one of the following secondary diagnoses:

• 204.00, Acute lymphoid leukemia without mention of remission
• 204.01, Acute lymphoid leukemia with remission
• 205.00, Acute myeloid leukemia without mention of remission
• 205.01, Acute myeloid leukemia with remission
• 206.00, Acute monocytic leukemia without mention of remission
• 206.01, Acute monocytic leukemia with remission
• 207.00, Acute erythremia and erythroleukemia without mention of remission
• 207.01, Acute erythremia and erythroleukemia with remission
• 208.00, Acute leukemia of unspecified cell type without mention of remission
• 208.01, Acute leukemia of unspecified cell type without mention of remission

We proposed to modify DRG 492 by adding new procedure code 00.15 to the logic. We indicated that assignment to this DRG would require the same two V-code principal diagnosis codes listed above (V58.1 and V67.2), but would require either one of the leukemia codes listed as a secondary diagnosis, or would require the procedure code 00.15. In addition, we proposed to change the title of DRG 492 to “Chemotherapy With Acute Leukemia or With Use of High Dose Chemotherapy Agent”.

In the proposed rule, we indicated that we would monitor cases with procedure code 00.15 as these data became available, and consider potential further refinements to DRG 492 as necessary.

Comment: Five commenters supported our proposed change. One commenter who opposed the proposed change believed that classifying high-dose IL-2 therapy as chemotherapy would be a violation of coding advice published in the American Hospital Association's coding publication, Coding Clinic for ICD-9-CM, because IL-2 therapy is a biologic response modifier and is considered immunotherapy, not chemotherapy. Therefore, the commenter asserted that the use of either V58.1 or V67.2 as principal diagnosis codes for these cases would result in erroneous coding advice. The commenter added that Coding Clinic, Fourth Quarter, page 51, indicates that when a patient is admitted for immunotherapy, the code for the neoplasm should be assigned as the principal diagnosis.

Response: We acknowledge the commenter's points concerning correct selection of principal diagnosis, as well as the advice published previously in Coding Clinic. However, the discussion of this topic has raised some concerns among the Cooperating Parties of AHA's Editorial Advisory Board. The advice given in the Fourth Quarter 1994 Coding Clinic predates the new treatment technology now available, which calls into question the correctness of the published advice. Therefore, this topic will be included on the agenda of an upcoming AHA Editorial Advisory Board meeting for further discussion and clarification. It is likely that new instructions will be issued in the next several months to clarify these coding instructions.

Therefore, in anticipation of this clarification, we are adopting as final the proposed changes to DRG 492. We will continue to monitor this DRG for shifts in resource consumption and validity of DRG assignment, and will specifically monitor code 00.15 for appropriate placement in DRG 492.

8. MDC 23 (Factors Influencing Health Status and Other Contacts With Health Services)

a. Implantable Devices

Prior to the publication of the proposed rule, we received a comment regarding three ICD-9-CM diagnosis codes that are currently assigned to MDC 23: V53.01 (Fitting and adjustment of cerebral ventricular (communicating) shunt); V53.02 (Neuropacemaker (brain) (peripheral nerve) (spinal cord)); and V53.09 (Fitting and adjustment of other devices related to nervous system and special senses). The commenter suggested that we move these three codes from MDC 23 to MDC 1 (Diseases and Disorders of the Nervous System) because these codes are used as the principal diagnosis for admissions involving removal, replacement, and reprogramming of devices such as cerebral ventricular shunts, neurostimulators, intrathecal infusion pumps and thalamic stimulators.

Currently, if these diagnosis codes are reported alone without an O.R. procedure, the case would be assigned to DRG 467 (Other Factors Influencing Health Status). However, if an O.R. procedure is reported with the principal diagnosis of V53.01, V53.02, or V53.09, the case would be assigned to DRG 461 (O.R. Procedure with Diagnoses of Other Contact with Health Services).

In our analysis of the MedPAR data, we found 30 cases assigned to DRG 467 and 179 cases assigned to DRG 461 with one of these codes as principal diagnosis. We found that the procedures reported with one of these diagnosis codes were procedures in MDC 1. The most frequent procedure was 86.06 (Insertion of totally implantable infusion pump).

Because the procedures that are routinely used with these codes are in MDC 1, we believe it would be appropriate to assign these diagnosis codes to MDC 1. As the commenter also stated, this assignment would be consistent with how fitting and adjustments of devices are handled within other MDCs, such as in MDC 5 (Diseases and Disorders of the Circulatory System) and MDC 11 (Diseases and Disorders of the Kidney and Urinary Tract). Diagnosis codes V53.31 (Cardiac pacemaker), V53.32 (Automatic implantable cardiac defibrillator), and V53.39 (Other cardiac device) are used for fitting and adjustment of cardiac devices and are assigned to MDC 5. Diagnosis code V53.6 (Urinary devices) is used for fitting and adjustment of urinary devices and is assigned to MDC 11.

Therefore, we proposed to move V53.01, V53.02, and V53.09 from MDC 23 to MDC 1 when an O.R. procedure is performed. If no O.R. procedure is performed, these diagnosis codes would be assigned to DRG 34 (Other Disorders of Nervous System With CC) or DRG 35 (Other Disorders of Nervous System Without CC). If an O.R. procedure is performed on a patient assigned with one of these codes as the principal diagnosis, the case would be assigned to the DRG in MDC 1 to which the O.R. procedure is assigned.

We received three comments that supported our proposal to move diagnosis codes V53.01, V53.02, and V53.09 from MDC 23 to MDC 1. Accordingly, we are adopting as final the proposed reassignment, effective for discharges occurring on or after October 1, 2003.

b. Malignancy Codes

Prior to the issuance of the proposed rule, we received correspondence that indicated that when we recognized code V10.48 (History of malignancy, epididymis) as a new code for FY 2002, we did not include the code as a history of malignancy code in DRG 465 (Aftercare with History of Malignancy as Secondary Diagnosis). All other history of malignancy codes were included in DRG 465.

We agree that code V10.48 should have been included in the list of history of malignancy codes within DRG 465. Therefore, we proposed to add it to the list of secondary diagnoses in DRG 465.

We received several comments that supported this DRG modification. Accordingly, we are adopting the proposal as final without modification.

9. Medicare Code Editor (MCE) Change

As explained under section II.B.1. of this preamble, the MCE is a software program that detects and reports errors in the coding of Medicare claims data.

We received a request to examine the MCE edit “Adult Diagnosis—Age Greater than 14” because currently the edit rejects claims for patients under age 15 who are being treated for gall bladder disease. We reviewed this issue with our pediatric consultants and determined that, although incidence is rare, gallbladder disease does occur in patients under age 15. Therefore, in the May 19, 2003 proposed rule, we proposed to modify the MCE by removing the following codes from the edit “Adult Diagnosis—Age Greater Than 14”:

• 574.00, Calculus of gallbladder with acute cholecystitis without mention of obstruction
• 574.01, Calculus of gallbladder with acute cholecystitis with obstruction
• 574.10, Calculus of gallbladder with other cholecystitis without mention of obstruction
• 574.11, Calculus of gallbladder with other cholecystitis with obstruction
• 574.20, Calculus of gallbladder without mention of cholecystitis without mention of obstruction
• 574.21, Calculus of gallbladder without mention of cholecystitis with obstruction
• 574.30, Calculus of bile duct with acute cholecystitis without mention of obstruction
• 574.31, Calculus of bile duct with acute cholecystitis with obstruction
• 574.40, Calculus of bile duct with other cholecystitis without mention of obstruction
• 574.41, Calculus of bile duct with other cholecystitis with obstruction
• 574.50, Calculus of bile duct without mention of cholecystitis without mention of obstruction
• 574.51, Calculus of bile duct without mention of cholecystitis with obstruction
• 574.60, Calculus of gallbladder and bile duct with acute cholecystitis without mention of obstruction
• 574.61, Calculus of gallbladder and bile duct with acute cholecystitis with obstruction)
• 574.70, Calculus of gallbladder and bile duct with other cholecystitis without mention of obstruction
• 574.71, Calculus of gallbladder and bile duct with other cholecystitis with obstruction
• 574.80, Calculus of gallbladder and bile duct with acute and chronic cholecystitis without mention of obstruction
• 574.81, Calculus of gallbladder and bile duct with acute and chronic cholecystitis with obstruction
• 574.90, Calculus of gallbladder and bile duct without cholecystitis without mention of obstruction
• 574.91, Calculus of gallbladder and bile duct without cholecystitis with obstruction
• 575.0, Acute cholecystitis
• 575.10, Cholecystitis, not otherwise specified
• 575.11, Chronic cholecystitis
• 575.12, Acute and chronic cholecystitis
• 575.2, Obstruction of gallbladder
• 575.3, Hydrops of gallbladder
• 576.0, Postcholecystectomy syndrome
• 577.1, Chronic pancreatitis

Comment: Four commenters agreed in general with our decision to remove the above listed codes from the MCE in the edit “Adult Diagnosis—Age Greater than 14.” However, one commenter recommended that all ICD-9-CM codes in the 575 through 577 range be removed from the edit and listed several codes that appeared to be missing from our list. These codes were 575.4 (Perforation of gallbladder), 577.0 (Acute pancreatitis), and 577.1 (Chronic pancreatitis). In addition, three commenters pointed out that code 574.90 had been erroneously listed twice with different narrative descriptions.

Response: We appreciate the commenters' interest in the correctness of the MCE. We also have received many telephone calls and e-mails concerning the typographical error with code 574.90. We have corrected the list above to reflect the correct code number, 574.91. As noted, the second narrative listing in the proposed rule correctly described code 574.91, not 574.90 (68 FR 27166).

With regard to the comment concerning the absence of codes 575.4 and 577.0 from the above list, we note that these codes are not included in the MCE edit. That is, these codes were never part of the MCE edit. With regard to code 577.1, this code is the last one on the list and was printed correctly in the proposed rule (68 FR 27166, third column).

Accordingly, we are adopting as final the proposal to remove the listed codes from the MCE edit “Adult Diagnosis—Age Greater than 14,” with the correction of the fifth digit of code 574.91 (Calculus of gallbladder and bile duct without cholecystitis with obstruction).

10. Surgical Hierarchies

Some inpatient stays entail multiple surgical procedures, each one of which, occurring by itself, could result in assignment of the case to a different DRG within the MDC to which the principal diagnosis is assigned. Therefore, it is necessary to have a decision rule within the GROUPER by which these cases are assigned to a single DRG. The surgical hierarchy, an ordering of surgical classes from most resource-intensive to least resource-intensive, performs that function. Application of this hierarchy ensures that cases involving multiple surgical procedures are assigned to the DRG associated with the most resource-intensive surgical class.

Because the relative resource intensity of surgical classes can shift as a function of DRG reclassification and recalibrations, we reviewed the surgical hierarchy of each MDC, as we have for previous reclassifications and recalibrations, to determine if the ordering of classes coincides with the intensity of resource utilization.

A surgical class can be composed of one or more DRGs. For example, in MDC 11, the surgical class “kidney transplant” consists of a single DRG (DRG 302) and the class “kidney, ureter and major bladder procedures” consists of three DRGs (DRGs 303, 304, and 305). Consequently, in many cases, the surgical hierarchy has an impact on more than one DRG. The methodology for determining the most resource-intensive surgical class involves weighting the average resources for each DRG by frequency to determine the weighted average resources for each surgical class. For example, assume surgical class A includes DRGs 1 and 2 and surgical class B includes DRGs 3, 4, and 5. Assume also that the average charge of DRG 1 is higher than that of DRG 3, but the average charges of DRGs 4 and 5 are higher than the average charge of DRG 2. To determine whether surgical class A should be higher or lower than surgical class B in the surgical hierarchy, we would weight the average charge of each DRG in the class by frequency (that is, by the number of cases in the DRG) to determine average resource consumption for the surgical class. The surgical classes would then be ordered from the class with the highest average resource utilization to that with the lowest, with the exception of “other O.R. procedures” as discussed below.

This methodology may occasionally result in assignment of a case involving multiple procedures to the lower-weighted DRG (in the highest, most resource-intensive surgical class) of the available alternatives. However, given that the logic underlying the surgical hierarchy provides that the GROUPER search for the procedure in the most resource-intensive surgical class, this result is unavoidable.

We note that, notwithstanding the foregoing discussion, there are a few instances when a surgical class with a lower average charge is ordered above a surgical class with a higher average charge. For example, the “other O.R. procedures” surgical class is uniformly ordered last in the surgical hierarchy of each MDC in which it occurs, regardless of the fact that the average charge for the DRG or DRGs in that surgical class may be higher than that for other surgical classes in the MDC. The “other O.R. procedures” class is a group of procedures that are only infrequently related to the diagnoses in the MDC but are still occasionally performed on patients in the MDC with these diagnoses. Therefore, assignment to these surgical classes should only occur if no other surgical class more closely related to the diagnoses in the MDC is appropriate.

A second example occurs when the difference between the average charges for two surgical classes is very small. We have found that small differences generally do not warrant reordering of the hierarchy because, as a result of reassigning cases on the basis of the hierarchy change, the average charges are likely to shift such that the higher-ordered surgical class has a lower average charge than the class ordered below it.

Based on the preliminary recalibration of the DRGs, in the May 19, 2003 proposed rule, we proposed modifications of the surgical hierarchy as set forth below.

We proposed to revise the surgical hierarchy for the pre-MDC DRGs, MDC 1 (Diseases and Disorders of the Nervous System), MDC 5 (Diseases and Disorders of the Circulatory System), MDC 8 (Diseases and Disorders of the Musculoskeletal System and Connective Tissue), and MDC 17 (Myeloproliferative Disease and Disorders, Poorly Differentiated Neoplasms for Lymphoma and Leukemia) as follows:

• In the pre-MDC DRGs, we proposed to reorder DRG 513 (Pancreas Transplant) above DRG 512 (Simultaneous Pancreas/Kidney Transplant).
• In MDC 1, we proposed to reorder DRG 3 (Craniotomy Age 0-17) above DRG 528 (Intracranial Vascular Procedures with Principal Diagnosis Hemorrhage); DRG 528 above DRGs 1 and 2 (Craniotomy Age >17 With and Without CC, respectively); DRGs 1 and 2 above DRGs 529 and 530 (Ventricular Shunt Procedures With and Without CC, respectively); DRGs 529 and 530 above DRGs 531 and 532 (Spinal Procedures With and Without CC, respectively); DRGs 531 and 532 above DRGs 533 and 534 (Extracranial Procedures With and Without CC, respectively); and DRGs 533 and 534 above DRG 6 (Carpal Tunnel Release).
• In MDC 5, we proposed to reorder DRG 535 (Cardiac Defibrillator Implant With Cardiac Catheterization With AMI, Heart Failure, or Shock) above DRG 536 (Cardiac Defibrillator Implant With Cardiac Catheterization Without AMI, Heart Failure, or Shock), and DRG 536 above DRG 515 (Cardiac Defibrillator Implant Without Cardiac Catheterization).
• In MDC 8, we proposed to reorder DRGs 537 and 538 (Local Excision and Removal of Internal Fixation Devices Except Hip and Femur With and Without CC, respectively) above DRG 230 (Local Excision and Removal of Internal Fixation Devices of Hip and Femur).
• In MDC 17, we proposed to reorder DRGs 539 and 540 (Lymphoma and Leukemia With Major O.R. Procedure With and Without CC, respectively) above DRGs 401 and 402 (Lymphoma and Non-Acute Leukemia With Other O.R. Procedures With and Without CC, respectively).

In the proposed rule, we were unable to test the effects of the proposed revisions to the surgical hierarchy and reflect these changes in the proposed relative weights because the revised GROUPER software was unavailable at the time the proposed rule was published. Rather, we simulated most major classification changes to approximate the placement of cases under the proposed reclassification, and then determined the average charge for each DRG. These average charges served as our best estimate of relative resources used for each surgical class. We have now tested the proposed surgical hierarchy changes using the revised GROUPER software, and are reflecting the final changes in the DRG relative weights in this final rule. Further, as discussed in section II.C. of the preamble of this final rule, the final recalibrated weights are different from the proposed weights because they were based on more complete data.

Based on a test of the proposed revisions using the March 2003 update of the FY 2002 MedPAR file and the revised GROUPER software, we have found that the proposed change in the pre-MDC DRGs to reorder DRG 513 (Pancreas Transplant) above DRG 12 (Simultaneous Pancreas/Kidney Transplant) was not supported by the data. If this proposal were finalized, no cases would be assigned to DRG 512. The other proposed revisions are still supported by the data.

Comment: Two commenters expressed support for the proposed change in the surgical hierarchy. Another commenter requested a change in the surgical hierarchy for a case in which a spinal fusion with subsequent debridement is performed during the same admission. This case is assigned to DRG 217 (Wound Debridement and Skin Graft Except Hand, for Musculoskeletal and Connective Tissue Disease). The commenter requested that this case be reassigned to DRG 497 (Spinal Fusion Except Cervical With CC) because it has a higher DRG weight than DRG 217.

Response: The surgical hierarchy places a patient with multiple procedures in the most resource intensive class, but this does not necessarily mean that the patient is assigned to the most resource intensive DRG. In this scenario, one surgical class is actually one DRG, and another surgical class is back and neck procedures. These classes encompass 7 DRGs (DRGs 496-500 and DRGs 519 and 520). The average charges for DRG 217 are approximately $15,000 more than the back and neck procedures class. DRG 217 is hierarchically ordered higher in the surgical group than DRG 497, which is the reason the case is assigned to DRG 217.

Therefore, we are adopting the proposed changes in MDCs 1, 5, 8, and 17 as final. We are not making any changes in the pre-MDC DRGs.

11. Refinement of Complications and Comorbidities (CC) List

In the September 1, 1987 final notice (52 FR 33143) concerning changes to the DRG classification system, we modified the GROUPER logic so that certain diagnoses included on the standard list of CCs would not be considered valid CCs in combination with a particular principal diagnosis. We created the CC Exclusions List for the following reasons: (1) To preclude coding of CCs for closely related conditions; (2) to preclude duplicative or inconsistent coding from being treated as CCs; and (3) to ensure that cases are appropriately classified between the complicated and uncomplicated DRGs in a pair. We developed this list of diagnoses, using physician panels, to include those diagnoses that, when present as a secondary condition, would be considered a substantial complication or comorbidity. In previous years, we have made changes to the list of CCs, either by adding new CCs or deleting CCs already on the list. As we proposed in the May 19, 2003 proposed rule, we are not deleting any of the diagnosis codes on the CC list.

As explained in the May 19, 1989 proposed rule (52 FR 18877) and the September 1, 1987 final notice (52 FR 33154), the excluded secondary diagnoses were established using the following five principles:

• Chronic and acute manifestations of the same condition should not be considered CCs for one another.
• Specific and nonspecific (that is, not otherwise specified (NOS)) diagnosis codes for the same condition should not be considered CCs for one another.
• Codes for the same condition that cannot coexist, such as partial/total, unilateral/bilateral, obstructed/unobstructed, and benign/malignant, should not be considered CCs for one another.
• Codes for the same condition in anatomically proximal sites should not be considered CCs for one another.
• Closely related conditions should not be considered CCs for one another.

The creation of the CC Exclusions List was a major project involving hundreds of codes. We have continued to review the remaining CCs to identify additional exclusions and to remove diagnoses from the master list that have been shown not to meet the definition of a CC.

We proposed a limited revision of the CC Exclusions List to take into account the proposed changes that will be made in the ICD-9-CM diagnosis coding system effective October 1, 2003. (See section II.B.13. of this preamble for a discussion of ICD-9-CM changes.) We proposed these changes in accordance with the principles established when we created the CC Exclusions List in 1987.

Tables 6G and 6H in the Addendum to this final rule contain the revisions to the 13 CC Exclusions List that will be effective for discharges occurring on or after October 1, 2003. Each table shows the principal diagnoses with changes to the excluded CCs. Each of these principal diagnoses is shown with an asterisk, and the additions or deletions to the CC Exclusions List are provided in an indented column immediately following the affected principal diagnosis.

CCs that are added to the list are in Table 6G—Additions to the CC Exclusions List. Beginning with discharges on or after October 1, 2003, the indented diagnoses will not be recognized by the GROUPER as valid CCs for the asterisked principal diagnosis.

CCs that are deleted from the list are in Table 6H—Deletions from the CC Exclusions List. Beginning with discharges on or after October 1, 2003, the indented diagnoses will be recognized by the GROUPER as valid CCs for the asterisked principal diagnosis.

Comment: One commenter indicated that it was unable to provide meaningful comments on Tables 6G and 6H because of formatting errors in the printed tables. In addition, the commenter suggested that the changes in the tables should not be effective until a revised version was made available for public comment.

Response: We apologize for the errors in the format of the tables, which were printer's errors. However, we note that the tables did contain the correct codes, even though the format of the columns was distorted. Therefore, we do not believe a delay in the effective date of the changes is warranted.

Copies of the original CC Exclusions List applicable to FY 1988 can be obtained from the National Technical Information Service (NTIS) of the Department of Commerce. It is available in hard copy for $133.00 plus shipping and handling. A request for the FY 1988 CC Exclusions List (which should include the identification accession number (PB) 88-133970) should be made to the following address: National Technical Information Service, United States Department of Commerce, 5285 Port Royal Road, Springfield, VA 22161; or by calling (800) 553-6847.

Users should be aware of the fact that all revisions to the CC Exclusions List (FYs 1989, 1990, 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000, 2002, and 2003) and those in Tables 6G and 6H of this final rule for FY 2004 must be incorporated into the list purchased from NTIS in order to obtain the CC Exclusions List applicable for discharges occurring on or after October 1, 2003. (Note: There was no CC Exclusions List in FY 2001 because we did not make changes to the ICD-9-CM codes for FY 2001.)

Alternatively, the complete documentation of the GROUPER logic, including the current CC Exclusions List, is available from 3M/Health Information Systems (HIS), which, under contract with CMS, is responsible for updating and maintaining the GROUPER program. The current DRG Definitions Manual, Version 20.0, is available for $225.00, which includes $15.00 for shipping and handling. Version 21.0 of this manual, which includes the final FY 2004 DRG changes, is available for $225.00. These manuals may be obtained by writing 3M/HIS at the following address: 100 Barnes Road, Wallingford, CT 06492; or by calling (203) 949-0303. Please specify the revision or revisions requested.

12. Review of Procedure Codes in DRGs 468, 476, and 477

Each year, we review cases assigned to DRG 468 (Extensive O.R. Procedure Unrelated to Principal Diagnosis), DRG 476 (Prostatic O.R. Procedure Unrelated to Principal Diagnosis), and DRG 477 (Nonextensive O.R. Procedure Unrelated to Principal Diagnosis) to determine whether it would be appropriate to change the procedures assigned among these DRGs.

DRGs 468, 476, and 477 are reserved for those cases in which none of the O.R. procedures performed are related to the principal diagnosis. These DRGs are intended to capture atypical cases, that is, those cases not occurring with sufficient frequency to represent a distinct, recognizable clinical group. DRG 476 is assigned to those discharges in which one or more of the following prostatic procedures are performed and are unrelated to the principal diagnosis:

• 60.0 Incision of prostate
• 60.12 Open biopsy of prostate
• 60.15 Biopsy of periprostatic tissue
• 60.18 Other diagnostic procedures on prostate and periprostatic tissue
• 60.21 Transurethral prostatectomy
• 60.29 Other transurethral prostatectomy
• 60.61 Local excision of lesion of prostate
• 60.69 Prostatectomy, not elsewhere classified
• 60.81 Incision of periprostatic tissue
• 60.82 Excision of periprostatic tissue
• 60.93 Repair of prostate
• 60.94 Control of (postoperative) hemorrhage of prostate
• 60.95 Transurethral balloon dilation of the prostatic urethra
• 60.99 Other operations on prostate

All remaining O.R. procedures are assigned to DRGs 468 and 477, with DRG 477 assigned to those discharges in which the only procedures performed are nonextensive procedures that are unrelated to the principal diagnosis. The original list of the ICD-9-CM procedure codes for the procedures we consider nonextensive procedures, if performed with an unrelated principal diagnosis, was published in Table 6C in section IV. of the Addendum to the September 30, 1988 final rule (53 FR 38591). As part of the final rules published on September 4, 1990 (55 FR 36135), August 30, 1991 (56 FR 43212), September 1, 1992 (57 FR 23625), September 1, 1993 (58 FR 46279), September 1, 1994 (59 FR 45336), September 1, 1995 (60 FR 45783), August 30, 1996 (61 FR 46173), and August 29, 1997 (62 FR 45981), we moved several other procedures From DRG 468 to DRG 477, and some procedures from DRG 477 to DRG 468. No procedures were moved in FY 1999, as noted in the July 31, 1998 final rule (63 FR 40962); in FY 2000, as noted in the July 30, 1999 final rule (64 FR 41496); in FY 2001, as noted in the August 1, 2000 final rule (65 FR 47064); or in FY 2002, as noted in the August 1, 2001 final rule (66 FR 39852). In the August 1, 2002 final rule (67 FR 49999), we did not move any procedures from DRG 477. However, we did move procedures codes from DRG 468 and placed them in more clinically coherent DRGs.

a. Moving Procedure Codes From DRG 468 or DRG 477 to MDCs

We annually conduct a review of procedures producing assignment to DRG 468 or DRG 477 on the basis of volume, by procedure, to see if it would be appropriate to move procedure codes out of these DRGs into one of the surgical DRGs for the MDC into which the principal diagnosis falls. The data are arrayed two ways for comparison purposes. We look at a frequency count of each major operative procedure code. We also compare procedures across MDCs by volume of procedure codes within each MDC.

We identify those procedures occurring in conjunction with certain principal diagnoses with sufficient frequency to justify adding them to one of the surgical DRGs for the MDC in which the diagnosis falls. Based on this year's review, we did not identify any necessary changes in procedures under DRG 477. Therefore, we did not propose moving any procedures from DRG 477 to one of the surgical DRGs in this final rule.

However, in the proposed rule, we identified a necessary proposed change under DRG 468 relating to code 50.29 (Other destruction of lesion of liver). We were contacted by a hospital about the fact that code 50.29 is not currently included in MDC 6 (Diseases and Disorders of the Digestive System). The hospital pointed out that it is not uncommon for patients to have procedures performed on the liver when they are admitted for a condition that is classified in MDC 6. For example, DRGs 170 and 171 (Other Digestive System O.R. Procedures With and Without CC, respectively) in MDC 6 currently include liver procedures such as biopsy of the liver. The hospital disagreed with the assignment of code 50.29 to DRG 468 when performed on a patient with a principal diagnosis in MDC 6. We believe that the commenter is correct. Therefore, we proposed to assign code 50.29 to DRGs 170 and 171 in MDC 6.

We received several comments of support for our proposal to assign code 50.29 to DRGs 170 and 171 in MDC 6. Therefore, we are adopting the proposal as final without modification. As a result, code 50.29 will not result in assignment to DRG 468 when this procedure is performed on patient with a principal diagnosis in MDC 6.

b. Reassignment of Procedures Among DRGs 468, 476, and 477

We also annually review the list of ICD-9-CM procedures that, when in combination with their principal diagnosis code, result in assignment to DRGs 468, 476, and 477, to ascertain if any of those procedures should be reassigned from one of these three DRGs to another of the three DRGs based on average charges and the length of stay. We look at the data for trends such as shifts in treatment practice or reporting practice that would make the resulting DRG assignment illogical. If we find these shifts, we would propose to move cases to keep the DRGs clinically similar or to provide payment for the cases in a similar manner. Generally, we move only those procedures for which we have an adequate number of discharges to analyze the data. Based on our review this year, we did not propose moving any procedures from DRG 476 to DRGs 468 or 477, or from DRG 477 to DRGs 468 or 476.

However, in the proposed rule, we identified several procedures that we proposed to move from DRG 468 and add to DRGs 476 and 477 because the procedures are nonextensive:

• 38.21, Biopsy of blood vessel
• 77.42, Biopsy of scapula, clavicle and thorax [ribs and sternum]
• 77.43, Biopsy of radius and ulna
• 77.44, Biopsy of carpals and metacarpals
• 77.45, Biopsy of femur
• 77.46, Biopsy of patella
• 77.47, Biopsy of tibia and fibula
• 77.48, Biopsy of tarsals and metatarsals
• 77.49, Biopsy of other bones
• 92.27, Implantation or insertion of radioactive elements

We note that the above codes being moved from DRG 468 to DRGs 476 and 477 were erroneously listed in the May 19, 2003 proposed rule under section II.B.12.c., which related to adding diagnosis or procedure codes to MDCs, instead of section II.B.12.b., which discussed the reassignment of procedures among DRGs 468, 476, and 477. We regret any inconvenience this inadvertent listing may have caused.

Comment: One commenter asked us to consider moving procedure code 51.23, Laparoscopic cholecystectomy, from DRG 468 and adding it to DRG 477. The commenter indicated that this procedure is often performed in the outpatient setting.

Response: We believe that the commenter's request has merit. We will perform the necessary data analysis and will consider proposing this change in next fiscal year's rule if we find that the data support this change.

c. Adding Diagnosis or Procedure Codes to MDCs

Based on our review this year, we did not propose adding any diagnosis codes to MDCs in this final rule. We did not receive any comments on the proposal.

13. Changes to the ICD-9-CM Coding System

As described in section II.B.1. of this preamble, the ICD-9-CM is a coding system that is used for the reporting of diagnoses and procedures performed on a patient. In September 1985, the ICD-9-CM Coordination and Maintenance Committee was formed. This is a Federal interdepartmental committee, co-chaired by the National Center for Health Statistics (NCHS) and CMS, charged with maintaining and updating the ICD-9-CM system. The Committee is jointly responsible for approving coding changes, and developing errata, addenda, and other modifications to the ICD-9-CM to reflect newly developed procedures and technologies and newly identified diseases. The Committee is also responsible for promoting the use of Federal and non-Federal educational programs and other communication techniques with a view toward standardizing coding applications and upgrading the quality of the classification system.

The ICD-9-CM Manual contains the list of valid diagnosis and procedure codes. (The ICD-9-CM Manual is available from the Government Printing Office on CD-ROM for $23.00 by calling (202) 512-1800.) The NCHS has lead responsibility for the ICD-9-CM diagnosis codes included in the Tabular List and Alphabetic Index for Diseases, while CMS has lead responsibility for the ICD-9-CM procedure codes included in the Tabular List and Alphabetic Index for Procedures.

The Committee encourages participation in the above process by health-related organizations. In this regard, the Committee holds public meetings for discussion of educational issues and proposed coding changes. These meetings provide an opportunity for representatives of recognized organizations in the coding field, such as the American Health Information Management Association (AHIMA), the American Hospital Association (AHA), and various physician specialty groups, as well as individual physicians, medical record administrators, health information management professionals, and other members of the public, to contribute ideas on coding matters. After considering the opinions expressed at the public meetings and in writing, the Committee formulates recommendations, which then must be approved by the agencies.

The Committee presented proposals for coding changes for implementation in FY 2004 at a public meeting held on December 6, 2002, and finalized the coding changes after consideration of comments received at the meetings and in writing by January 10, 2003. Those coding changes are announced in Tables 6A and 6B of this final rule. Copies of the minutes of the procedure codes discussions at the Committee's 2002 meetings can be obtained from the CMS Web site: http://www.cms.gov/paymentsystems/icd9/. The minutes of the diagnoses codes discussions at the 2002 meetings are found at: http://www.cdc.gov/nchs/icd9.htm. Paper copies of these minutes are no longer available and the mailing list has been discontinued.

The first of the 2003 public meetings was held on April 3, 2003. In the September 7, 2001 final rule implementing the IPPS new technology add-on payments (66 FR 46906), we indicated we would attempt to include all proposals discussed and approved at the April meeting as part of the code revisions effective the following October. Because the proposed rule was published after the April meeting, we were able to include all new procedure codes that were approved subsequent to that meeting in Table 6B of the Addendum to the proposed rule, including the DRG assignments. However, the National Center for Health Statistics (NCHS) created and finalized three new severe acute respiratory syndrome (SARS) related codes after the proposed rule was published. These new codes, which were not listed in Table 6A of the Addendum to the proposed rule, have been included in Table 6A of the Addendum to this final rule. The new codes are as follows:

• 079.82, SARS-associated coronavirus
• 480.3, Pneumonia due to SARS-associated coronavirus
• V01.82, Exposure to SARA-associated coronavirus

These new codes have been identified with a footnote (1) in Table 6A of the Addendum to this final rule.

For a report of procedure topics discussed at the April 2003 meeting, see the Summary Report at: http://www.cms.hhs.gov/paymentsystems/icd9/. For a report of the diagnosis topics discussed at the April 2003 meeting, see the Summary Report at: http://www.cdc.gov/nchs/icd9.htm.

We encourage commenters to address suggestions on coding issues involving diagnosis codes to: Donna Pickett, Co-Chairperson; ICD-9-CM Coordination and Maintenance Committee; NCHS; Room 2404, 3311 Toledo Road, Hyattsville, MD 20782. Comments may be sent by E-mail to: dfp4@cdc.gov.

Questions and comments concerning the procedure codes should be addressed to: Patricia E. Brooks, Co-Chairperson; ICD-9-CM Coordination and Maintenance Committee; CMS, Center for Medicare Management, Hospital and Ambulatory Policy Group, Division of Acute Care; C4-08-06; 7500 Security Boulevard; Baltimore, MD 21244-1850. Comments may be sent by E-mail to: pbrooks1@cms.hhs.gov.

The ICD-9-CM code changes that have been approved will become effective October 1, 2003. The new ICD-9-CM codes are listed, along with their DRG classifications, in Tables 6A and 6B (New Diagnosis Codes and New Procedure Codes, respectively) in the Addendum to this final rule. As we stated above, the code numbers and their titles were presented for public comment at the ICD-9-CM Coordination and Maintenance Committee meetings. Both oral and written comments were considered before the codes were approved. Accordingly, in the May 19, 2003 proposed rule, we only solicited comments on the proposed DRG classification of these new codes.

Comment: One commenter expressed concern about the MDC and DRG designations for new diagnosis code 752.89 (Other specified anomalies of genital organs) that was included in Table 6A of the Addendum to the proposed rule. We had proposed assigning this new code to MDC 12 (Diseases and Disorders of the Male Reproductive System), and DRG 352 (Other Male Reproductive System Diagnoses). The commenter pointed out that this new code could apply to both males and females. Its predecessor code was assigned to MDC 12, DRG 352, as well as to MDC 13 (Diseases and Disorders of the Female Reproductive System) and DRGs 358 (Uterine and Adnexa Procedure for Non-Malignancy with CC), 359 (Uterine and Adnexa Procedure for Non-Malignancy without CC), and 369 (Menstrual and Other Female Reproductive System Disorders).

Response: The commenter is correct. Diagnosis code 752.89 would apply to both males and females and should have been included in both MDC 12 and MDC 13. In this final rule, we are assigning diagnosis code 752.89 to MDC 13 under DRGs 358, 359, and 369 and have modified Table 6A of the Addendum to this final rule accordingly.

Comment: One commenter pointed out a typographical error for the code title for V15.87. The commenter indicated that the word “membrance” should be changed to “membrane”; that is, the title should read “History of Extracorporeal Membrane Oxygenation (ECMO).”

Response: We agree with the commenter and have corrected the title in Table 6A of the Addendum to this final rule.

For codes that have been replaced by new or expanded codes, the corresponding new or expanded diagnosis codes are included in Table 6A. New procedure codes are shown in Table 6B. Diagnosis codes that have been replaced by expanded codes or other codes or have been deleted are in Table 6C (Invalid Diagnosis Codes). These invalid diagnosis codes will not be recognized by the GROUPER beginning with discharges occurring on or after October 1, 2003. Table 6D contains invalid procedure codes. Revisions to diagnosis code titles are in Table 6E (Revised Diagnosis Code Titles), which also includes the DRG assignments for these revised codes. Table 6F includes revised procedure code titles for FY 2004.

The Department of Health and Human Services has been actively working on the development of new coding systems to replace the ICD-9-CM. In December 1990, the National Committee on Vital and Health Statistics (NCVHS) issued a report noting that, while the ICD-9-CM classification system had been responsive to changing technologies and identifying new diseases, there was concern that the ICD classification might be stressed to a point where the quality of the system would soon be compromised. The ICD-10-CM (for diagnoses) and the ICD-10-PCS (for procedures) were developed in response to these concerns. These efforts have become increasingly important because of the growing number of problems with the ICD-9-CM, which was implemented 24 years ago.

Implementing ICD-10-PCS as a national standard was discussed at the December 6, 2002, ICD-9-CM Coordination and Maintenance Committee meeting. A complete report of the meeting, including examples of letters supporting and opposing ICD-10-PCS, can be found at the CMS Web site: http://www.cms.hhs.gov/paymentsystems/icd9/. Also, the Secretary has asked the NCVHS to recommend whether or not the country should replace ICD-9-CM as a national coding standard with ICD-10-CM and ICD-10-PCS. A complete report on the activities of this committee can be found at: http://www.ncvhs.hhs.gov.

Comment: Several commenters supported the move to ICD-10-CM and ICD-10-PCS as national coding standards. One commenter representing hospitals supported moving to these systems expeditiously. The commenter stated that ICD-10-CM and ICD-10-PCS are a vast improvement over ICD-9-CM and would provide greater specificity and detail in coding. Another commenter believed that the new systems would offer immediate and long-term benefits for specifying illness severity and accommodating a diverse array of new technologies that warrant expedited assignment under the DRG system.

Response: We appreciate the support from many in the health care industry for ICD-10-CM and ICD-10-PCS. We agree with the importance of having and maintaining medical coding systems that accurately capture the patient's conditions and medical procedures. We also agree that ICD-9-CM is seriously constrained because of its structure and space limitations. We recognize that over 30 countries have implemented ICD-10 to better capture medical conditions. Countries such as Canada and Australia have successfully implemented ICD-10 without serious ramifications to their data or reimbursement systems. We agree that it is important to capture information on new technologies. It is becoming increasingly difficult to do so using ICD-9-CM. We will continue working with NCVHS and the health care industry to determine if these new systems should be named as national coding standards.

14. Other Issues

In addition to the specific topics discussed in section II.B.1. through 13. of this preamble, we considered a number of other DRG-related issues in the May 19, 2003 proposed rule. Below is a summary of the issues that were addressed.

a. Cochlear Implants

Cochlear implants were first covered by Medicare in 1986 and were assigned to DRG 49 (Major Head and Neck Procedures) in MDC 3 (Diseases and Disorders of the Ear, Nose, Mouth, and Throat). This is the highest weighted surgical DRG in MDC 3. However, prior to the publication of the proposed rule, commenters contended that this DRG assignment is clinically and economically inappropriate for cochlear implants and requested a more specific DRG. The commenters contend that, like heart assist systems (for which we created a new DRG last year, DRG 525 (Heart Assist System Implant) in MDC 5), cochlear implants are low incidence procedures with disproportionately high costs compared to other procedures within DRG 49.

As we stated in the FY 2003 final rule in our discussion regarding the creation of DRG 525 (67 FR 49989), we found 185 heart assist system cases in DRG 104 (Cardiac Valve and Other Major Cardiothoracic Procedures with Cardiac Catheterization) and 90 cases in DRG 105 (Cardiac Valve and Other Major Cardiothoracic Procedures without Cardiac Catheterization). The average charges for these cases were approximately $36,000 and $85,000 higher than the average charges for cases in DRGS 104 and 105, respectively. However, these cases represented only a small fraction of all cases in these DRGs (1.3 percent and 0.5 percent, respectively). Therefore, despite the drastically higher average charges for heart assist systems, the relative volume was insufficient to affect the DRG weight to any great degree.

In our analysis of the FY 2002 MedPAR file, we found 134 cochlear implant cases out of 1,637 cases assigned to DRG 49, which represent more than 8 percent of the total cases in DRG 49. Compared to the situation with the heart assist system implant cases in DRGs 104 and 105, cochlear implants do have a greater effect on the relative weight for DRG 49. Also, while average charges for cochlear implant cases are significantly more than other cases in DRG 49 (average charges for cochlear implant cases were $51,549 compared to $25,052 for noncochlear implant cases), this difference is much less than the $36,000 and $85,000 differences for heart assist systems cited above.

Although we are concerned about the disparity between the average costs and payments for cochlear implant patients, we also have concerns about establishing a separate DRG for these cases. Doing so could create an incentive for some of these procedures to be shifted from outpatient settings, where most are currently performed. Even among current cochlear implant cases, our analysis found the average length of stay for Medicare patients receiving this procedure in the inpatient setting was just over 1 day, indicating minimal inpatient care is necessary for these cases. It is unclear whether a shift toward more inpatient stays would be appropriate.

We also are concerned whether the volume of cochlear implant cases across all hospitals performing this procedure warrants establishing a new DRG. The DRG relative weights reflect an average cost per case, with the costs of some procedures above the DRG mean costs and some below the mean. It is expected that hospitals will offset losses for certain procedures with payment gains for other procedures, while responding to incentives to maintain efficient operations. An excessive proliferation of new DRGs for specific technologies would fundamentally alter this averaging concept.

Accordingly, for the reasons cited above, we did not propose to change the DRG assignment of cochlear implants in the May 19, 2003 proposed rule. However, we did encourage public comments as to whether a new DRG for cochlear implants (or some other solution) is warranted.

Comment: Several commenters urged CMS to reassign cochlear implantation procedures to a DRG that has a weight appropriate to reflect the costs of cochlear implantation. The commenters stated that while a hospital's acquisition cost of the device itself averages approximately $23,800, the proposed payment for FY 2004 is approximately $8,233. While most cochlear implants have been and will continue to be performed on an outpatient basis, a small, but significant portion, particularly for Medicare beneficiaries, need to be conducted as an inpatient procedure. The commenters stated that the low volume of inpatient cases is a direct result of the inadequate payment rate.

The commenters stated that cochlear implantation is clinically incongruent and economically inconsistent with the other procedures in DRG 49. The commenters believed that cochlear implants do not meaningfully affect the weighting of DRG 49 and proposed two options: Create a new DRG specifically for cochlear implants, or reassign cochlear implants cases to DRG 482 (Tracheostomy for Face, Mouth, and Neck Diagnoses).

Response: We requested public input on possible solutions for these cases because we recognize the data indicate the charges for these cases are much higher than for other cases in DRG 49. However, we are concerned that the options suggested by commenters are not workable solutions. As we alluded to in the proposed rule, we have concerns about creating a new DRG for this procedure. We appreciate the point made by commenters that only those patients requiring inpatient care would receive the procedure in an inpatient setting, even if the DRG payment were increased. However, as we have stated previously, we are reluctant to create new DRGs for specific, low-volume procedures. Doing so would create a proliferation of DRGs and a loss of some of the efficiency incentives inherent in the current system. Hospitals are generally able to offset any losses on such procedures through corresponding payment advantages from other, less expensive procedures.

The second option suggested, to reassign these cases to DRG 482, is inconsistent with the structure of that DRG, which requires that a tracheostomy be performed in order to be assigned to this DRG. Assigning cochlear implants to this DRG would fundamentally alter its structure, which could not be done without first proposing such a change for public review and comment.

However, as we indicated above, we recognize the disparity in average charges for these cases compared to other cases in DRG 49, and will continue to evaluate possible reclassification options for FY 2005.

b. Burn Patients on Mechanical Ventilation

Prior to the publication of the proposed rule, concerns were raised by hospitals treating burn patients that the current DRG payment for burn patients on mechanical ventilation is not adequate. The DRG assignment for these cases depends on whether the hospital performed the tracheostomy, or the tracheostomy was performed prior to transfer to the hospital. If the hospital does not actually perform the tracheostomy, the case is assigned to one of the burn DRGs in MDC 22 (Burns). If the hospital performs a tracheostomy, the case is assigned to DRG 482 (Tracheostomy for Face, Mouth, and Neck Diagnoses) or DRG 483 (Tracheostomy with Mechanical Ventilation 96 + Hours, Except Face, Mouth and Neck Diagnoses).

In the August 1, 2002 final rule, we modified DRGs 482 and 483 to recognize code 96.72 (Continuous mechanical ventilation for 96 consecutive hours or more) for the first time in the DRG assignment (67 FR 49996). We noted that many patients assigned to DRG 483 did not have code 96.72 recorded. We believed this was due, in part, to the limited number of procedure codes (six) that can be submitted on the current billing form, and the fact that code 96.72 did not affect the DRG assignment (prior to FY 2003). We stated that we would give future consideration to further modifying DRGs 482 and 483 based on the presence of code 96.72. We anticipate that cases of patients receiving 96 or more hours of continuous mechanical ventilation are more expensive than other tracheostomy patients. Once code 96.72 is reported more frequently, we will be better able to assess the need for future revisions to DRGs 482 and 483.

To assess the payment for burn patients on mechanical ventilation when the hospital did not perform the tracheostomy, we analyzed data on cases reporting both code 96.72 and diagnosis code V44.0 (Tracheostomy status). We had hoped that these cases would show patients on long-term ventilation who were admitted to the hospital with a tracheostomy in place. Our data did not include any cases reported in any of the burn DRGs with codes 96.72 and V44.0. We then analyzed data on the frequency of cases reporting code 96.72 along with diagnosis code V46.1 (Respirator dependence). We found only 5 of these cases in the burn DRGs. With so few cases reporting code 96.72, it is difficult for us to determine the effect of long-term ventilation on reimbursement for burn cases.

All hospitals, including those that treat burn patients, are encouraged to increase the reporting of code 96.72 for patients who are on continuous mechanical ventilation for 96 or more hours. With better data, we would be able to determine how best to make any future DRG modification for all patients on long-term mechanical ventilation.

We received one comment from an organization representing coders that agreed with the importance of reporting code 96.72 and the need for further education on this issue. We will continue to monitor our data to assess the payment for burn patients on mechanical ventilation in the future.

c. Multiple Level Spinal Fusion

Prior to the publication of the proposed rule, we received a comment recommending the establishment of new DRGs that would differentiate between the number of levels of vertebrae involved in a spinal fusion procedure. The commenter noted that the ICD-9-CM Coordination and Maintenance Committee discussed adding a new series of codes to identify multiple levels of spinal fusions at its December 6, 2002 meeting.

The following codes were approved by the Committee, effective for October 1, 2003, and are listed in Table 6B in the Addendum to this final rule:

• 81.62, Fusion or refusion of 2-3 vertebrae
• 81.63, Fusion or refusion of 4-8 vertebrae
• 81.64, Fusion or refusion of 9 or more vertebrae

The commenter conducted an analysis to support redefining the spinal fusion DRGs using these new ICD-9-CM codes. Using the CMS FY 2001 Standard Analytical File data for physicians and hospitals as the basis for its analysis, the commenter linked a 5-percent sample of hospital spinal fusion cases with the corresponding physician claims. Because there were no ICD-9-CM codes to identify multiple level fusions in 2001, multiple level fusions were identified using Current Procedural Terminology (CPT) codes on the physician claims.

The analysis found that increasing the levels fused from 1 to 2 levels to 3 or more levels increased the mean standardized charges by 38 percent for lumbar/thoracic fusions, and by 47 percent for cervical fusions. The commenter then recommended redefining the spinal fusion DRGs to differentiate between 1 to 2 level spinal fusions and multilevel spinal fusions.

The following current spinal fusion DRGs separate cases based on whether or not a CC is present: DRG 497 (Spinal Fusion Except Cervical With CC) and DRG 498 (Spinal Fusion Except Cervical Without CC); and DRG 519 (Cervical Spinal Fusion With CC) and DRG 520 (Cervical Spinal Fusion Without CC). The difference in charges associated with the current CC split is only slightly greater than the difference attributable to the number of levels fused as found by the commenter's analysis. Therefore, in the May 19, 2003 proposed rule, we did not propose to redefine these DRGs to differentiate on the basis of the number of levels fused.

We note that adopting the commenter's recommendation would necessitate adjusting the DRG relative weights using non-MedPAR data, because Medicare claims data with the new ICD-9-CM codes will not be available until the FY 2003 MedPAR file. Although we considered this possibility, we believe the more prudent course, given that the current DRG structure actually appears to differentiate appropriately among these cases, is to wait until sufficient data with the new multilevel spinal fusion codes are available before making a final determination on whether multilevel spinal fusions should be incorporated into the DRG structure.

Comment: Several commenters supported our proposal to wait for data using the new ICD-9-CM procedure codes for multiple level spinal fusions prior to making revisions to the spinal fusion DRGs. One commenter representing hospitals supported our proposal to continue with the current DRG classification system until sufficient data are available to evaluate a potential DRG change. Several commenters expressed their appreciation for the creation of the new codes for multiple level spinal fusion. They recognized the difficult challenge that was involved in developing this new classification system as part of ICD-9-CM.

One commenter requested us to proceed with a DRG revision for multiple level spinal fusion without waiting for data using the new codes. This commenter stated that there are significant costs involved with increased instrumentation and hardware when multiple level spinal fusions are performed, and requested that we consider using non-MedPAR data to establish relative weights for new DRGs based on the levels of vertebrae involved. In addition, the commenter stated that there is a need to distinguish between fusions and refusions within the DRGs. The commenter stated that refusions vary significantly due to the existence of scar tissue and implants that need to be removed and replaced. Further, the commenter recommended that we split DRG 496 Combined anterior/posterior spinal fusion based on the presence or absence of a complication or comorbidity.

Response: We appreciate the support of commenters that we wait for data from the reporting of the new codes for multiple level spinal fusion prior to proposing revisions to the spinal DRGs (rather than using non-MedPAR data prior to the availability of data using the new codes). We also appreciate the comments concerning the extensive effort it took on our part to develop a set of ICD-9-CM codes that could capture this type of information. We believe it is important to carefully examine hospital data prior to making any revisions for multiple level spinal fusions. Therefore, we will look at this data as we receive it and evaluate any need for DRG revisions. We will consider all the points raised by the commenters as we consider additional DRG revisions for spinal fusions in the future.

d. Heart Assist System Implant

During the comment period for the FY 2003 IPPS proposed rule on which the FY 2003 IPPS final rule was based, we received a suggestion from a commenter that we develop a new heart transplant DRG entitled “Heart Transplant with Left Ventricular Assist Device (LVAD).” The commenter stated that, because a great number of LVAD cases remain inpatients until heart transplant occurs, there is a disparity in costs between heart transplant patients who receive LVADs during the stay and those who do not. Cases in which heart transplantation occurs during the hospitalization are assigned to DRG 103 (Heart Transplant). Therefore, the costs of these LVAD cases where a heart transplant is also performed during the same hospitalization are included in the DRG relative weight for DRG 103. Accordingly, we did not create a new DRG for these cases. However, we noted that we would continue to monitor these types of cases.

When we reviewed the FY 2002 MedPAR data, we identified only 21 cases in DRG 103 that listed a procedure code indicating the use of any heart assist system. We do not believe that 21 cases is a sufficient number of cases to support creation of an additional DRG. Therefore, in the May 19, 2003 proposed rule, we did not propose a change to the structure of either DRG 103 or DRG 525.

Comment: Two commenters argued that procedure code 37.66 (Implant of an implantable, pulsatile heart assist system) does not fit clinically or financially with the following other procedure codes in DRG 525:

• 37.62, Implant of other heart assist system,
• 37.63, Replacement and repair of heart assist system,
• 37.65, Implant of an external, pulsatile heart assist system
• 37.66, Implant of an implantable, pulsatile heart assist system.

One commenter indicated that, according to an analysis that it performed, Medicare data on procedure code 37.66 demonstrates that average charges ($342,725) and length of stay (40.1 days) are significantly higher than data on all other procedures in DRG 525 (average charges ranging from $112,748 to $190,672) and (average length of stay ranging from 10.9 to 16.7). According to the commenter, the implantable pulsatile technology represents a different class of device and procedure (long-term support) compared to the less resource intensive, short-term devices used in other procedures in DRG 525.

The commenters requested three possible alternatives for the reclassification of procedure code 37.66: (1) Create a unique DRG for this procedure; (2) add this procedure code to DRG 103 (Heart Transplant); or (3) add a new technology add-on payment for code 37.66 to DRG 525.

Response: In response to comments we received on the creation of new DRG 525 last year, we noted that these four codes represent the most expensive cases in MDC 5 (67 FR 49991). However, the specific point made by the commenters this year, that procedure code 37.66 is significantly different in terms of clinical procedures and resource utilization from the other procedures in DRG 525, was not raised prior to this year's proposed rule.

While we recognize the significant disparities referenced by the commenter warrant further consideration, the potential solutions suggested by the commenter are significant changes to the DRG system that warrant public comment. In particular, the reassignment of code 37.66 to DRG 103 would result in inclusion of nontransplant cases in this existing single-procedure DRG. Therefore, in light of the significant impacts of each of the commenters' suggestions on the structure of the DRGs involved and the need to submit any such significant impacts to public review and comment, we are not changing DRG 525 for FY 2004. We appreciate the commenter bringing this issue to our attention. We will evaluate whether to make further changes to DRG 525 in light of the information that there is significant disparity in the costs of the different procedures included in the DRG. We note that the outlier payment policy will help to offset extraordinarily expensive costs.

Furthermore, the volume and mix of cases in this DRG is likely to change over the next year. Currently, CMS has approved the use of LVADs in two instances. They can be used as either a bridge to heart transplant or for support of blood circulation postcardiotomy (the period following open-heart surgery). In these two applications, the LVAD is used as temporary mechanical circulatory support. CMS is currently reviewing a request for expanded coverage for these devices as destination (or permanent) therapy for end-stage heart failure patients who are not candidates for heart transplantation. Destination therapy means that the patient will use the LVAD for the remainder of his or her life.

We believe it will be helpful to have data on the resources and volume associated with any potential destination therapy cases prior to revising DRG 525.

e. Drug-Eluting Stents

In the August 1, 2002 final rule, we created two new temporary DRGs to reflect cases involving the insertion of a drug-eluting coronary artery stent as signified by the presence of code 36.07 (Insertion of drug-eluting coronary artery stent): DRG 526 (Percutaneous Cardiovascular Procedure With Drug-Eluting Stent With AMI); and DRG 527 (Percutaneous Cardiovascular Procedure With Drug-Eluting Stent Without AMI). We expect that when claims data are available that reflect the use of these stents, we will combine drug-eluting stent cases with other cases in DRGs 516 and 517.

In the absence of MedPAR data reflecting the use of drug-eluting stents, it was necessary to undertake several calculations to establish the FY 2003 DRG relative weights for these two new DRGs. First, based on prices in countries where drug-eluting stents were already being used compared to the average price of nondrug-eluting stents in those countries, we calculated a price differential of approximately $1,200. When we apply average overall hospital charge markups to this technology (based on weighted average cost-to-charge ratios), we estimated that the charge differential between nondrug-eluting and drug-eluting stents would be approximately $2,664 per stent. However, we recognize that some cases involve more than one stent. Using an average of 1.5 stents per procedure, we estimated that the net incremental charge for cases that would receive drug-eluting stents is $3,996.

In order to determine accurately the DRG relative weights for these two new DRGs relative to all other DRGs, we also must estimate the volume of drug-eluting stent cases likely to occur. We used the manufacturer's estimate that as many as 43 percent of current stent patients will receive drug-eluting stents during FY 2003 to calculate the FY 2003 DRG relative weights, although we prorated this percentage since the new DRGs did not become active until April 1, 2003.

In determining the FY 2004 DRG relative weights for DRGs 526 and 527, we assumed that 43 percent of coronary stent cases (those with code 36.06 (Insertion of nondrug-eluting coronary artery stent)) from DRGs 516 and 517 would be reassigned to new DRGs 526 and 527 (with code 36.07), and the charges for these cases would be increased $3,996 per case, to approximate the higher charges associated with the drug-eluting stents in DRGs 526 and 527. The relative weights for DRGs 516 and 517 are calculated based on the charges of the cases estimated to remain in these two DRGs.

Comment: In response to our statement in the proposed rule that we would use the best available data to establish the FY 2004 relative weights for DRGs 526 and 527, one commenter (the manufacturer of the only FDA-approved drug-eluting stents at this time) commissioned an independent accounting firm to collect costs, charges, and utilization data from hospitals on drug-eluting and nondrug-eluting stents.

The data were collected from a randomized, statistically significant sample of United States hospitals with interventional cardiac catherization laboratories. First, the firm identified those hospitals that performed coronary angioplasty on Medicare beneficiaries. The method used to identify these hospitals was first to review MedPAR data to isolate those hospitals with average volume in DRGs with a placement of coronary artery stent, ICD-9-CM procedure code (36.06). From this list of hospitals, it was necessary to eliminate those that appeared to have quality issues with the data. This resulted in a list of 1,033 hospitals for the “population” group from which the sample was drawn.

A sample size sufficient to achieve a confidence level of 95 percent that the results would be within 5 percent of the actual distribution (assuming a normal distribution) was then determined, and a randomized selection within each state identified 279 hospitals. An additional 30 hospitals from a preliminary phase of the study were added because these hospitals had already supplied nondrug-eluting stent data and had committed to supply drug-eluting stent data. Therefore, the total sample size for the survey instrument was 309 hospitals.

At the time of the survey, 83 of the selected hospitals had not yet received shipments of the drug-eluting stents and, hence, were not able to complete the survey because they had no cost or charge data for drug-eluting stents. The final number of completed surveys was 119 (or 53 percent of the sample).

The survey was designed to collect data regarding costs, charges, and utilization for drug-eluting stents at three different points in time: currently; October 1, 2003; and at full-maturity (defined as that point in time in which the hospital has achieved a stable and consistent usage of the drug-eluting stent). The data were submitted (including a sample of invoices) under a request for confidential treatment under the Freedom of Information Act.

Based on the data collected, the commenter recommended that CMS increase the harge differential between nondrug-eluting and drug-eluting stents to create a payment differential of $3,024. This represents the cost per case differential between nondrug-eluting stent and drug-eluting stent cases anticipated by surveyed hospitals on October 1, 2003. The current cost differential reported by the sample of hospitals was $2,721. The commenter estimated that our proposed methodology results in a payment differential of $1,451 and $1,495 between DRGs 516 and 526, and DRGs 517 and 527, respectively. The surveyed hospitals reported average current and anticipated stents used per case of 1.4 and 1.5, respectively. Average projected utilization of drug-eluting stents relative to all stents was reported in the survey to currently be 33 percent, and by October 1, 2003, utilization is projected to be 69 percent.

Another commenter noted that the actual cost per stents is 59 percent higher than our projection of $1,200. The commenter also noted that most cases use 2 stents instead of the projected 1.5 stents, and, therefore, the net incremental charge difference should be $5,554 instead of the $3,996 projected by CMS.

Response: The data submitted was extensively detailed and helped us better understand the costs, charges, and utilization for all types of stents. As noted above, we stated in the proposed rule that we would use the best available data at the time of the final rule to establish the FY 2004 relative weights for DRGs 526 and 527, and these data are much more detailed and current than any other sources available to us at this time. These data are extremely useful to assess the appropriateness of our proposed methodology to determine the relative weights for DRGs 526 and 527.

The commenter recommended that CMS establish a payment differential between DRGs for nondrug-eluting stents and drug-eluting stents of $3,024 to account for the estimated cost difference between the two types of stents. However, the DRG relative weights are established using the average charges per case of each DRG relative to the national average. Therefore, we examined the charge per case data from the sample.

The commenter referred to a mean charge differential per case of $5,721, based on anticipated costs per drug-eluting stent on October 1, 2003. However, we do not believe it is appropriate to use anticipated October 1, 2003 charges for several reasons. First, these data cannot be substantiated. As noted above, we received a sampling of current invoices that allowed us to verify the current costs per drug-eluting stent. These invoices cannot verify the $300 average per stent cost increase that reportedly will occur between the time the survey was conducted and October 1, 2003. Second, for all other DRGs, we are using charge data reflective of FY 2002 charges. Although we are establishing the FY 2004 relative weights in this final rule, using anticipated FY 2004 charge data would result in 2-year later charge data being used to establish the DRG 526 and 527 relative weights, while FY 2002 charge data are used to establish all other relative weights. Therefore, we believe the current data more closely approximate the data used to determine the FY 2004 relative weights for the remainder of the DRGs. Finally, hospitals must rely upon the manufacturer of the only currently available drug-eluting stents for information on future pricing. We believe this raises questions as to the validity of the data due to the lack of independently verifiable pricing data for the future.

Therefore, we are basing our evaluation of our proposed methodology on the sample data from the current period. The commenter reported a mean differential in charges per case of $4,859 for the current period. However, we are concerned that the mean differential in charges per case is unduly influenced by extraordinarily high charge markups reported on the part of some hospitals. For example, one hospital reported charging $28,000 per drug-eluting stent, while its costs per stent were only $3,023. This same hospital reported charges of $9,500 for nondrug-eluting stents, with costs per stent of $1,010. To control the distorting impact such a hospital would have on the mean charge differential, we examined the geometric mean charge differential based on current charges per case.

The survey data showed that, for seven hospitals, the charge per case was higher for nondrug-eluting stent cases. In order to calculate the geometric mean differential charge per case, it was necessary to remove these seven negative differentials. The result was a current geometric mean differential charge per case of $4,186. As an alternative to removing these seven negative numbers, we set them to a $1 differential, and calculated a geometric mean differential charge per case of $2,291. Based on the range of these results, we believe our proposed charge differential of $3,996 represents a reasonable approximation of the differential in charges per case, and we are proceeding to establish the DRG relative weights for DRGs 526 and 527 for FY 2004 using this amount.

We note that there is a difference between CMS and the commenter on the current cost difference between drug-eluting stents and nondrug-eluting stents (our estimate began with a $1,200 per stent differential, while the survey found a $2,721 current differential). It appears that the reason our charges per case for drug-eluting stents and nondrug-eluting stents are not substantially different from the charges in the survey data, despite the discrepancy in the cost differential, is due to the fact that hospitals are not marking up drug-eluting stents by the same proportion as nondrug-eluting stents. From the data submitted by the commenter, we found the average charge increase for nondrug-eluting stents is 183 percent. The average charge increase for drug-eluting stents is 124 percent. This lower markup reduces the differential in charges relative to the actual costs hospitals may incur.

Based on data submitted to us last year by the commenter, we proposed that 43 percent of stent cases from DRGs 516 and 517 would be reassigned to DRGs 526 and 527. However, based on the survey data, for FY 2004 we are changing our estimate to assume that 69 percent of coronary stent cases will be reassigned from DRGs 516 and 517 to DRGs 526 and 527, respectively. We note that, although this percentage is based on anticipated utilization on October 1, 2003, it is not based on data that is only available from the manufacturer. We are continuing to assume a utilization rate of 1.5 stents per case.

Comment: Many commenters argued that the proposed payment for drug-eluting stents is inadequate and asked that CMS consider the data it has received to date from hospital claims to determine whether the proposed FY 2004 payment rate for drug-eluting stents is adequate. Other commenters requested that CMS use the most current United States data available (as opposed to data from the United Kingdom) to establish the DRG weights for FY 2004.

Some commenters noted that current DRG weights account for 1.5 stents per case, but that the number of stents per case is expected to rise because the insertion of drug-eluting stents is more technically challenging in comparison to competitive products. The commenters also noted that because drug-eluting stents are able to treat smaller vessels, more diffuse disease in diabetics, and longer lesions, a rise is expected in the stent per patient ratio. The commenters asked that CMS adjust its ratio of 1.5 stents per case to an amount closer to 2 stents per case when recalibrating the DRG weights. Another commenter explained that, based on their analysis, an average of 1.7 drug-eluting stents is used per procedure and the average cost per drug-eluting stent is $3,195. The commenter requested that these amounts be used to compute the relative weights for DRGs 526 and 527. The commenter also noted that the payment rates for FY 2003 are higher than the payment rates for FY 2004 due to the decline in the DRG relative weights.

One commenter suggested as an alternative to increasing the weights for drug-eluting stents that payment be contingent on the type and number of stents used per procedure. The commenter recommended that CMS set up revenue codes to indicate the type and number of stents used per case and make payment approximately $1,000 above the cost per stent.

Another commenter also noted that the demand from hospitals for drug-eluting stents is much higher than the projected 43 percent of coronary artery stent cases. The commenter estimated that 85 to 90 percent of all stent cases should be reassigned from DRGs 516 and 517 to DRGs 526 and 527. Another commenter explained that drug-eluting stents, compared with nondrug-eluting stents, have already been shown to decrease angiographic restenosis in coronary arteries by more than half, which should reduce the need for repeat procedure rates from 20 percent of cases to less than 5 percent. As a result, demand for drug-eluting stents is expected to increase and the commenter estimated that 70 percent of all coronary artery stent cases will involve the use of drug-eluting stents. Therefore, 70 percent of all stent cases should be moved to DRGs 526 and 527 to account for drug-eluting stents instead of the 43 percent proposed by CMS.

One commenter explained that there are many added costs of using drug-eluting stents, such as that the area of blockage to be treated is to be predilated with an angioplasty balloon before and after implanting the stent, the use of intravascular ultrasound to ensure proper positioning and deployment of stents in certain cases, and increased length of time a patient spends in the cardiac catheterization laboratory. The commenter also added that percutaneous transluminal coronary angioplasty volume is expected to increase due to obesity, smoking, sedentary lifestyle, and diabetes. Therefore, the commenter recommended that CMS ensure that drug-eluting stents are adequately paid.

Response: As described above, we used data submitted to us from a survey of U.S. hospitals to evaluate our proposed methodology. Our analysis indicates that the proposed charge differential and the number of stents per procedure in our methodology are appropriate. However, we have increased our assumed utilization rate of drug-eluting stents to 69 percent from 43 percent, based on these data.

With respect to the decline in the proposed FY 2004 DRG relative weights compared to FY 2003, every year we recalibrate the DRG weights comparing the average charge per DRG to all other DRGs. The weights of one DRG can change for numerous reasons (for example, increase or decrease in total cases or increase or decrease in charges) and cause weights from other DRGs to increase or decrease due to budget neutrality.

As we proposed, we are maintaining DRGs 526 and 527 for FY 2004, and adopting the same methodology to establish the relative weights as we used for FY 2003. We have used the best available data to establish the final FY 2004 relative weights for DRGs 526 and 527 included in this final rule. We will continue to evaluate the appropriate assignment of these cases in the future.

Comment: One commenter recommended that CMS move drug-eluting stents to DRGs 516 and 517 and adjust the weights, because CMS should not provide a financial incentive for hospitals to favor one therapy when other alternatives with equal or better outcomes are available. The commenter stated further that CMS should not create an incentive that promotes a more expensive treatment for which risks and benefits are not yet completely known. Another commenter suggested that drug-eluting stents should receive add-on payments for new technology instead of receiving their own DRG payment.

Response: We explained our rationale for creating new DRGs 525 and 526 (instead of assigning these cases to DRGs 516 or 517 or approving a new technology add-on) in the August 1, 2002 IPPS final rule (67 FR 50005) and refer the commenters to that rule for our response. We appreciate the commenter's continual input and interest in these issues.

f. Artificial Anal Sphincter

The ICD-9-CM Coordination and Maintenance Committee created two new codes to describe procedures involving an artificial anal sphincter for use for discharges occurring on or after October 1, 2002. One code (49.75, Implantation or revision of artificial anal sphincter) is used to identify cases involving implantation or revision of an artificial anal sphincter. The second code (49.76, Removal of artificial anal sphincter) is used to identify cases involving the removal of the device. In Table 6B of the August 1, 2002 IPPS final rule (67 FR 50242), we assigned both codes to one of four MDCs based on principal diagnosis, and to one of six DRGs within those MDCs as follows: MDC 6 (Diseases and Disorders of the Digestive System), DRG 157 (Anal and Stomal Procedures With CC) and DRG 158 (Anal and Stomal Procedures Without CC); MDC 9 (Diseases and Disorders of the Skin, Subcutaneous Tissue and Breast), DRG 267 (Perianal and Pilonidal Procedures); MDC 21 (Injuries, Poisonings, and Toxic Effect of Drugs), DRG 442 (Other O.R. Procedures for Injuries With CC) and DRG 443 (Other O.R. Procedures for Injuries Without CC); and MDC 24 (Multiple Significant Trauma), DRG 486 (Other O.R. Procedures for Multiple Significant Trauma).

Prior to the publication of the proposed rule, we received a request that we review these DRG assignments. According to the requester, the artificial anal sphincter procedures are expensive and the payment does not adequately cover a hospital's costs in the most likely occurring DRGs: DRG 157 and DRG 158. The requester submitted data showing cases involving artificial anal sphincters with average charges of $44,000, and suggested that we assign codes 49.75 and 49.76 in MDC 6 to DRG 170 (Other Digestive System O.R. Procedures With CC) and DRG 171 (Other Digestive System O.R. Procedures Without CC) because DRG 170 and DRG 171 are higher weighted than DRGs 157 and 158.

In the May 19, 2003 proposed rule, we did not propose to assign these cases to DRGs 170 and 171. Although we recognized that the data submitted by the commenter appear to show this procedure is associated with above average costs in the DRGs to which these cases are assigned, we stated that we believe the current assignment is the most clinically appropriate at this time. As noted above, the procedure codes to identify the implantation, revision, or removal of these devices were effective beginning on October 1, 2002. Therefore, we proposed to monitor the costs of these cases using actual Medicare cases with these codes included from the FY 2003 MedPAR that will be used for the FY 2004 DRG relative weights.

Comment: Two commenters expressed concern that the procedures for insertion and removal of an artificial anal sphincter are assigned to DRG groupings that do not cover the cost of the device. In addition, one commenter stated that, as the surgeon must operate on two distinct areas of the patient's body, these procedures are more resource-intensive and, therefore, are not clinically coherent with other procedures of low complexity in DRGs 157 and 158.

Response: As noted above, the codes describing the implantation, revision, or removal of artificial anal sphincters were created for use beginning on October 1, 2002. Therefore, we do not have data on cases assigned to codes 49.75 and 49.76. Accordingly, we are not making any changes to the DRG assignments of these codes at this time. However, we will continue to monitor this procedure in the upcoming MedPAR data and will, in the future, consider modifications relating to DRG assignment(s) if warranted.

C. Recalibration of DRG Weights

As we proposed, in this final rule we used the same basic methodology for the FY 2004 recalibration as we did for FY 2003 (August 1, 2002 IPPS final rule (67 FR 50008). That is, we recalibrated the DRG weights based on charge data for Medicare discharges using the most current charge information available (the FY 2002 MedPAR file).

The MedPAR file is based on fully coded diagnostic and procedure data for all Medicare inpatient hospital bills. The FY 2002 MedPAR data used in this final rule include discharges occurring between October 1, 2001 and September 30, 2002, based on bills received by CMS through March 31, 2003, from all hospitals subject to the IPPS and short-term acute care hospitals in Maryland (which is under a waiver from the IPPS under section 1814(b)(3) of the Act). The FY 2002 MedPAR file includes data for approximately 11,496,239 Medicare discharges. Discharges for Medicare beneficiaries enrolled in a Medicare+Choice managed care plan are excluded from this analysis. The data excludes CAHs, including hospitals that subsequently became CAHs after the period from which the data were taken. This is a change from the recalibration methodology in the proposed rule, where hospitals that subsequently became CAHs were included in the data. In this final rule, we changed the recalibration methodology for consistency with our change that excluded these CAHs from the data used to construct the wage index.

The methodology used to calculate the DRG relative weights from the FY 2002 MedPAR file is as follows:

• To the extent possible, all the claims were regrouped using the DRG classification revisions discussed in section II.B. of this preamble.
• The transplant cases that were used to establish the relative weight for heart and heart-lung, liver, and lung transplants (DRGs 103, 480, and 495) were limited to those Medicare-approved transplant centers that have cases in the FY 2000 MedPAR file. (Medicare coverage for heart, heart-lung, liver, and lung transplants is limited to those facilities that have received approval from CMS as transplant centers.)
• Organ acquisition costs for kidney, heart, heart-lung, liver, lung, pancreas, and intestinal (or multivisceral organs) transplants continue to be paid on a reasonable cost basis. Because these acquisition costs are paid separately from the prospective payment rate, it is necessary to subtract the acquisition charges from the total charges on each transplant bill that showed acquisition charges before computing the average charge for the DRG and before eliminating statistical outliers.
• Charges were standardized to remove the effects of differences in area wage levels, indirect medical education and disproportionate share payments, and, for hospitals in Alaska and Hawaii, the applicable cost-of-living adjustment.
• The average standardized charge per DRG was calculated by summing the standardized charges for all cases in the DRG and dividing that amount by the number of cases classified in the DRG. A transfer case is counted as a fraction of a case based on the ratio of its transfer payment under the per diem payment methodology to the full DRG payment for nontransfer cases. That is, a transfer case receiving payment under the transfer methodology equal to half of what the case would receive as a nontransfer would be counted as 0.5 of a total case.
• Statistical outliers were eliminated by removing all cases that are beyond 3.0 standard deviations from the mean of the log distribution of both the charges per case and the charges per day for each DRG.
• The average charge for each DRG was then recomputed (excluding the statistical outliers) and divided by the national average standardized charge per case to determine the relative weight.

The new weights are normalized by an adjustment factor (1.45726) so that the average case weight after recalibration is equal to the average case weight before recalibration. This adjustment is intended to ensure that recalibration by itself neither increases nor decreases total payments under the IPPS.

As noted below in section IV.A.2. of the preamble of this final rule, we are expanding the transfer policy applicable to postacute care transfers to a total of 29 DRGs (the current 10 DRGs, minus 2, plus 21 additional DRGs), beginning in FY 2004. Because we count a transfer case as a fraction of a case as described above in the recalibration process, the expansion of the postacute care transfer policy to additional DRGs affects the relative weights for those DRGs. Therefore, we calculated the final FY 2004 normalization factor comparing: the case-mix using the final FY 2004 DRG relative weights in which we treated postacute care transfer cases in the additional DRGs for the postacute transfer policy for FY 2004 as a fraction of a case with the case-mix using the FY 2003 DRG relative weights without treating cases in these additional DRGs as transfer cases.

When we recalibrated the DRG weights for previous years, we set a threshold of 10 cases as the minimum number of cases required to compute a reasonable weight. We used that same case threshold in recalibrating the final DRG weights for FY 2004. Using the FY 2002 MedPAR data set, there are 42 DRGs that contain fewer than 10 cases. We computed the weights for these low-volume DRGs by adjusting the FY 2003 weights of these DRGs by the percentage change in the average weight of the cases in the other DRGs.

Comment: Commenters questioned the fact that the proposed weights for several DRGs declined from the prior fiscal year.

Response: As described above, the relative weight for each DRG is calculated by comparing the average charge for cases within each DRG (after removing statistical outliers) with the national average charge per case. Therefore, there are several factors that can cause a shift in the relative weight of a DRG from one fiscal year to the next. For example, even though the average charges of cases within a particular DRG may have increased, if they did not increase by an equal or greater percentage than the national average, the DRG relative weight would decline. In this final rule, the weights for 223 DRGs for FY 2004 decline from those for FY 2003 (all but 38 DRGs by less than 5 percent), while the weights for 299 DRGs for FY 2004 increased from those for FY 2003 (all but 39 DRGs by less than 5 percent).

Section 1886(d)(4)(C)(iii) of the Act requires that, beginning with FY 1991, reclassification and recalibration changes be made in a manner that assures that the aggregate payments are neither greater than nor less than the aggregate payments that would have been made without the changes. Although normalization is intended to achieve this effect, equating the average case weight after recalibration to the average case weight before recalibration does not necessarily achieve budget neutrality with respect to aggregate payments to hospitals because payments to hospitals are affected by factors other than average case weight. Therefore, as we have done in past years and as discussed in section II.A.4.a. of the Addendum to this final rule, we are making a budget neutrality adjustment to ensure that the requirement of section 1886(d)(4)(C)(iii) of the Act is met.

Comment: One commenter expressed concern that the impact of the proposed DRG recalibration is a $3 million decrease in payments to its hospitals. The commenter was hopeful that the budget neutrality adjustment to ensure that the normalization of DRG weights is achieved will somehow restore the estimated negative impact.

Response: As explained above and in the proposed rule, section 1886(d)(4)(C)(iii) of the Act requires that the changes made through DRG reclassification and recalibration be made in a manner that assures that the aggregate payments are neither greater than nor less than the aggregate payment that would have been made without the changes. However, this requirement refers to aggregate national payments. Therefore, for individual hospitals, the impacts of these changes may be either positive or negative.

D. LTC-DRG Reclassifications and Relative Weights for LTCHs for FY 2004

1. Background

In the June 6, 2003 LTCH PPS final rule (68 FR 34122) we changed the LTCH PPS annual payment rate update cycle to be effective July 1 through June 30 instead of October 1 through September 30. In addition, since the patient classification system utilized under the LTCH PPS is based directly on the DRGs used under the IPPS for acute care hospitals, in that same final rule, we explained that the annual update of the long-term care diagnosis-related group (LTC-DRG) classifications and relative weights will continue to remain linked to the annual reclassification and recalibration of the CMS-DRGs under the IPPS.

The annual update to the IPPS DRGs is based on the annual revisions to the ICD-9-CM codes and is effective each October 1. In the health care industry, annual changes to the ICD-9-CM codes are effective for discharges occurring on or after October 1 each year. The use of the ICD-9-CM coding system is also compliant with the requirements of the Health Insurance Portability and Accountability Act (HIPAA), Pub. L. 104-191, under 45 CFR parts 160 and 162. Therefore, the manual and electronic versions of the GROUPER software, which are based on the ICD-9-CM codes, are also revised annually and effective for discharges occurring on or after October 1 each year. Because the LTC-DRGs are based on the patient classification system used under the IPPS (CMS-DRGs), which is updated annually and effective for discharges occurring on or after October 1 through September 30 each year, in the June 6, 2003 LTCH PPS final rule (68 FR 34128), we specified that we will continue to update the LTC-DRG classifications and relative weights to be effective for discharges occurring on or after October 1 through September 30 each year. Furthermore, we stated that we will publish the annual update of the LTC-DRGs in the proposed and final rules for the IPPS.

As we explained in the May 19, 2003 IPPS proposed rule (68 FR 27173), we proposed revisions to the LTC-DRG classifications and relative weights and indicated that we would finalize them in the IPPS final rule, to be effective October 1, 2003 through September 30, 2004. The final LTC-DRGs and relative weights for FY 2004 in this final rule are based on the IPPS DRGs (GROUPER version 21.0) discussed in section II. of this final rule.

2. Changes in the LTC-DRG Classifications

a. Background

Section 123 of Pub. L. 106-113 specifically requires that the PPS for LTCHs be a per discharge system with a DRG-based patient classification system reflecting the differences in patient resources and costs in LTCHs while maintaining budget neutrality. Section 307(b)(1) of Pub. L. 106-554 modified the requirements of section 123 of Pub. L. 106-113 by specifically requiring that the Secretary examine “the feasibility and the impact of basing payment under such a system [the LTCH PPS] on the use of existing (or refined) hospital diagnosis-related groups (DRGs) that have been modified to account for different resource use of long-term care hospital patients as well as the use of the most recently available hospital discharge data.”

In accordance with section 307(b)(1) of Pub. L. 106-554 and § 412.515 of our existing regulations, the LTCH PPS uses information from LTCH patient records to classify patient cases into distinct LTC-DRGs based on clinical characteristics and expected resource needs. The LTC-DRGs used as the patient classification component of the LTCH PPS correspond to the DRGs under the IPPS for acute care hospitals. Thus, under this final rule, we will use the IPPS version 21.0 GROUPER for FY 2004 to process LTCH PPS claims. The changes to the IPPS DRG classification system for FY 2004 (Grouper 21.0) are discussed in section II.B. of this preamble.

Under the LTCH PPS, we determine relative weights for each of the IPPS DRGs to account for the difference in resource use by patients exhibiting the case complexity and multiple medical problems characteristic of LTCH patients. In a departure from the IPPS, as we discussed in both the May 19, 2003 proposed rule (68 FR 27174) and the June 6, 2003 LTCH PPS final rule (68 FR 34132), we use low volume quintiles in determining the LTC-DRG weights for LTC-DRGs with less than 25 LTCH cases, since LTCHs do not typically treat the full range of diagnoses as do acute care hospitals. In order to deal with the large number of low volume LTC-DRGs (LTC-DRGs with fewer than 25 cases), as we discussed in the May 19, 2003 proposed rule (68 FR 27176), we group those low volume LTC-DRGs into 5 quintiles based on average charge per discharge. (A listing of the composition of low volume quintiles for the FY 2004 LTC-DRGs (based on FY 2002 MedPAR data) appears in section II.D.3. of this final rule.) We also adjust for cases in which the stay at the LTCH is less than or equal to five-sixths of the geometric average length of stay; that is, short-stay outlier cases (§ 412.529), as discussed in section II.D.4. of this preamble.

b. Patient Classifications Into DRGs

Generally, under the LTCH PPS, Medicare payment is made at a predetermined specific rate for each discharge; that is, payment varies by the LTC-DRG to which a beneficiary's stay is assigned. Similar to case classification for acute care hospitals under the IPPS (see section II.B. of this preamble), cases are classified into LTC-DRGs for payment under the LTCH PPS based on the principal diagnosis, up to eight additional diagnoses, and up to six procedures performed during the stay, as well as age, sex, and discharge status of the patient. The diagnosis and procedure information is reported by the hospital using codes from the ICD-9-CM.

As discussed above in section II.B. of this preamble, the DRGs are organized into 25 major diagnostic categories (MDCs), most of which are based on a particular organ system of the body; the remainder involve multiple organ systems (such as MDC 22, Burns). Accordingly, the principal diagnosis determines MDC assignment. Within most MDCs, cases are then divided into surgical DRGs and medical DRGs. Some surgical and medical DRGs are further differentiated based on the presence or absence of CCs. (See section II.B. of this preamble for further discussion of surgical DRGs and medical DRGs.)

Because the assignment of a case to a particular LTC-DRG will help determine the amount that is paid for the case, it is important that the coding is accurate. As used under the IPPS, classifications and terminology used under the LTCH PPS are consistent with the ICD-9-CM and the Uniform Hospital Discharge Data Set (UHDDS), as recommended to the Secretary by the National Committee on Vital and Health Statistics (“Uniform Hospital Discharge Data: Minimum Data Set, National Center for Health Statistics, April 1980”) and as revised in 1984 by the Health Information Policy Council (HIPC) of the U.S. Department of Health and Human Services. We wish to point out again that the ICD-9-CM coding terminology and the definitions of principal and other diagnoses of the UHDDS are consistent with the requirements of the Administrative Simplification Act of 1996 of the HIPAA (45 CFR Parts 160 and 162).

The emphasis on the need for proper coding cannot be overstated. Inappropriate coding of cases can adversely affect the uniformity of cases in each LTC-DRG and produce inappropriate weighting factors at recalibration and result in inappropriate payments under the LTCH PPS. LTCHs are to follow the same coding guidelines used by the acute care hospitals to ensure accuracy and consistency in coding practices. There will be only one LTC-DRG assigned per long-term care hospitalization; it will be assigned at the discharge. Therefore, it is mandatory that the coders continue to report the same principal diagnosis on all claims and include all diagnostic codes that coexist at the time of admission, that are subsequently developed, or that affect the treatment received. Similarly, all procedures performed during that stay are to be reported on each claim.

Upon the discharge of the patient from a LTCH, the LTCH must assign appropriate diagnosis and procedure codes from the ICD-9-CM. As of October 16, 2002, a LTCH that was required to comply with the HIPAA Administrative Simplification Standards and that had not obtained an extension in compliance with the Administrative Compliance Act (Pub. L. 107-105) is obligated to comply with the standards at 45 CFR 162.1002 and 45 CFR 162.1102. Completed claim forms are to be submitted to the LTCH's Medicare fiscal intermediary.

Medicare fiscal intermediaries enter the clinical and demographic information into their claims processing systems and subject this information to a series of automated screening processes called the Medicare Code Editor (MCE). These screens are designed to identify cases that require further review before assignment into a LTC-DRG can be made.

After screening through the MCE, each LTCH claim will be classified into the appropriate LTC-DRG by the Medicare LTCH GROUPER. The LTCH GROUPER is specialized computer software based on the same GROUPER used under the IPPS. After the LTC-DRG is assigned, the Medicare fiscal intermediary determines the prospective payment by using the Medicare PRICER program, which accounts for LTCH hospital-specific adjustments. As provided for under the IPPS, we provide an opportunity for the LTCH to review the LTC-DRG assignments made by the fiscal intermediary and to submit additional information within a specified timeframe (§ 412.513(c)).

The GROUPER is used both to classify past cases in order to measure relative hospital resource consumption to establish the LTC-DRG weights and to classify current cases for purposes of determining payment. The records for all Medicare hospital inpatient discharges are maintained in the MedPAR file. The data in this file are used to evaluate possible DRG classification changes and to recalibrate the DRG weights during our annual update (as discussed in section II. of this preamble). The LTC-DRG weights are based on data for the population of LTCH discharges, reflecting the fact that LTCH patients represent a different patient mix than patients in short-term acute care hospitals.

3. Development of the FY 2004 LTC-DRG Relative Weights

a. General Overview of Development of the LTC-DRG Relative Weights

As we stated in the August 30, 2002 LTCH PPS final rule (67 FR 55981), one of the primary goals for the implementation of the LTCH PPS is to pay each LTCH an appropriate amount for the efficient delivery of care to Medicare patients. The system must be able to account adequately for each LTCH's case-mix in order to ensure both fair distribution of Medicare payments and access to adequate care for those Medicare patients whose care is more costly. To accomplish these goals, we adjust the LTCH PPS standard Federal prospective payment system rate by the LTC-DRG relative weights in determining payment to LTCHs for each case.

Under the LTCH PPS, relative weights for each LTC-DRG are a primary element used to account for the variations in cost per discharge and resource utilization among the payment groups (§ 412.515). To ensure that Medicare patients classified to each LTC-DRG have access to an appropriate level of services and to encourage efficiency, we calculate a relative weight for each LTC-DRG that represents the resources needed by an average inpatient LTCH case in that LTC-DRG. For example, cases in a LTC-DRG with a relative weight of 2 will, on average, cost twice as much as cases in a LTC-DRG with a weight of 1.

b. Data

To calculate the LTC-DRG relative weights for FY 2004 in this final rule, we obtained total Medicare allowable charges from FY 2002 Medicare hospital bill data from the December 2002 update of the MedPAR file, and we used Version 21.0 of the CMS GROUPER for IPPS, as discussed in section II.B. of this preamble, to classify cases. Consistent with the methodology under the IPPS, we recalculated the FY 2004 LTC-DRG relative weights based on the best available data for this final rule.

As we discussed in the May 19, 2003 proposed rule (68 FR 27151), we have excluded the data from LTCHs that are all-inclusive rate providers and LTCHs that are reimbursed in accordance with demonstration projects authorized under section 402(a) of Pub. L. 90-248 (42 U.S.C. 1395b-1) or section 222(a) of Pub. L. 92-603 (42 U.S.C. 1395b-1). Therefore, in the development of the FY 2004 LTC-DRG relative weights, we have excluded the data of the 22 all-inclusive rate providers and the 3 LTCHs that are paid in accordance with demonstration projects.

In addition, as we discussed in that same proposed rule, a data problem regarding the proposed FY 2003 LTC-DRG relative weight values that were determined using MedPAR (claims) data for FYs 2000 and 2001 was brought to our attention. Following notification of this problem, we researched the commenter's claims and determined that, given the long stays at LTCHs, some providers had submitted multiple bills for payment under the reasonable cost-based reimbursement system for the same stay. Based upon our research, we became aware of the following situation: In certain LTCHs, hospital personnel apparently reported a different principal diagnosis on each bill since, under the reasonable cost-based reimbursement system, payment was not dependent upon principal diagnosis, as it is under a DRG-based system. These claims from the MedPAR file were run through the LTCH GROUPER and used in determining the proposed FY 2003 relative weights for each LTC-DRG.

After this issue was brought to our attention, we discovered that only data from the final bills were being extracted for the MedPAR file. Therefore, it was possible that the original MedPAR file was not receiving the correct principal diagnosis. In the August 30, 2002 final rule (67 FR 55989), we addressed the problem by identifying all LTCH cases in the FY 2001 MedPAR file for which multiple bills were submitted. For each of these cases, beginning with the first bill and moving forward consecutively through subsequent bills for that stay, we recorded the first unique diagnosis codes up to 10 and the first unique procedure codes up to 10. We then used these codes to appropriately group each LTCH case to a LTC-DRG for FY 2003.

As we noted above, we are using LTCH claims data from the FY 2002 MedPAR file for the determination of the FY 2004 LTC-DRG relative weights. Since at the time (FY 2002) LTCHs were still reimbursed under the reasonable cost-based system, some LTCHs also had submitted multiple bills for Medicare payment for the same stay. Thus, in certain LTCHs, hospital personnel were apparently still reporting a different principal diagnosis on each bill since, under the reasonable cost-based reimbursement system in FY 2002, payment was not dependent upon principal diagnosis as it is under a DRG-based system. Therefore, as we explained in the May 19, 2003 proposed rule (68 FR 27151), we are following the same methodology outlined above to determine the appropriate diagnosis and procedure codes for those multiple bill LTCH cases in the FY 2002 MedPAR files, and we are using these codes to group each LTCH case to a LTC-DRG for FY 2004. Since the LTCH PPS was implemented for cost reporting periods beginning on or after October 1, 2002 (FY 2003), we believe that this problem will be self-correcting as LTCHs submit more completely coded data in the future.

c. Hospital-Specific Relative Value Methodology

By nature LTCHs often specialize in certain areas, such as ventilator-dependent patients and rehabilitation and wound care. Some case types (DRGs) may be treated, to a large extent, in hospitals that have, from a perspective of charges, relatively high (or low) charges. Such nonarbitrary distribution of cases with relatively high (or low) charges in specific LTC-DRGs has the potential to inappropriately distort the measure of average charges. To account for the fact that cases may not be randomly distributed across LTCHs, we use a hospital-specific relative value method to calculate the LTC-DRG relative weights instead of the methodology used to determine the DRG relative weights under the IPPS described above in section II.C. of this preamble. We believe this method will remove this hospital-specific source of bias in measuring LTCH average charges. Specifically, we reduce the impact of the variation in charges across providers on any particular LTC-DRG relative weight by converting each LTCH's charge for a case to a relative value based on that LTCH's average charge.

Under the hospital-specific relative value method, we standardize charges for each LTCH by converting its charges for each case to hospital-specific relative charge values and then adjusting those values for the LTCH's case-mix. The adjustment for case-mix is needed to rescale the hospital-specific relative charge values (which, by definition, averages 1.0 for each LTCH). The average relative weight for a LTCH is its case-mix, so it is reasonable to scale each LTCH's average relative charge value by its case-mix. In this way, each LTCH's relative charge value is adjusted by its case-mix to an average that reflects the complexity of the cases it treats relative to the complexity of the cases treated by all other LTCHs (the average case-mix of all LTCHs).

In accordance with the methodology established under § 412.523, we standardize charges for each case by first dividing the adjusted charge for the case (adjusted for short-stay outliers under § 412.529 as described in section II.D.4. (step 3) of this preamble) by the average adjusted charge for all cases at the LTCH in which the case was treated. Short-stay outliers under § 412.529 are cases with a length of stay that is less than or equal to five-sixths the average length of stay of the LTC-DRG. The average adjusted charge reflects the average intensity of the health care services delivered by a particular LTCH and the average cost level of that LTCH. The resulting ratio is multiplied by that LTCH's case-mix index to determine the standardized charge for the case.

Multiplying by the LTCH's case-mix index accounts for the fact that the same relative charges are given greater weight in a LTCH with higher average costs than they would at a LTCH with low average costs which is needed to adjust each LTCH's relative charge value to reflect its case-mix relative to the average case-mix for all LTCHs. Because we standardize charges in this manner, we count charges for a Medicare patient at a LTCH with high average charges as less resource intensive than they would be at a LTCH with low average charges. For example, a $10,000 charge for a case in a LTCH with an average adjusted charge of $17,500 reflects a higher level of relative resource use than a $10,000 charge for a case in a LTCH with the same case-mix, but an average adjusted charge of $35,000. We believe that the adjusted charge of an individual case more accurately reflects actual resource use for an individual LTCH because the variation in charges due to systematic differences in the markup of charges among LTCHs is taken into account.

d. Low Volume LTC-DRGs

In order to account for LTC-DRGs with low volume (that is, with fewer than 25 LTCH cases), in accordance with the methodology discussed in the May 19, 2003 proposed rule (68 FR 27176), we group those low volume LTC-DRGs into one of five categories (quintiles) based on average charges, for the purposes of determining relative weights. For this final rule, using LTCH cases from the FY 2002 MedPAR file, we identified 173 LTC-DRGs that contained between 1 and 24 cases. This list of LTC-DRGs was then divided into one of the five low volume quintiles, each containing a minimum of 34 LTC-DRGs (173/5 = 34 with 3 LTC-DRGs as the remainder). For FY 2004, as we described in that same proposed rule, we are making an assignment to a specific low volume quintile by sorting the 173 low volume LTC-DRGs in ascending order by average charge. Since the number of LTC-DRGs with less than 25 LTCH cases is not evenly divisible by five, the average charge of the low volume LTC-DRG was used to determine which low volume quintile received the additional LTC-DRG. After sorting the 173 low volume LTC-DRGs in ascending order, we grouped the first fifth (34) of low volume LTC-DRGs with the lowest average charge into Quintile 1. The highest average charge cases are grouped into Quintile 5. Since the average charge of the 69th LTC-DRG in the sorted list is closer to the previous LTC-DRG's average charge (assigned to Quintile 2) than to the average charge of the 70th LTC-DRG in the sorted list (to be assigned to Quintile 3), we placed it into Quintile 2. This process was repeated through the remaining low volume LTC-DRGs so that 3 low volume quintiles contain 35 LTC-DRGs and 2 low volume quintiles contain 34 LTC-DRGs.

In order to determine the relative weights for the LTC-DRGs with low volume for FY 2004, in accordance with the methodology described in the May 19, 2003 proposed rule (68 FR 27176), we used the five low volume quintiles described above. The composition of each of the five low volume quintiles shown below in Table 1 is used in determining the LTC-DRG relative weights for FY 2004. We determine a relative weight and (geometric) average length of stay for each of the five low volume quintiles using the formula that we apply to the regular LTC-DRGs (25 or more cases), as described below in section II.D.4. of this preamble. We assign the same relative weight and average length of stay to each of the LTC-DRGs that make up that low volume quintile. We note that as this system is dynamic, it is possible that the number and specific type of LTC-DRGs with a low volume of LTCH cases will vary in the future. We use the best available claims data in the MedPAR file to identify low volume LTC-DRGs and to calculate the relative weights based on our methodology.

4. Steps for Determining the FY 2004 LTC-DRG Relative Weights

As we noted previously, the FY 2004 LTC-DRG relative weights are determined in accordance with the methodology described in the May 19, 2003 proposed rule (68 FR 27179). In summary, LTCH cases must be grouped in the appropriate LTC-DRG, while taking into account the low volume LTC-DRGs as described above, before the FY 2004 LTC-DRG relative weights can be determined. After grouping the cases in the appropriate LTC-DRG, we calculate the relative weights for FY 2004 in this final rule by first removing statistical outliers and cases with a length of stay of 7 days or less. Next, we adjust the number of cases in each LTC-DRG for the effect of short-stay outlier cases under § 412.529. The short-stay adjusted discharges and corresponding charges are used to calculate “relative adjusted weights” in each LTC-DRG using the hospital-specific relative value method described above.

Below we discuss in detail the steps for calculating the FY 2004 LTC-DRG relative weights.

Step 1—Remove Statistical Outliers

The first step in the calculation of the FY 2004 LTC-DRG relative weights is to remove statistical outlier cases. As we discussed in the May 19, 2003 proposed rule (68 FR 27179), we define statistical outliers as cases that are outside of 3.0 standard deviations from the mean of the log distribution of both charges per case and the charges per day for each LTC-DRG. These statistical outliers are removed prior to calculating the relative weights. We believe that they may represent aberrations in the data that distort the measure of average resource use. Including those LTCH cases in the calculation of the relative weights could result in an inaccurate relative weight that does not truly reflect relative resource use among the LTC-DRGs.

Step 2—Remove Cases With a Length of Stay of 7 Days or Less

The FY 2004 LTC-DRG relative weights reflect the average of resources used on representative cases of a specific type. Generally, as we discussed in the May 19, 2003 proposed rule (68 FR 27179), cases with a length of stay 7 days or less do not belong in a LTCH because such stays do not fully receive or benefit from treatment that is typical in a LTCH stay, and full resources are often not used in the earlier stages of admission to a LTCH. If we were to include stays of 7 days or less in the computation of the FY 2004 LTC-DRG relative weights, the value of many relative weights would decrease and, therefore, payments would decrease to a level that may no longer be appropriate.

We do not believe that it would be appropriate to compromise the integrity of the payment determination for those LTCH cases that actually benefit from and receive a full course of treatment at a LTCH, in order to include data from these very short-stays. Thus, in determining the FY 2004 LTC-DRG relative weights, we remove LTCH cases with a length of stay of 7 days or less.

Step 3—Adjust Charges for the Effects of Short-Stay Outliers

The third step in the calculation of the FY 2004 LTC-DRG relative weights is to adjust each LTCH's charges per discharge for short-stay outlier cases (that is, a patient with a length of stay that is less than or equal to five-sixths the average length of stay of the LTC-DRG).

As we discussed in the May 19, 2003 proposed rule (68 FR 27179), we make this adjustment by counting a short-stay outlier as a fraction of a discharge based on the ratio of the length of stay of the case to the average length of stay for the LTC-DRG for nonshort-stay outlier cases. This has the effect of proportionately reducing the impact of the lower charges for the short-stay outlier cases in calculating the average charge for the LTC-DRG. This process produces the same result as if the actual charges per discharge of a short-stay outlier case were adjusted to what they would have been had the patient's length of stay been equal to the average length of stay of the LTC-DRG.

As we explained in that same proposed rule, counting short-stay outlier cases as full discharges with no adjustment in determining the LTC-DRG relative weights would lower the LTC-DRG relative weight for affected LTC-DRGs because the relatively lower charges of the short-stay outlier cases would bring down the average charge for all cases within a LTC-DRG. This would result in an “underpayment” to nonshort-stay outlier cases and an “overpayment” to short-stay outlier cases. Therefore, in this final rule, we adjust for short-stay outlier cases under § 412.529 in this manner since it results in more appropriate payments for all LTCH cases.

Step 4—Calculate the FY 2004 LTC-DRG Relative Weights on an Iterative Basis

As we discussed in the May 19, 2003 proposed rule (68 FR 27180), the process of calculating the LTC-DRG relative weights using the hospital specific relative value methodology is iterative. First, for each LTCH case, we calculate a hospital-specific relative charge value by dividing the short-stay outlier adjusted charge per discharge (see step 3) of the LTCH case (after removing the statistical outliers (see step 1)) and LTCH cases with a length of stay of 7 days or less (see step 2) by the average charge per discharge for the LTCH in which the case occurred. The resulting ratio is then multiplied by the LTCH's case-mix index to produce an adjusted hospital-specific relative charge value for the case. An initial case-mix index value of 1.0 is used for each LTCH.

For each LTC-DRG, the FY 2004 LTC-DRG relative weight is calculated by dividing the average of the adjusted hospital-specific relative charge values (from above) for the LTC-DRG by the overall average hospital-specific relative charge value across all cases for all LTCHs. Using these recalculated LTC-DRG relative weights, each LTCH's average relative weight for all of its cases (case-mix) is calculated by dividing the sum of all the LTCH's LTC-DRG relative weights by its total number of cases. The LTCHs' hospital-specific relative charge values above are multiplied by these hospital specific case-mix indexes. These hospital-specific case-mix adjusted relative charge values are then used to calculate a new set of LTC-DRG relative weights across all LTCHs. In this final rule, this iterative process is continued until there is convergence between the weights produced at adjacent steps, for example, when the maximum difference is less than 0.0001.

Step 5—Adjust the FY 2004 LTC-DRG Relative Weights to Account for Nonmonotonically Increasing Relative Weights

As explained in section II.B. of this preamble, the FY 2004 CMS DRGs, upon which the FY 2004 LTC-DRGs are based, contain “pairs” that are differentiated based on the presence or absence of CCs. The LTC-DRGs with CCs are defined by certain secondary diagnoses not related to or inherently a part of the disease process identified by the principal diagnosis, but the presence of additional diagnoses does not automatically generate a CC. As we discussed in the May 19, 2003 proposed rule (68 FR 27180), the value of monotonically increasing relative weights rises as the resource use increases (for example, from uncomplicated to more complicated). The presence of CCs in a LTC-DRG means that cases classified into a “without CC” LTC-DRG are expected to have lower resource use (and lower costs). In other words, resource use (and costs) are expected to decrease across “with CC”/“without CC” pairs of LTC-DRGs.

For a case to be assigned to a LTC-DRG with CCs, more coded information is called for (that is, at least one relevant secondary diagnosis), than for a case to be assigned to a LTC-DRG “without CCs” (which is based on only one principal diagnosis and no relevant secondary diagnoses). Currently, the LTCH claims data include both accurately coded cases without complications and cases that have complications (and cost more) but were not coded completely. Both types of cases are grouped to a LTC-DRG “without CCs” since only one principal diagnosis was coded. Since LTCHs were previously paid under cost-based reimbursement, which is not based on patient diagnoses, coding by LTCHs for these cases may not have been as detailed as possible.

Thus, in developing the FY 2003 LTC-DRG relative weights for the LTCH PPS based on FY 2001 claims data, as we discussed in the August 30, 2002 LTCH PPS final rule (67 FR 55990), we found on occasion that the data suggested that cases classified to the LTC-DRG “with CCs” of a “with CC”/“without CC” pair had a lower average charge than the corresponding LTC-DRG “without CCs.” Similarly, based on FY 2002 claims data, we also found on occasion that the data suggested that cases classified to the LTC-DRG “with CCs” of a “with CC”/“without CC” pair have a lower average charge than the corresponding LTC-DRG “without CCs” for FY 2004.

We believe this anomaly may be due to coding that may not have fully reflected all comorbidities that were present. Specifically, LTCHs may have failed to code relevant secondary diagnoses, which resulted in cases that actually had CCs being classified into a “without CC” LTC-DRG. It would not be appropriate to pay a lower amount for the “with CC” LTC-DRG. Therefore, as we discussed in the May 19, 2003 proposed rule (68 FR 27180), we grouped both the cases “with CCs” and “without CCs” together for the purpose of calculating the FY 2004 LTC-DRG relative weights in this final rule. We continue to employ this methodology to account for nonmonotonically increasing relative weights until we have adequate data to calculate appropriate separate weights for these anomalous LTC-DRG pairs. We expect that, as was the case when we first implemented the IPPS, this problem will be self-correcting, as LTCHs submit more completely coded data in the future.

There are three types of “with CC” and “without CC” pairs that could be nonmonotonic, that is, where the “without CC” LTC-DRG would have a higher average charge than the “with CC” LTC-DRG. For this final rule, using the LTCH cases in the December 2002 update of the FY 2002 MedPAR file, we identified three of the types of nonmonotonic LTC-DRG pairs.

The first category of nonmonotonically increasing relative weights for FY 2004 LTC-DRG pairs “with and without CCs” contains 1 pair of LTC-DRGs in which both the LTC-DRG “with CCs” and the LTC-DRG “without CCs” had 25 or more LTCH cases and, therefore, did not fall into one of the 5 low volume quintiles. For that type of nonmonotonic LTC-DRG pair, as discussed in the May 19, 2003 proposed rule (68 FR 27180), we combine the LTCH cases and compute a new relative weight based on the case-weighted average of the combined LTCH cases of the LTC-DRGs. The case-weighted average charge is determined by dividing the total charges for all LTCH cases by the total number of LTCH cases for the combined LTC-DRG. This new relative weight is then assigned to both of the LTC-DRGs in the pair. In this final rule, for FY 2004, LTC-DRGs 180 and 181 are in this category.

The second category of nonmonotonically increasing relative weights for LTC-DRG pairs with and without CCs consists of 7 pairs of LTC-DRGs that has fewer than 25 cases, and each LTC-DRG is grouped to different low volume quintiles in which the “without CC” LTC-DRG is in a higher-weighted low volume quintile than the “with CC” LTC-DRG. For those pairs, as we discussed in the May 19, 2003 proposed rule (68 FR 27181), we combine the LTCH cases and determine the case-weighted average charge for all LTCH cases. The case-weighted average charge is determined by dividing the total charges for all LTCH cases by the total number of LTCH cases for the combined LTC-DRG. Based on the case-weighted average LTCH charge, we determine which low volume quintile the “combined LTC-DRG” is grouped. Both LTC-DRGs in the pair are then grouped into the same low volume quintile, and thus would have the same relative weight. For FY 2004, in this final rule, the following LTC-DRGs are in this category: LTC-DRGs 31 and 32 (low volume quintile 3); LTC-DRGs 193 and 194 (low volume quintile 2); LTC-DRGs 226 and 227 (low volume quintile 4); LTC-DRGs 449 and 450 (low volume quintile 3); LTC-DRGs 493 and 494 (low volume quintile 4); LTC-DRGs 497 and 498 (low volume quintile 3); and LTC-DRGs 506 and 507 (low volume quintile 2).

The third category of nonmonotonically increasing relative weights for LTC-DRG pairs with and without CCs consists of 6 pairs of LTC-DRGs where one of the LTC-DRGs has fewer than 25 LTCH cases and is grouped to a low volume quintile and the other LTC-DRG has 25 or more LTCH cases and has its own LTC-DRG relative weight, and the LTC-DRG “without CCs” has the higher relative weight. As we discussed in the May 19, 2003 proposed rule (68 FR 27181), we remove the low volume LTC-DRG from the low volume quintile and combine it with the other LTC-DRG for the computation of a new relative weight for each of these LTC-DRGs. This new relative weight is assigned to both LTC-DRGs, so they each have the same relative weight. For FY 2004, in this final rule, the following LTC-DRGs are in this category: LTC-DRGs 7 and 8; LTC-DRGs 205 and 206; LTC-DRGs 207 and 208; LTC-DRGs 265 and 266; LTC-DRGs 346 and 347; and LTC-DRGs 478 and 479.

Step 6—Determine a FY 2004 LTC-DRG Relative Weight for LTC-DRGs With No LTCH Cases

As we stated above, we determine the relative weight for each LTC-DRG using charges reported in the December 2002 update of the FY 2002 MedPAR file. Of the 518 LTC-DRGs for FY 2004, we identified 167 LTC-DRGs for which there were no LTCH cases in the database. That is, based on data from the FY 2002 MedPAR file used in this final rule, no patients who would have been classified to those LTC-DRGs were treated in LTCHs during FY 2002 and, therefore, no charge data were reported for those LTC-DRGs. Thus, in the process of determining the LTC-DRG relative weights, we are unable to determine weights for these 167 LTC-DRGs using the methodology described in steps 1 through 5 above. However, since patients with a number of the diagnoses under these LTC-DRGs may be treated at LTCHs beginning in FY 2004, we assign relative weights to each of the 167 “no volume” LTC-DRGs based on clinical similarity and relative costliness to one of the remaining 354 (518−167 = 351) LTC-DRGs for which we are able to determine relative weights, based on FY 2002 claims data.

As there are currently no LTCH cases in these “no volume” LTC-DRGs, as we discussed in the May 19, 2003 proposed rule (68 FR 27181), we determine relative weights for the 167 LTC-DRGs with no LTCH cases in the FY 2002 MedPAR file used in this final rule by grouping them to the appropriate low volume quintile. This methodology is consistent with our methodology used in determining relative weights to account for the low volume LTC-DRGs described above.

Our methodology for determining relative weights for the “no volume” LTC-DRGs is as follows: First, we crosswalk the no volume LTC-DRGs by matching them to other similar LTC-DRGs for which there were LTCH cases in the FY 2002 MedPAR file based on clinical similarity and intensity of use of resources as determined by care provided during the period of time surrounding surgery, surgical approach (if applicable), length of time of surgical procedure, post-operative care, and length of stay. We assign the relative weight for the applicable low volume quintile to the no volume LTC-DRG if the LTC-DRG to which it is crosswalked is grouped to one of the low volume quintiles. If the LTC-DRG to which the no volume LTC-DRG is crosswalked is not one of the LTC-DRGs to be grouped to one of the low volume quintiles, we compare the relative weight of the LTC-DRG to which the no volume LTC-DRG is crosswalked to the relative weights of each of the five quintiles and we assign the no volume LTC-DRG the relative weight of the low volume quintile with the closest weight. For this final rule, a list of the no volume FY 2004 LTC-DRGs and the FY 2004 LTC-DRG to which it is crosswalked in order to determine the appropriate low volume quintile for the assignment of a relative weight for FY 2004 is shown below in Table 2.

To illustrate this methodology for determining the relative weights for the 164 LTC-DRGs with no LTCH cases, we are providing the following examples, which refer to the no volume LTC-DRGs crosswalk information for FY 2004 provided above in Table 2:

Example 1: There were no cases in the FY 2002 MedPAR file used for this final rule for LTC-DRG 163 (Hernia Procedures Age 0-17). Since the procedure is similar in resource use and the length and complexity of the procedures and the length of stay are similar, we determined that LTC-DRG 178 (Uncomplicated Peptic Ulcer Without CC), which is assigned to low volume quintile 1 for the purpose of determining the FY 2004 relative weights, would display similar clinical and resource use. Therefore, we assign the same relative weight of LTC-DRG 178 of 0.4964 (Quintile 1) for FY 2004 (Table 11 in the Addendum to this final rule) to LTC-DRG 163.

Example 2: There were no LTCH cases in the FY 2002 MedPAR file used in this final rule for LTC-DRG 91 (Simple Pneumonia and Pleurisy Age 0-17). Since the severity of illness in patients with bronchitis and asthma is similar in patients regardless of age, we determined that LTC-DRG 90 (Simple Pneumonia and Pleurisy Age >17 Without CC) would display similar clinical and resource use characteristics and have a similar length of stay to LTC-DRG 91. There were over 25 cases in LTC-DRG 90. Therefore, it would not be assigned to a low volume quintile for the purpose of determining the LTC-DRG relative weights. However, under our established methodology, LTC-DRG 91, with no LTCH cases, would need to be grouped to a low volume quintile. We identified that the low volume quintile with the closest weight to LTC-DRG 90 (0.7318; see Table 11 in the Addendum to this final rule) would be low volume quintile 2 (0.7372; see Table 11 in the Addendum to this final rule). Therefore, we assign LTC-DRG 91 a relative weight of 0.7372 for FY 2004.

Furthermore, we are providing LTC-DRG relative weights of 0.0000 for heart, kidney, liver, lung, pancreas, and simultaneous pancreas/kidney transplants (LTC-DRGs 103, 302, 480, 495, 512, and 513, respectively) for FY 2004 because Medicare will only cover these procedures if they are performed at a hospital that has been certified for the specific procedures by Medicare and presently no LTCH has been so certified.

Based on our research, we found that most LTCHs only perform minor surgeries, such as minor small and large bowel procedures, to the extent any surgeries are performed at all. Given the extensive criteria that must be met to become certified as a transplant center for Medicare, we believe it is unlikely that any LTCHs would become certified as a transplant center. In fact, in the nearly 20 years since the implementation of the IPPS, there has never been a LTCH that even expressed an interest in becoming a transplant center.

However, if in the future a LTCH applies for certification as a Medicare-approved transplant center, we believe that the application and approval procedure would allow sufficient time for us to determine appropriate weights for the LTC-DRGs affected. At the present time, we are only including these six transplant LTC-DRGs in the GROUPER program for administrative purposes. Since we use the same GROUPER program for LTCHs as is used under the IPPS, removing these LTC-DRGs would be administratively burdensome.

Again, we note that as this system is dynamic, it is entirely possible that the number of LTC-DRGs with a zero volume of LTCH cases based on the system will vary in the future. We used the best most recent available claims data in the MedPAR file to identify zero volume LTC-DRGs and to determine the relative weights in this final rule.

Table 11 in the Addendum to this final rule lists the LTC-DRGs and their respective relative weights, geometric mean length of stay, and five-sixths of the geometric mean length of stay (to assist in the determination of short-stay outlier payments under § 412.529) for FY 2004.

E. Add-On Payments for New Services and Technologies

1. Background

Sections 1886(d)(5)(K) and (L) of the Act establish a process of identifying and ensuring adequate payment for new medical services and technologies under the IPPS. Section 1886(d)(5)(K)(ii)(I) of the Act specifies that the process must apply to a new medical service or technology if, “based on the estimated costs incurred with respect to discharges involving such service or technology, the DRG prospective payment rate otherwise applicable to such discharges under this subsection is inadequate.” Section 1886(d)(5)(K)(vi) of the Act specifies that a medical service or technology will be considered “new” if it meets criteria established by the Secretary after notice and opportunity for public comment.

Section 412.87(b)(1) of our existing regulations provides that a new technology will be an appropriate candidate for an additional payment when it represents an advance in medical technology that substantially improves, relative to technologies previously available, the diagnosis or treatment of Medicare beneficiaries (see the September 7, 2001 final rule (66 FR 46902)). Section 412.87(b)(3) provides that, to receive special payment treatment, new technologies meeting this clinical definition must be demonstrated to be inadequately paid otherwise under the DRG system. As discussed below, for applicants for new technology add-on payments for FY 2005, we are establishing the criteria that will be applied to assess whether technologies would be inadequately paid under the DRGs 75 percent of 1 standard deviation (based on the logarithmic values of the charges and transformed back to charges) beyond the geometric mean standardized charge for all cases in the DRGs to which the new technology is assigned (or the case-weighted average of all relevant DRGs, if the new technology occurs in many different DRGs). Table 10 in the Addendum to this final rule lists the qualifying criteria by DRG, based on the discharge data that we used to calculate the FY 2004 DRG weights. The thresholds that are published in this final rule for FY 2004 will be used to evaluate applicants for new technology add-on payments during FY 2005.

In addition to the clinical and cost criteria, we established that, in order to qualify for the new technology add-on payments, a specific technology must be “new” under the requirements of § 412.87(b)(2) of our regulations. The statutory provision contemplated the special payment treatment for new technologies until such time as data are available to reflect the cost of the technology in the DRG weights through recalibration (no less than 2 years and no more than 3 years). There is a lag of 2 to 3 years from the point a new technology is first introduced on the market and when data reflecting the use of the technology are used to calculate the DRG weights. For example, data from discharges occurring during FY 2002 are used to calculate the FY 2004 DRG weights in this final rule.

Technology may be considered “new” for purposes of this provision within 2 or 3 years after the point at which data begin to become available reflecting the costs of the technology. After we have recalibrated the DRGs to reflect the costs of an otherwise new technology, the special add-on payment for new technology will cease (§ 412.87(b)(2)). For example, an approved new technology that received FDA approval in October 2002 would be eligible to receive add-on payments as a new technology at least until FY 2005 (discharges occurring before October 1, 2004), when data reflecting the costs of the technology would be used to recalibrate the DRG weights. Because the FY 2005 DRG weights will be calculated using FY 2003 MedPAR data, the costs of such a new technology would likely be reflected in the FY 2005 DRG weights.

Similar to the timetable for applying for new technology add-on payments during FY 2004, applicants for FY 2005 must submit a formal request, including a full description of the clinical applications of the technology and the results of any clinical evaluations demonstrating that the new technology represents a substantial clinical improvement, along with a significant sample of data to demonstrate the technology meets the high-cost threshold, no later than early October 2003. Applicants must submit a complete database no later than mid-December 2003. Complete application information is available at our Web site at: http://www.cms.hhs.gov/providers/hipps/default.asp. To allow interested parties to identify the technologies under review before the publication of the annual proposed rule, the Web site also lists the tracking forms completed by each applicant.

The new technology add-on payment policy provides additional payments for cases with high costs involving eligible new technologies while preserving some of the incentives under the average-based payment system. The payment mechanism is based on the cost to hospitals for the new technology. Under § 412.88, Medicare pays a marginal cost factor of 50 percent for the costs of the new technology in excess of the full DRG payment. If the actual costs of a new technology case exceed the DRG payment by more than the estimated costs of the new technology, Medicare payment is limited to the DRG payment plus 50 percent of the estimated costs of the new technology.

The report language accompanying section 533 of Pub. L. 106-554 indicated Congressional intent that the Secretary implement the new mechanism on a budget neutral basis (H.R. Conf. Rep. No. 106-1033, 106th Cong., 2nd Sess. at 897 (2000)). Section 1886(d)(4)(C)(iii) of the Act requires that the adjustments to annual DRG classifications and relative weights must be made in a manner that ensures that aggregate payments to hospitals are not affected. Therefore, we account for projected payments under the new technology provision during the upcoming fiscal year at the same time we estimate the payment effect of changes to the DRG classifications and recalibration. The impact of additional payments under this provision would then be included in the budget neutrality factor, which is applied to the standardized amounts and the hospital-specific amounts.

Because any additional payments directed toward new technology under this provision must be offset to ensure budget neutrality, it is important to consider carefully the extent of this provision and ensure that only technologies representing substantial advances are recognized for additional payments. In that regard, we indicated that we would discuss in the annual proposed and final rules those technologies that were considered under this provision; our determination as to whether a particular technology meets our criteria to be considered new; whether it is determined further that cases involving the new technology would be inadequately paid under the existing DRG payment; and any assumptions that went into the budget neutrality calculations related to additional payments for that new technology, including the expected number, distribution, and costs of these cases.

To balance appropriately the Congress' intent to increase Medicare's payments for eligible new technologies with concern that the total size of those payments not result in significantly reduced payments for other cases, we set a target limit for estimated add-on payments for new technology under the provisions of sections 1886(d)(5)(K) and (L) of the Act at 1.0 percent of estimated total operating prospective payments.

If the target limit is exceeded, we would reduce the level of payments for approved technologies across the board, to ensure estimated payments do not exceed the limit. Using this approach, all cases involving approved new technologies that would otherwise receive additional payments would still receive special payments, albeit at a reduced amount. Although the marginal payment rate for individual technologies would be reduced, this reduction would be offset by large overall payments to hospitals for new technologies under this provision.

Comment: Some commenters asked that CMS ensure that the necessary software changes be made to accommodate newly approved technologies so that hospitals experience no delay in receiving add-on payments for new technologies. Commenters noted that, at the time they prepared their comments, it was unclear whether hospitals were receiving any new technology add-on payments for FY 2003. Given that $74.8 million was carved out of the FY 2003 standardized amount, it is critical that a reliable system be put in place to ensure that hospitals receive these add-on payments.

Response: We regret the delay any hospital may be experiencing in receiving add-on payments for FY 2003. On December 13, 2002, we issued Program Memorandum A-02-124 that requested fiscal intermediaries to implement the new technology payment mechanism into the claims processing system by April 1, 2003. The changes outlined in this program memorandum were delayed until July 16, 2003, in order to ensure that the claims processing system could properly process these add-on payments.

Comment: Several commenters pointed out that new ICD-9-CM codes are being created for procedures that were not typically captured and reported using ICD-9-CM coding. The commenters specifically mentioned the creation of new codes for types of drugs. Commenters are concerned about the types of medical record documentation that may be required for the administration of these drugs to be assigned an ICD-9-CM code. They asked if a physician order for a drug and a notation on a medical sheet that a nurse had in fact injected the drug were sufficient documentation. The commenters indicated that further guidance is needed regarding documentation requirements for ICD-9-CM codes for new services and technologies that have not traditionally been reported through the use of ICD-9-CM coding.

One commenter recommended that the approval process for new technologies be revised to include a requirement that the applicant must barcode such item with appropriate detailed information. The commenter stated that the use of barcoding would reduce medical errors. The commenter also was concerned that the limit of 6 procedure codes that can be reported on the billing form may become problematic as more new technologies are approved in the future.

Response: We have asked the AHA to schedule this topic for discussion by the Cooperating Parties for ICD-9-CM and the Editorial Advisory Board for Coding Clinic for ICD-9-CM. AHA agrees that this is a timely topic and has scheduled it for discussion in one of its upcoming ICD-9-CM meetings.

We would like to explore further the commenter's suggestion to require applicants for new technology add-on payments to barcode the technology. We recognize the potential limitations of the current claims form, as well as the overall limitations of ICD-9-CM. As we have stated previously, we believe ICD-10-PCS offers great potential improvement for more specific coding that may limit the use of multiple ICD-9-CM codes to identify certain classes of patients.

Comment: Commenters asked that CMS present a full and clear accounting for estimated and actual new technology add-on payments and their impact on the DRG base rate in each proposed and final rule in order to ensure that hospitals receive these add-on payments in full. Another commenter recommended that, similar to outlier payments, CMS should report every year on the extent to which the actual add-on payments per case exceeded or were lower than the amount removed from the standardized amounts.

One commenter was concerned that additional payments might be carved out of the standardized amount for new technologies to ensure budget neutrality, and those payments might not be made because CMS’ projection of spending for the add-on payments was too high or because hospitals failed to bill properly for add-on payments. The commenter recommended that CMS split the budget neutrality adjustment for DRG reclassification and recalibration into two components in order to isolate the reduction associated with add-on payments for new technologies.

Commenters did not agree that add-on payments for new technology should be budget neutral, and explained that the purpose of having additional payments for high-cost items was to compensate a hospital for its unrecovered cost. Because of budget neutrality, these high-cost items are not being properly paid. The commenter also noted that these high-cost items are also the cause of a higher than expected outlier payment.

One commenter recommended that CMS develop a separate pool of money to fund new technology and remove it from the budget neutrality calculation. The commenter explained that, while the technology is new, there should be money set aside and accessed only by those hospitals utilizing that technology.

Response: When we approve a new technology for add-on payments, we conduct an analysis based on the latest data available to estimate the total add-on payments that will be made for the new technology during the upcoming fiscal year and include the results in the annual proposed and final rules. Analyses of technologies approved for add-on payments for FY 2004 are presented below. These analyses include our analysis of available FY 2003 MedPAR data on the utilization of Xigris® and the basis for our estimated payments for new technologies approved for FY 2004. We also discuss this analysis in our description of budget neutrality in section II.A.4.a. of the Addendum to this final rule. We note that, based on our analysis, we have reduced considerably our estimate of add-on payments for Xigris® from the FY 2003 level, which led to a smaller budget neutrality offset to the standardized amounts.

As we stated above, the Congressional Report language accompanying section 533 of Pub. L. 106-554 clearly indicated Congress' intent that this provision be implemented in a budget neutral manner. Therefore, Congress is the appropriate body to consider concerns about the budget neutrality of this provision.

We do not believe it necessary to establish a separate budget neutrality calculation or pool for these payments. The amount of the payments is clearly identified in the final rule. Like all of the budget neutrality calculations, it is a prospective estimate.

Comment: Commenters recommended that CMS eliminate the use of case-weighted averages in the calculation of the cost threshold for technologies that occur in more than one DRG. The commenter believed that the goal of add-on payments is to provide adequate payment for new technologies in the DRGs in which the technology is used. The commenter added that the use of a case-weighted average biases the cost threshold against technologies that occur in more than one DRG and places hospitals at a disadvantage in DRGs where the threshold would otherwise be met except for application of the case-weighted average.

Commenters argued that our criteria for what is considered a new technology is not consistent with section 1886(d)(5)(K)(ii)(II) of the Act. The commenter stated that this provision was intended to provide for the collection of data with respect to the costs of a new medical service or technology for a period of not less than 2 years and not more than 3 years, “beginning on the date on which an inpatient hospital code is issued with respect to the service or technology.” Therefore, the commenter recommended that, instead of no longer considering technologies new because the related charges are already captured in the MedPAR data, CMS should only view a technology as ineligible on the grounds that it is no longer new if the agency can specifically identify a significant sample of cases involving use of the technology in the MedPAR data. One commenter noted that sufficient charge data to assess whether the new technology meets the cost threshold criterion are often only available through the MedPAR data after the new ICD-9-CM code becomes effective. Some commenters also recommended that CMS raise the add-on payment amount from 50 percent of the cost of the new technology to an 80-percent or 100-percent marginal cost factor.

Another commenter asked CMS to provide established clinical requirements or criteria that would control substantial clinical improvement determinations.

One commenter recommended that CMS deem products that fall within one of the following categories designated by the FDA to have met the substantial clinical improvement criterion: Drugs or biologicals that obtain fast track or accelerated approval; and drugs or biologicals approved after priority review or approved for orphan indication. The commenter recommended that CMS defer to the clinical expertise of the FDA with respect to these products and find that any product falling in the above categories satisfy the substantial clinical improvement criterion without further CMS analysis.

In addition, many commenters addressed the proposed change to the cost threshold criterion. (We are addressing these comments in our discussion of specific proposals later in this section of the preamble.)

Response: We appreciate the interest of the many stakeholders in ensuring that Medicare beneficiaries have full access to improvements in medical technology. We have previously discussed our position on each of the issues raised by the commenters on the proposed rule in detail in the September 7, 2001 final rule (66 FR 46905) and the August 1, 2002 final rule (67 FR 50009). Our rationales for these policies have not changed since we discussed them in those final rules, and we did not propose changes to these policies in the May 19, 2003 proposed rule. Therefore, readers are referred to the September 7, 2001 final rule and the August 1, 2002 final rule for our responses to these comments. However, we will continue to assess each of these policies and would appreciate the commenters' continued input on these issues.

Comment: One commenter suggested that CMS conduct a historical review of technologies that would have likely met the “new” and substantial improvement criteria and determine the relationship between the costs of those items and the new technology cost threshold. The commenter noted that such an analysis might provide useful insights as to whether a more flexible cost criterion is needed.

Response: We will take this suggestion under consideration.

2. FY 2004 Status of Technology Approved for FY 2003 Add-On Payments: Drotrecogin Alfa (Activated)—Xigris®

In the August 1, 2002 IPPS final rule, we stated that cases involving the administration of Xigris® (a biotechnology product that is a recombinant version of naturally occurring Activated Protein C (APC)) as identified by the presence of code 00.11 (Infusion of drotrecogin alfa (activated)) are eligible for additional payments of up to $3,400 (50 percent of the average cost of the drug) (67 FR 50013). (The August 1, 2002 final rule contains a detailed discussion of this technology.) Although Xigris® was approved by the FDA in November 2001, it did not qualify for add-on payments until discharges on or after October 1, 2002. Consequently, FY 2002 discharges (between October 1, 2001 and September 30, 2002) may not reflect full utilization of the technology due to the absence of the add-on payment.

Therefore, for FY 2004, we will continue to make add-on payments for cases involving the administration of Xigris® as identified by the presence of code 00.11. Based on preliminary analysis of the incidence of Xigris® in the first quarter FY 2003 MedPAR file, in the May 19, 2003 proposed rule, we proposed to revise downward our estimate of total add-on payments for Xigris®. For FY 2003, we estimated that total add-on payments would be approximately $74.8 million (22,000 Medicare patients who would be eligible for a $3,400 add-on payment). For FY 2004, we estimated in the proposed rule the total add-on payments would be approximately $50 million (based on 14,000 Medicare patients who would be eligible for a $3,400 add-on payment). We indicated that this proposed additional payment would be included in the DRG reclassification and recalibration budget neutrality factor, which is applied to the standardized amounts and the hospital-specific amounts. However, we indicated that, before the publication of the FY 2004 IPPS final rule, we would reevaluate our assumptions regarding this estimate based on preliminary claims data from the FY 2003 MedPAR file.

We have analyzed the claims from the March 2003 update to the FY 2003 MedPAR file. We identified claims that had received Xigris® based on the inclusion of procedure code 00.11. We identified only 1,500 claims from this file. Although the March 2003 update of the FY 2003 MedPAR probably only realistically includes about 5 months' worth of claims, it appears that a lower than expected number of cases are receiving this new technology at the present time.

Therefore, in this final rule for FY 2004, we are lowering the total payments in proportion to the cases that have actually received this drug. We are doubling the number of cases in our March 2003 MedPAR update to an estimated 3,000 cases that will receive Xigris® in FY 2003. We recognize there may actually be more cases than this by the end of the year, as only about 5 months of data are accounted for in our analysis. Also, this estimate does not account for future increased use of the drug. However, these potential underestimates are offset by the fact that we are assuming all cases will qualify for the full $3,400 add-on payment. We believe these effects will largely offset one another. Therefore, the final projected costs for add-on payments are estimated to be $10 million. We will use this estimate in our budget neutrality calculations.

Comment: One commenter supported our decision to continue paying add on payments for Xigris®, but disagreed with the proposed estimated decline in add-on payments in FY 2004 from $74.8 million to $50 million. The commenter explained that this conclusion was made using only first quarter FY 2003 MedPAR data and, since this technology is still in its infancy, the commenter believed FY 2003 MedPAR data will reflect an upward trend in its use and overall availability.

Some commenters were concerned that first year utilization of any new technology is an inappropriate measure for CMS to rely on in determining the full extent of use of a new technology. They asserted that the gradual adoption of new technology and the time required for hospitals to adapt their coding and charge structures to new technologies make it difficult to base projections of the ultimate utilization and costs of new technology immediately following its introduction. In addition, one commenter explained that CMS’ system delays in processing claims have led to a negative impact on both uptake of the technology and the data collection associated with its use.

Also, the commenter explained that Congress required data relating to the cost of the technology be collected for not less than 2 years and not more than 3 years after an appropriate inpatient hospital service code is established. The commenter added that, because CMS publishes its proposed and final rules before the completion of a fiscal year, CMS would make its decision for FY 2005 with less than 2 full year's worth of data. As a result, the commenters recommended that CMS make additional payments for the full 3 years so when it moves a new technology into a DRG, it does so based on accurate and reliable information about its cost and clinical use.

Response: Before each fiscal year, we use the latest available data to determine if we should continue to pay add-on payments for approved new technologies. As stated above, we are continuing to pay for Xigris® for FY 2004 because FY 2002 discharges may not reflect full utilization of the technology. Based on the March update of the FY 2003 MedPAR file, we lowered our cost estimates from the proposed rule because a lower than projected number of cases is receiving this technology at the present time. Before FY 2005, we will again use the latest available data to determine whether we would propose to continue to make add-on payments for Xigris® for FY 2005.

3. FY 2004 Applicants for New Technology Add-On Payments

We received two applications for new technologies to be designated eligible for inpatient add-on payments for new technology for FY 2004. A discussion of these applications and our determinations appear below.

a. Bone Morphogenetic Proteins (BMPs) for Spinal Fusions

An application was submitted for the InFUSE^TM Bone Graft/LT-CAGE^TM Lumbar Tapered Fusion Device (InFUSE^TM) for approval as a new technology eligible for add-on payments. A similar application was submitted last year. However, we denied it because, based on the available data, the technology did not exceed the 1 standard deviation threshold above the average charges for the DRGs to which the technology is assigned.

The product is applied through use of an absorbable collagen sponge and an interbody fusion device, which is then implanted at the fusion site. The patient undergoes a spinal fusion, and the product is placed at the fusion site to promote bone growth. This procedure is done in place of the more traditional use of autogenous iliac crest bone graft. For a more detailed discussion about InFUSE^TM, see the August 1, 2002 IPPS final rule (67 FR 50016).

On July 2, 2002, the FDA approved InFUSE^TM for spinal fusion procedures in skeletally mature patients at one level. Therefore, based on the FDA's approval, multilevel use of this technology would be off-label. In the August 1, 2002 IPPS final rule (67 FR 50017), we stated this technology would meet the cost threshold only if the added costs of multilevel fusions were taken into account. Because the FDA had not approved this technology for multilevel fusions, and the applicant had not submitted data to demonstrate this technology is a substantial clinical improvement for multilevel fusions (the clinical trial upon which the application was based was a single-level fusion trial), we could not issue a substantial clinical improvement determination for multilevel fusions and, consequently, did not consider the costs associated with multilevel fusions in our analysis of whether this technology met the cost threshold. Therefore, because the average charges for this new technology, when used for single-level spinal fusions, did not exceed the threshold to qualify for new technology add-on payment, we denied this application for add-on payments for FY 2003. For similar reasons, we did not consider data on the charges for multilevel fusions in our analysis of whether this technology meets the cost threshold for FY 2004.

In its application for add-on payments for FY 2004, the applicant used data from the CMS FY 2001 Standard Analytical File for physicians and hospitals. The analysis linked a 5-percent sample of hospital spinal fusions cases with the corresponding physician claims. Because there were no ICD-9-CM codes to identify multilevel fusions in 2001, multilevel fusions were identified using CPT codes on the physician claims. Average charges were taken from actual cases used in clinical trials.

After grouping these cases into one, two, and three or more levels fused in DRGs 497 and 498 (Spinal Fusion Except Cervical With and Without CC, respectively), the applicant then calculated average charges assuming the use of the InFUSE^TM for these cases. For DRG 497, the estimated single-level fusion average charge was $41,321; for DRG 498, the estimated single-level fusion average charge was $37,200. Because these DRGs are not currently split for different numbers of fusion levels involved, Medtronic has calculated its own standard deviation of average charges to determine the threshold for these DRGs using the 5-percent sample data. For DRG 497, the threshold (calculated by Medtronic) was $45,646, which is greater than the estimated average charge of $41,321 for single-level fusions noted above. For DRG 498, the threshold (calculated by Medtronic) was $36,935, which is less than the average charges for single-level fusions in this DRG as noted above.

However, we note the thresholds to qualify for the new technology add-on payments for FY 2003 published in Table 10 of the August 1, 2002 IPPS final rule for DRGs 497 and 498 were $58,040 and $41,923, respectively. These thresholds were computed based on all cases assigned to these DRGs, and do not differentiate between the number of spinal levels fused. Because we are not redefining these DRGs to differentiate cases on the basis of the number of levels of the spine fused in the manner suggested by the applicant's analysis, the thresholds published in last year's final rule are applicable for a new technology to qualify for add-on payments in these DRGs for FY 2004. Therefore, because the averages calculated by the applicant for single-level fusions do not exceed the published thresholds, as proposed, we did not approve this technology on the basis of this analysis.

The applicant also submitted data from actual cases involving the InFUSE^TM with single level fusions only. The data submitted included 31 claims from 4 hospitals (only one Medicare patient was included in the sample). All 31 cases were from DRG 498. The average standardized charge for these cases was $47,172. Based on these data, the average standardized charge exceeds the threshold for DRG 498. However, we note that this limited sample excludes any cases from DRG 497.

For discharges occurring on or after October 1, 2002, ICD-9-CM codes 84.51 (Insertion of interbody spinal fusion device) and 84.52 (Insertion of recombinant bone morphogenetic protein) are effective to identify cases involving this technology. Therefore, in an effort to resolve the difficulties in obtaining sufficient data upon which to determine whether this technology exceeds the applicable threshold in the May 19, 2003 proposed rule, we stated our intention to review available MedPAR data for the first several months of FY 2003 to identify these cases and calculate their average standardized charges to compare with the thresholds. We noted that some of these cases would involve multilevel spinal fusions, and that it would be necessary to adjust for those cases in order to remove them from the calculation of the average charges.

We have analyzed data from the March update of FY 2003 MedPAR, containing claims data for the first 6 months of FY 2003. As discussed above, accounting for a lag time in claims processing, we are assuming that this data accounts for approximately 5 months of FY 2003 discharges. We identified InFUSE^TM cases by the presence of the two new ICD-9-CM codes 84.51 and 84.52, used in combination with each other. We identified 117 and 88 cases in the March 2003 MedPAR data for DRGs 497 and 498, respectively.

We standardized the charges to remove the effects of differences in area wage levels, indirect medical education and disproportionate share payments, and, for hospitals in Alaska and Hawaii, the applicable cost-of-living adjustment, and calculated an average standardized charge of $64,931 for the 117 cases in DRG 497. For DRG 498, the average standardized charge was $58,266 for the 88 cases in our data. The average standardized charge across both DRGs was $62,752. As we noted in the proposed rule, we anticipate that some of these cases will involve multilevel spinal fusions. Based on the applicant's analysis of FY 2001 Standard Analytical File data in which they were able to distinguish between one, two, and three or more levels fused by using CPT codes on the physician claims, we determined that the average charges of single level fusions were about 78 percent of the average charges across all spinal fusions in the analysis. (It was not possible to independently match records from the Standard Analytical File in the time available after we attained the March 2003 MedPAR data.) However, as noted above, these data are from FY 2001 and did not include any cases involving InFUSE^TM. Therefore, we anticipate more of the cases in our data will be single-fusion cases, consistent with the FDA approval, and that the total charges in our data for single-level fusion cases will be higher than 78 percent of the average for all InFUSE^TM cases in our data. Given the relatively recent approval by the FDA of this product, we anticipate the majority of uses are in accordance with the FDA's approval criteria. Therefore, to estimate the average standardized charges of the single-level spinal fusion cases in our data, we estimated 90 percent of the average standardized charges of all the InFUSE^TM cases in our data would approximate the charges for single-level cases.

Finally, because these were FY 2003 cases compared to FY 2002 thresholds (based on FY 2001 cases), we adjusted the average charges (by the market basket) to be consistent with the FY 2002 thresholds. The resulting average standardized charge for the cases from our FY 2003 MedPAR data for all InFUSE^TM cases across both DRGs 497 and 498 was $53,376.

We then calculated the case-weighted threshold amount across DRGs 497 and 498 based on the proportion of cases in our data in each DRG. Since 57 percent of the cases we identified in our database were in DRG 497, we applied this percentage to the threshold amount for DRG 497 of $58,040. We then added this amount to 43 percent of the threshold amount for DRG 498, for a combined threshold amount of $51,121. Because our data indicates that the average standardized charge for single-level InFUSE^TM cases exceeds this threshold amount, this technology has met the cost criteria to qualify for new technology add-on payments.

Because the technology meets the cost threshold based on the MedPAR data, we evaluated whether it qualifies as a substantial clinical improvement. According to the applicant:

“InFUSE^TM Bone Graft is more appropriate to use and has been proven more effective in its use than autogenous iliac crest bone graft, when either is placed in the LT-Cage^TM Lumbar Tapered Fusion Device for anterior lumbar interbody fusion. Use of InFUSE^TM Bone Graft instead of autogenous iliac crest bone graft:

• Obviates iliac crest bone graft donor site morbidity.
• Reduces operative time, blood loss and hospitalization.
• Results in greater fusion success.
• We found that the Oswestry Low Back Pain Disability score and SF-36 Physical Component and Pain Index score were consistently 10 percent better in the InFUSE^TM Bone Graft group than the autogenous iliac bone graft group.
• Enables earlier return to work.”

As indicated in the May 19, 2003 proposed rule, among the issues we planned to consider were: does avoiding the complications associated with the iliac crest bone harvesting procedure constitute a substantial clinical improvement; and, with the increased rate of osteoarthritis and osteoporosis in the Medicare population, is there evidence that the technology represents a substantial clinical improvement in spinal fusions among this population? In the May 19, 2003 proposed rule, we indicated we were particularly interested in data on the results of aged Medicare patients who have been treated with BMP, and any basic biology bench data on the results of using BMP in osteoporotic bones.

Since the May 19, 2003 proposed rule, we received from the sponsor of this application an analysis, prepared by an orthopedic surgeon, that showed limited evidence of results in a series of patients older than 65, all with good or better fusion results than the younger age group. That analysis presented evidence that older patients typically have better results than younger patients in the standard iliac crest bone harvesting fusion procedure. Finally, it included the results of bench testing of mesenchymal and osteoblastic cells that demonstrated response to rhBMP-2, including cells from elderly patients.

The sum of this evidence does not preclude generalizing the results of InFUSE^TM trials to Medicare aged beneficiaries. In addition, the small series of Medicare-aged patients treated with InFUSE^TM technology, as well as the bench science on the response of elderly mesenchymal cells to rhBMP-2, do provide some positive, though limited, evidence for generalizability. These results, combined with the benefits of the elimination of the need to harvest bone from the iliac crest (and the associated complications), lead us to conclude that InFUSE^TM does meet the substantial improvement criteria. Therefore, we are approving InFUSE^TM for add-on payments under § 412.88, to be effective for FY 2004.

This approval is on the basis of using InFUSE^TM for a single-level, lumbar spinal fusions, consistent with the FDA's approval and the data presented to us by the applicant. Therefore, we intend to limit the add-on payment to cases using this technology for anterior lumbar fusions in DRGs 497 and 498. Cases involving InFUSE^TM that are eligible for the new technology add-on payment will be identified by assignment to DRGs 497 or 498 as a lumbar spinal fusion, with the combination of ICD-9-CM procedure codes 84.51 and 84.52.

As explained above, we are limiting our approval of this technology to uses consistent with our substantial clinical improvement decision. Therefore, add-on payments are only available for use of the technology at a single-level. The average cost of the InFUSE^TM is reported to be $8,900, and a single level fusion requires two of the products. Therefore, the total cost for the InFUSE^TM for a single-level fusion is expected to be $17,800. Under § 412.88(a)(2), new technology add-on payments are limited to the lesser of 50 percent of the average cost of the device or 50 percent of the costs in excess of the DRG payment for the case. As a result, the maximum add-on payment for a case involving the InFUSE^TM is $8,900.

For purposes of budget neutrality, it is necessary to estimate the additional payments that would be made under this provision during FY 2004. We identified 205 cases in DRGs 497 and 498 in the March 2003 update of the FY 2003 MedPAR data. For our FY 2004 budget neutrality estimate, we are projecting this number will grow to 500. Given this estimate and the maximum add-on payment of $8,900, we estimate the total amount of the add-on payments for the InFUSE^TM for FY 2004 will be $4.4 million dollars.

Comment: One commenter asked that CMS reconsider the decision to exclude multilevel fusions with InFUSE^TM from the cost threshold calculation. The commenter noted that excluding multilevel fusions with InFUSE^TM is inconsistent with FDA guidance, clinical practice and other CMS payment decisions for new technologies (notably the creation of DRGs for drug-eluting stents based on the presence of a condition not indicated on the product label, that is, acute myocardial infarction).

Response: As stated previously, because the FDA has not approved this technology for multilevel fusions and the applicant has not submitted data to demonstrate this technology is a substantial clinical improvement for multilevel fusions, we cannot issue a substantial clinical improvement for multilevel fusions. In the September 7, 2001 final rule implementing this provision (66 FR 46913), we stated our position that the special payments under this provision should be limited to those new technologies that have been demonstrated to represent a substantial improvement in caring for Medicare beneficiaries. Where such an improvement is not demonstrated, we continue to believe the incentives of the DRG system provide a useful balance to the introduction of new technologies, and no new technology add-on payment is necessary.

Comment: In the proposed rule, we stated that, if InFUSE^TM meet the cost threshold, we would evaluate whether it qualifies as a substantial clinical improvement. One commenter noted that, assuming InFUSE^TM does meet the cost threshold, CMS would make a determination on whether the technology meets the substantial clinical improvement criterion without public input or the opportunity to address concerns that CMS may have. The commenter noted that these actions are inconsistent with the Administrative Procedure Act and CMS's pledge to be more open in its policy making.

Response: Because of the many questions that remained at the time of the proposed rule, we were unable to determine if InFUSE^TM qualified as a substantial clinical improvement. However, in order to receive comments on this determination, we indicated certain issues we would consider when determining if InFUSE^TM qualifies as a substantial clinical improvement. As noted above, we received additional information that enabled us to approve this technology as a substantial clinical improvement. Therefore, we believe interested parties had sufficient information to provide informed comments.

Comment: One commenter, a designer, manufacturer, and supplier of orthopedic devices and supplies, explained that the applicant's analysis probably includes cases for both posterior approaches or posterior instrumentation, or both, which are considered off-label uses from the indications approved by the FDA. Therefore, the commenter requested that cases that do not meet FDA approved indications, once identified, be eliminated from the analysis.

The commenter also noted that once claims of InFUSE^TM can be identified with MedPAR data, DRG weights become eligible for recalibration in order to reflect the appropriate payment within the assigned DRG. Once the weights of a DRG can be evaluated, a technology should no longer be classified as new. Also, the commenter stated that clinical trial results counter the claim of significant improvement, because information presented at the FDA Orthopedics and Rehabilitation Devices Panel public meeting on January 20, 2002, indicated that the InFUSE^TM product resulted in an equivalency to that of traditional bone grafting techniques. Although there was a decrease in donor site pain in a small number of subjects in the BMP group, compared with the control group, the commenter questioned whether this factor meets the criteria of substantial clinical improvement. The commenter also questioned the results of a published article on this technology.

Response: One of the criteria for a substantial clinical improvement classification is avoidance of surgery. CMS determined that InFUSE^TM should be classified as a substantial improvement if the results of the clinical trials demonstrated outcomes at least equivalent to bone grafting, and the bone harvesting procedure was avoided. CMS clinical staff reviewed the literature and concluded that the current evidence did support grafting equivalence for the FDA approved indications and, therefore, InFUSE^TM met the substantial improvement standard. As described above, we did not rely on the applicant's analysis to determine the technology met the high-cost threshold, but conducted direct analysis of available FY 2003 MedPAR data.

b. GLIADEL^® Wafer

Glioblastoma Multiforme (GBM) is the most common and most aggressive of the primary brain tumors. Standard care for patients diagnosed with GBM is surgical resection and radiation. According to the manufacturer, the GLIADEL^® Wafer is indicated for use as an adjunct to surgery to prolong survival in patients with recurrent GBM. Implanted directly into the cavity that is created when a brain tumor is surgically removed, GLIADEL^® delivers chemotherapy directly to the site where tumors are most likely to recur.

The FDA approved GLIADEL_® Wafer on September 23, 1996, for use as an adjunct to surgery to prolong survival in patients with recurrent GBM for whom surgical resection is indicated. In announcing its approval, the FDA indicated that GLIADEL^® was approved:

“* * * based on the results of a multi-center placebo controlled study in 222 patients who had recurrent malignant glioma after initial treatment with surgery and radiation therapy. Following surgery to remove the tumor, half of the patients were treated with GLIADEL^® implants and half with placebo. In patients with glioblastoma multiforme, the 6-month survival rate increased from 36 percent with placebo to 56 percent with GLIADEL^® Median survival increased from 20 weeks with placebo to 28 weeks with GLIADEL^®. In patients with pathologic diagnoses other than glioblastoma multiforme, GLIADEL^® had no effect on survival.”

Guilford Pharmaceuticals has requested that GLIADEL^® still be considered new because, until a new ICD-9-CM code (00.10 Implementation of Chemotherapeutic Agent) was established on October 1, 2002, it was not possible to identify specifically these cases in the MedPAR data. However, as noted previously, technology will no longer be considered new after the costs of the technology are reflected in the DRG weights. Because the costs of GLIADEL^® are currently reflected in the DRG weights (despite the absence of a specific code), GLIADEL^® does not meet our criterion that a medical service or technology be “new”. That is, FY 2002 MedPAR data used to calculate the DRG weights for FY 2004 in this final rule include cases where GLIADEL^® was administered (and the corresponding charges of these cases include charges associated with GLIADEL^®). On February 26, 2003, the FDA approved GLIADEL^® for use in newly diagnosed patients with high-grade malignant glioma as an adjunct to surgery and radiation. However, our understanding is that many newly diagnosed patients were already receiving this therapy. To the extent this is true, the charges associated with this use of GLIADEL^® are also reflected in the DRG relative weights.

According to Guilford's application, the current average wholesale price of GLIADEL^® is $10,985. Guilford submitted charge data for 23 Medicare patients at 7 hospitals from FY 2000. The charges were then standardized and adjusted for inflation using the hospital market basket inflation factor (from 2000 to 2003) in order to determine an inflated average standardized charge of $33,002. Guilford points out that this charge narrowly misses the DRG 2 threshold published in Table 10 of the August 1, 2002 IPPS final rule of $34,673. However, we note that, according to the manufacturer, as many as 60 percent of current GLIADEL^® cases may be assigned to DRG 1 based on the presence of CCs. Based on this assumption, the qualifying threshold for GLIADEL^® would be $54,312 (60 percent of the DRG 1 threshold of $67,404, and 40 percent of the DRG 2 threshold of $34,673).

As mentioned in section II.B.3.a of the May 19, 2003 proposed rule and above in this final rule, we examined the definitions of DRGs 1 and 2 to determine whether they could be improved. As proposed, we are creating a new DRG for patients with an intracranial vascular procedure and an intracranial hemorrhage and two new DRGs for patients with only a vascular shunt procedure (splitting on the presence or absence of a CC). We also compared the data submitted in the application for add-payments regarding the charges for GLIADEL^® cases with the charges of other procedures in DRGs 1 and 2. We found that, although the $33,002 average standardized charge reported is just below the qualifying threshold in DRG 2, it is actually well below the mean average standardized charge for DRG 1 ($42,092). As noted previously, as many as 60 percent of current GLIADEL^® cases may be assigned to DRG 1 based on the presence of CCs. Therefore, we do not believe that any change to the DRG assignment of cases receiving GLIADEL^® is warranted at this time. However, we will continue to monitor our data to determine whether a change is warranted in the future.

Comment: One commenter supported CMS’ determination that this technology is currently reflected within the DRG weights and does not meet the criteria of being called “new.” Another commenter commented that CMS’ interpretation of whether a technology is “new” is inconsistent with the current statute. The commenter explained that section 1886 (d)(5)(K)(ii)(II) of Act states that CMS should collect data on new technologies “for a period of not less than 2 years and not more than 3 years beginning on the date on which an inpatient hospital code is issued for the technology.” Accordingly, the commenter believed it is inconsistent with the intent of Congress to deny new technology status to a product that has been on the market but for which there is no unique ICD-9 code that allows CMS to track the costs of cases in which it is utilized. The commenter urged CMS to reconsider its interpretation of the statute and approve GLIADEL^® as a new technology, making clear that a technology will be considered new for 2 to 3 years from the date that an ICD-9-CM code, specific to the technology, becomes available.

Response: As stated above, we discussed our position on this issue in detail in the September 7, 2001 final rule (66 FR 46905). Our rationale for this policy has not changed since we discussed it in that final rule, and we did not propose changes to this policy in the May 19, 2003 proposed rule. Therefore, we are denying this application for add-on payments for FY 2004.

4. Review of the High-Cost Threshold

The current cost threshold for a new technology to qualify for add-on payments is that the average standardized charges of cases involving the new technology must be demonstrated to exceed 1 standard deviation beyond the geometric mean of the standardized charges of the DRG to which the new technology will be assigned. If the new technology is assigned to more than one DRG, the qualifying threshold is equal to the case-weighted (based on the proportion of cases involving the new technology estimated to be assigned to each DRG) average threshold across all relevant DRGs. When we established this threshold in the September 7, 2001 final rule, we expressed our belief that it is important to establish a threshold that recognizes the variability in costs per case within DRGs and maintains the fundamental financial incentives of the IPPS (66 FR 46917).

In commenting on this approach, MedPAC and some hospital associations supported the 1 standard deviation threshold. However, others, particularly representatives of the manufacturers of new technology, have argued this threshold is too high, and that virtually no new technology would qualify for the special payment provision.

We are concerned that establishing higher payments for a great number of new technologies may be inflationary because the add-on payments reduce the efficiency incentives hospitals face when new technologies must otherwise be financed out of current payments for similar cases. Traditionally, under the IPPS, new technologies were required to compete with existing treatment methods on clinical and cost criteria. Add-on payments are intended to give new technologies a competitive boost relative to existing treatment methods with the goal of encouraging faster and more widespread adoption of new technologies.

Much of the current variation around the mean within any particular DRG is due to the range of procedures contained within each DRG. Generally, some of these procedures will be more expensive than the mean and some will be less expensive. The threshold should be set high enough to ensure that it identifies truly high-cost technologies. If the threshold were set too low (for example, at $2,500, as some have suggested), additional technologies may qualify merely by association with a procedure only slightly more costly than the mean for the DRG.

For example, consider a DRG with five different procedures and mean charges of $15,000. The mean charges for each procedure are distributed around $15,000, as illustrated in the following table. A qualifying threshold of $2,500 would result in any new technology that is only used for the fifth procedure automatically qualifying for new technology add-on payments (unless the new technology had the unlikely effect of lowering the mean cost for cases with this procedure by at least $2,500). This is because the average charge of $20,000 for cases in this procedure already exceeds the mean charges for the DRG plus $2,500.

At the same time, we recognize that the very limited number of applications that have been submitted the past 2 years (five for FY 2003; two for FY 2004) may indicate that only a very small number of the new technologies that come onto the market every year are costly enough even to apply for new technology add-on payments. Therefore, for FY 2005 and subsequent fiscal years, in the May 19, 2003 proposed rule, we proposed to reduce the threshold to 75 percent of 1 standard deviation beyond the geometric mean standardized charge for all cases in the DRG to which the new medical service or technology is assigned (§ 412.87(b)(3)).

Based on our analysis of the thresholds for FY 2004, this proposed change would reduce the average threshold across all DRGs to qualify for the add-on payments from approximately $9,900 above the mean standardized charges for each DRG to approximately $7,400. This reduction would maintain the averaging principles of the IPPS while easing the requirement somewhat to allow more technologies to qualify. Furthermore, the situation illustrated above, where a technology qualifies on the basis of its association with a high cost procedure, is much less likely to occur as a result of this reduction than if the threshold were reduced dramatically.

Comment: Some commenters were concerned that the revised threshold of 75 percent of the standard deviation remains too high. The commenters noted that even with the revised cost threshold, few technologies would qualify for add-on payments.

On the assumption that the vast majority of technologies that would qualify for add-on payments would be identified by a new ICD-9-CM procedure code, one commenter identified a total of 26 ICD-9-CM procedure codes issued between the years of 1998 and 2001. The commenter then analyzed 2001 MedPAR data and found that only 2 of the 26 procedures will exceed either the current 1 standard deviation threshold, and 4 would exceed the a threshold at 75 percent of 1 standard deviation. The commenter also explained that the proposed reduction of the threshold is only an 8-percent reduction, and continues to block eligibility for add-on-payments for important new technologies, even where costs increase by 70 percent. The commenter recommended that CMS use a threshold based upon 75 percent of the standardized amount inflated to charges, plus the geometric mean charges for the DRG. The commenter identified 13 of the 26 procedures that would qualify using this threshold.

Another commenter asked that CMS consider adopting separate criteria for biologics and devices, because they have different price levels and pricing patterns relative to drugs and relative to DRG standardized amounts. Other commenters recommended a threshold where the cost of the technology must exceed the cost of existing technologies by at least 50 percent of the DRG standardized amount, multiplied by the DRG weight, but not to exceed $7,500.

One commenter was concerned that, because of budget neutrality, any reduction to the threshold for new technologies would allow more technologies to qualify for add-on payments and would therefore reduce payments for all other hospital inpatient services. The commenter explained that shifting money within the IPPS leaves some hospitals without additional money they need to ensure beneficiaries have access to the newest medical tests and treatments. Therefore, the commenter recommended that add-on payments continue to be limited to new, cutting-edge, breakthrough technologies with significant cost implications.

Response: As stated in the August 1, 2002 final rule (67 FR 50011), it is our intention to implement this provision without fundamentally disrupting the IPPS. A substantial number of cases receiving extra cost-based payments (or substantial disaggregation of the DRGs into smaller units of payment) would undermine the efficiency incentives of the DRG payment system. Also, we continue to believe a threshold based on the standard deviation is appropriate for this purpose. (For further reading on this, see the September 7, 2001 final rule (66 FR 46917).)

The DRG system is an average-based system under which hospitals expect to finance costly cases through less costly cases. We believe the add-on policy envisioned by some commenter, that would reduce the maximum threshold across all DRGs to 75 percent of the standardized amount (approximately $3,300) adjusted to charges, would significantly disrupt the averaging principles of the IPPS. By assuming only 26 new technologies over a 4-year span, the analysis presented by the commenter dramatically underestimates the annual volume of new technologies that would be likely to meet such a reduced threshold. Industry sources cite over 1,000 companies producing medical devices, diagnostic products, and medical information systems in the U.S., producing over $70 billion worth of products annually. A very limited number of these products receive specific ICD-9-CM procedure codes, particularly in years prior to the establishment of the IPPS new technology add-on policy. A more accurate estimate of the number of technologies likely to be approved under this revised threshold could be attained by listing the technologies approved during that period with the average wholesale price.

As stated above, we recognize the limited number of applications for add-on payments that have been submitted in the past 2 years and, therefore, we are lowering the threshold. We believe this new threshold is a fair balance that maintains the averaging principles of the IPPS while easing the qualifying requirement. Therefore, for FY 2005 and subsequent fiscal years, we are reducing the threshold to 75 percent of 1 standard deviation (based on the logarithmic values of the charges) beyond the geometric mean standardized charges for all cases in the DRG to which the new medical service or technology is assigned, transformed back to charges.

We disagree with the commenter's suggestion that we establish separate thresholds for biologics and devices. We believe the IPPS is intended to pay hospitals for their costs to treat patients, and physicians select from a range of options based on the medical needs of the patients. The payment system should be neutral with respect to those options. We are concerned that establishing separate thresholds for biologics and devices would indicate an inappropriate payment preference for one or the other option.

Comment: Other commenters representing hospitals approved of the threshold proposed by CMS. One commenter explained that a threshold that limits the number of new technologies is necessary, as the administrative burden for hospitals and the program is significant for each additional item qualifying. Given the finite pool of funds, an abundance of qualifying technologies could result in prorata reductions, such as those that were experienced under the outpatient prospective payment system. With that in mind, the commenter asked that CMS look at other approval mechanisms that would direct the funds to be focused on significantly expensive new technologies that also have significant volumes nationally. For example, national expenditures projected by CMS for each technology seeking approval should exceed $30 million. Assuming national total expenditures of $75 billion with a 1 percent set aside at $750 million, and a marginal cost at 50 percent, 25 technologies could be approved by CMS.

As an alternative, the commenter recommended that CMS incorporate new technologies into the appropriate DRG without having to specifically code the new technology. The DRG weights would then be adjusted to reflect the increased costs associated with such new technologies rather than making a separate add-on payment. The commenter believed this would be a reasonable compromise between the need to incorporate new technologies into the DRGs, while avoiding an unduly burdensome coding and billing process.

Response: We believe the incremental costs to hospitals associated with this provision should be minimal. Specifically, the additional payments are triggered by the presence of an ICD-9-CM code on the bill, information already required to process the claim for normal DRG payments. Accordingly, there should be little need for training or other operational changes in response to the approval of a new technology for add-on payments.

Also, adding further criteria as suggested by the commenter would make it even more difficult for new technologies to qualify for add-on payments. In this final rule, it is our intention to lower the threshold in order to increase the number of applications we receive each year for add-on payments. With respect to the commenter's suggestion to incorporate a new technology in a DRG and raise the weight of the DRG based on the increased cost of the new technology, we are concerned that this suggestion would have the potential to create possibly large imbalances in the DRG weights if the predicted volume of a particular technology turns out to be inaccurate. We believe an add-on payment is the most appropriate methodology to provide additional payments for qualifying high cost new technologies, while still maintaining the overall integrity of the DRG system.

5. Technical Changes

Subpart H of part 412 describes payments to hospitals under IPPS. We have become aware of references to the calculation of IPPS payments in this subpart that inadvertently omit references to new technology add-on payments. For example, § 412.112(c) describes the basis for per case payments. This section refers to outlier payments under subpart F, but was not revised to reflect the implementation of the new technology add-on payments. Therefore, in the May 19, 2003 proposed rule, we proposed to amend § 412.112(c) to add a new paragraph (d) to include a reference to additional payments for new medical services or technologies under subpart F.

We did not receive any comments on this proposal and, therefore, are adopting it as final.

Section 412.116(e) currently states that payments for outlier cases are not made on an interim basis. That is, for hospitals receiving payments under a biweekly, lump-sum payment methodology, outlier payments are not included in the calculation of the lump-sum payment amounts. Rather, outlier payments are calculated on a case-by-case basis. Similarly, due to the unique nature of the new technology add-on payments, in the May 19, 2003 proposed rule, we proposed that they would also be calculated on a case-by-case basis rather than included in the calculation of interim payment amounts. Therefore, we proposed to revise § 412.116(e) to include this policy.

We did not receive any comments on this proposal. Therefore, in this final rule, we are adopting the proposal as final without modification.

III. Changes to the Hospital Wage Index

A. Background

Section 1886(d)(3)(E) of the Act requires that, as part of the methodology for determining prospective payments to hospitals, the Secretary must adjust the standardized amounts “for area differences in hospital wage levels by a factor (established by the Secretary) reflecting the relative hospital wage level in the geographic area of the hospital compared to the national average hospital wage level.” In accordance with the broad discretion conferred under the Act, we currently define hospital labor market areas based on the definitions of Metropolitan Statistical Areas (MSAs), Primary MSAs (PMSAs), and New England County Metropolitan Areas (NECMAs) issued by the Office of Management and Budget (OMB). OMB also designates Consolidated MSAs (CMSAs). A CMSA is a metropolitan area with a population of one million or more, comprising two or more PMSAs (identified by their separate economic and social character). For purposes of the hospital wage index, we use the PMSAs rather than CMSAs because they allow a more precise breakdown of labor costs. If a metropolitan area is not designated as part of a PMSA, we use the applicable MSA. For purposes of the IPPS wage index, rural areas are counties outside a designated MSA, PMSA, or NECMA. For purposes of the wage index, we combine all of the rural counties in a State to calculate a rural wage index for that State.

We note that, effective April 1, 1990, the term Metropolitan Area (MA) replaced the term MSA (which had been used since June 30, 1983) to describe the set of metropolitan areas consisting of MSAs, PMSAs, and CMSAs. The terminology was changed by OMB in the March 30, 1990 Federal Register to distinguish between the individual metropolitan areas known as MSAs and the set of all metropolitan areas (MSAs, PMSAs, and CMSAs) (55 FR 12154). For purposes of the IPPS, we continue to refer to these areas as MSAs.

Under section 1886(d)(8)(B) of the Act, hospitals in certain rural counties adjacent to one or more MSAs are considered to be located in one of the adjacent MSAs if certain standards are met. Under section 1886(d)(10) of the Act, the Medicare Geographic Classification Review Board (MGCRB) considers applications from hospitals for geographic reclassification from a rural area to a MSA, from one rural area to another rural area, or from one MSA to another MSA for purposes of payment under the IPPS.

On June 6, 2003, the Office of Management and Budget (OMB) issued OMB Bulletin No. 03-04, announcing revised definitions of Metropolitan Statistical Areas and new definitions of Micropolitan Statistical Areas and Combined Statistical Areas. A copy of the bulletin may be obtained at the following Internet address: http://www.whitehouse.gov/omb/bulletins/b03-04.html. According to OMB, “(t)his bulletin provides the definitions of all Metropolitan Statistical Areas, Metropolitan Divisions, Micropolitan Statistical Areas, Combined Statistical Areas, and New England City and Town Areas in the United States and Puerto Rico based on the standards published on December 27, 2000, in the Federal Register (65 FR 82228-82238) and Census 2000 data.”

In the proposed rule, we stated that we would evaluate the new area designations and their possible effects on the Medicare hospital wage index. In addition, we proposed that the earliest usage of these new definitions would be the FY 2005 wage index.

The new definitions recognize 49 new Metropolitan Statistical Areas and 565 new Micropolitan Statistical Areas, as well as extensively revising the construct of many of the existing Metropolitan Areas. For example, according to OMB's previous definition of the Asheville, NC MSA, this Metropolitan Statistical Area was comprised of Buncombe and Madison counties. When we apply the new definitions, Asheville's Metropolitan Statistical Area includes both Buncombe and Madison counties, as well as Henderson and Haywood counties. An example of a Micropolitan Statistical Area is that of Elizabeth City, NC which includes Camden, Pasquotank, and Perquimans counties. These were non-Metropolitan Statistical Area counties in previous OMB definitions.

In order to implement these changes for the IPPS, it is necessary to identify the new area designation for each county and hospital in the country. Because this process will have to be extensively reviewed and verified, we are unable to undertake it before publication of this final rule. In addition, because we wish to engage in notice and comment rulemaking, prior to adopting these changes, it would be impractical to have done so prior to this final rule. (We note that the OMB Bulletin was issued during the comment period and we did not receive any comments regarding whether the new definitions should be applied to the FY 2004 wage index or objecting to our proposed policy of implementing the changes in FY 2005 at the earliest.)

Finally, geographic reclassification decisions for FY 2004 have already been made based on the previous Metropolitan Statistical Area definitions. These decisions would have to be individually reevaluated if we were to adopt the new OMB definitions for FY 2004. This would not be possible to accomplish while complying with the requirement of section 1886(d)(6) of the Act to publish this annual IPPS update final rule by August 1. For these reasons, at this time, we are not applying these new definitions to the FY 2004 wage index.

Comment: Several commenters recommended that when CMS does implement OMB's new definitions, it should adopt the new 49 MSAs as outlined in the OMB Bulletin. However, the commenters mentioned that the adoption of the MSAs for FY 2004 would be premature, given the magnitude of the policy change. One commenter encouraged CMS to issue a rule or to elaborate on plans for the new Metropolitan and Micropolitan Statistical Area definition changes as soon as possible to allow time for impact analysis, as well as public comments and input. One commenter raised concerns with respect to the criteria that OMB used to define the new MSAs.

Response: We indicated in the proposed IPPS rule that we would need to assess these new definitions before adopting them. In order to implement such a change, it will be necessary to identify the new area designation for each county and hospital in the country, requiring extensive review and verification. We will undertake this analysis as soon as possible. We intend to move very deliberately and expeditiously regarding these potentially vast changes. Any changes would be made through notice and comment rulemaking. Therefore, we are not addressing technical comments relating to the new MSAs in this document.

Beginning October 1, 1993, section 1886(d)(3)(E) of the Act requires that we update the wage index annually. Furthermore, this section provides that the Secretary base the update on a survey of wages and wage-related costs of short-term, acute care hospitals. The survey should measure, to the extent feasible, the earnings and paid hours of employment by occupational category, and must exclude the wages and wage-related costs incurred in furnishing skilled nursing services. This provision also requires us to make any updates or adjustments to the wage index in a manner that ensures that aggregate payments to hospitals are not affected by the change in the wage index. This adjustment is discussed in section II.4.a. of the Addendum to this final rule.

As discussed below in section III.F. of this preamble, we also take into account the geographic reclassification of hospitals in accordance with sections 1886(d)(8)(B) and 1886(d)(10) of the Act when calculating the wage index. Under section 1886(d)(8)(D) of the Act, the Secretary is required to adjust the standardized amounts so as to ensure that aggregate payments under the IPPS after implementation of the provisions of sections 1886(d)(8)(B) and (C) and 1886(d)(10) of the Act are equal to the aggregate prospective payments that would have been made absent these provisions. This adjustment is discussed in section II.4.b. of the Addendum to this final rule.

Section 1886(d)(3)(E) of the Act also provides for the collection of data every 3 years on the occupational mix of employees for each short-term, acute care hospital participating in the Medicare program, in order to construct an occupational mix adjustment to the wage index. The initial collection of these data must be completed by September 30, 2003, for application beginning October 1, 2004 (the FY 2005 wage index). In the April 4, 2003 Federal Register (68 FR 16516), we published a notice of intent to collect calendar year 2002 data from hospitals.

Many commenters on the April 4, 2003 notice requested that CMS publish a more detailed proposed methodology, illustrating how the occupational mix index will be calculated and how it will be used to adjust the overall wage index. Other comments on the April 4, 2003 notice included: CMS should develop or expand more categories to include all hospital employees; CMS should develop and publish a more reasonable timeframe for the hospitals to complete the survey, and a more reasonable timeframe for fiscal intermediaries to audit the occupational mix survey; CMS should clarify the relationship between the current annual cost report wage index schedule and the proposed occupational mix survey.

We plan to publish a final notice of intent in the Federal Register, with a 30-day comment period. The notice will include any revisions to the survey published on April 4, 2003 based on the comments we received, a detailed timetable, and all audit guidelines. Subsequent to that, we plan to send the surveys to all IPPS hospitals (and hospitals in Maryland that are under a waiver from the IPPS) through the fiscal intermediaries, with the intent to collect these data to be incorporated in the FY 2005 wage index.

Comment: In response to the May 19, 2003 IPPS proposed rule, commenters requested that we publish a detailed proposed methodology, for comment, illustrating how the occupational mix index will be calculated and how it will be used to adjust the overall wage index.

Response: Although our approach will not be finalized until publication of the FY 2005 rule, one possible approach to computing an occupational mix adjusted index is to first calculate, based on the hours collected for each occupational category from all hospitals nationally, a national average percentage attributable to each occupational category. Next, for each hospital, the total dollars and hours for each category would be summed, and an average hourly wage would be determined for each category by dividing dollars by hours. Each hospital's occupational mix adjusted average hourly wage would be calculated by multiplying each category's average hourly wage by the applicable weighting factors and then summing the results across all categories. Similar calculations would then be performed at the labor market level and the national level to construct an index.

We intend to analyze the impacts of implementing an occupational mix adjusted index in the proposed rule for FY 2005. Based on the estimated impacts, we will also evaluate at that time the possibilities for blending such an index with the FY 2005 wage index calculated using our current methodology based on data from the Worksheet S-3, Part II of the Medicare cost report.

B. FY 2004 Wage Index Update

The FY 2004 wage index values (effective for hospital discharges occurring on or after October 1, 2003 and before October 1, 2004) in section VI. of the Addendum to this final rule are based on the data collected from the Medicare cost reports submitted by hospitals for cost reporting periods beginning in FY 2000 (the FY 2003 wage index was based on FY 1999 wage data).

The data for the FY 2004 wage index were obtained from Worksheet S-3, Parts II and III of the FY 2000 Medicare cost reports. Instructions for completing the Worksheet S-3, Parts II and III are in the Provider Reimbursement Manual, Part I, sections 3605.2 and 3605.3. The FY 2004 wage index includes the following categories of data associated with costs paid under the IPPS (as well as outpatient costs), which were also included in the FY 2003 wage index:

• Salaries and hours from short-term, acute care hospitals.
• Home office costs and hours.
• Certain contract labor costs and hours (includes direct patient care, certain top management, pharmacy, laboratory, and nonteaching physician Part A services).
• Wage-related costs (The September 1, 1994 Federal Register included a list of core wage-related costs that are included in the wage index, and discussed criteria for including other wage-related costs (59 FR 45356)).

Consistent with the wage index methodology for FY 2003, the wage index for FY 2004 also excludes the direct and overhead salaries and hours for services not subject to IPPS payment, such as skilled nursing facility (SNF) services, home health services, costs related to GME (teaching physicians and residents) and certified registered nurse anesthetists (CRNAs), and other subprovider components that are not paid under the IPPS.

These wage data are also currently used to calculate wage indexes applicable to other providers, such as SNFs, home health agencies, and hospices. They are also used for prospective payments to rehabilitation and long-term care hospitals, and for hospital outpatient services.

C. FY 2004 IPPS Wage Index

1. Elimination of Wage Costs Associated With Rural Health Clinics and Federally Qualified Health Centers

In the FY 2001 IPPS final rule, we discussed removing from the wage index the salaries, hours, and wage-related costs of hospital-based rural health clinics (RHCs) and Federally qualified health centers (FQHCs) because Medicare pays for these costs outside of the IPPS (65 FR 47074). We noted that because RHC and FQHC costs were not previously separately reported on Worksheet S-3 of the Medicare cost report, we could not exclude these costs from the prior wage indexes. We further noted that we would evaluate the exclusion of RHC and FQHC wage data in developing the FY 2004 wage index.

We revised the FY 2000 Worksheet S-3 so that it now allows for the separate reporting of RHC and FQHC wage costs and hours. In the May 19, 2003 proposed rule, we proposed to exclude the wage and hours data for RHCs and FQHCs from the hospital wage index calculation beginning with the FY 2004 wage index.

We received several comments, all supporting this proposal. Therefore, beginning with the FY 2004 wage index, we are excluding the salaries, hours and wage-related costs associated with RHCs and FQHCs. This change is consistent with others we have implemented in our continuous effort to limit the wage index as much as possible to costs for which hospitals receive payment under IPPS. An analysis of the effects of this change is included in the Appendix A of this final rule.

2. Paid Hours

It has been the longstanding policy of CMS to calculate the wage index using paid hours rather than hours worked (see the September 1, 1993 Federal Register, 58 FR 46299). This policy reflects our belief that paid hours more appropriately reflect a hospital's total wage costs, which include amounts paid for actual time worked and for covered leave periods (for example, annual, sick, and holiday leave). Therefore, the inclusion of paid lunch hours in the wage index is consistent with our inclusion of other paid nonworking hours.

Several hospitals have requested that we exclude paid lunch or meal break hours from the wage index calculation. At these hospitals, the typical workday is 71/2 working hours, plus a 1/2 hour paid meal break, for a total of 8 paid hours. These hospitals, some of which are municipal-owned and required by their overarching union contracts to provide paid lunch hours, believe they are disadvantaged by a wage index policy that requires paid lunch hours to be included in calculating the wage index.

The hospitals argue that their practice of paying employees for meal breaks is not substantially different, in practice, from other hospitals whose employees do not receive paid lunch hours but who are on call during their lunch periods. These hospitals further argue that this policy causes them, in some cases due to union contracts beyond their control, to be the only hospitals with this category of nonproductive hours included in their wage index.

In the May 19, 2003 proposed rule, we solicited comments on our policy that paid lunch hours should be excluded from the wage index. Specifically, we were interested in a broader understanding of the issue of whether some hospitals may, in fact, be truly disadvantaged by this policy through no fault of their own. We indicated that any change in our policy would not be implemented until, at the earliest, the FY 2005 wage index.

Some hospitals and associations have also recommended that we exclude the paid hours associated with military and jury duty leave from the wage index calculation. They state that, unlike other paid leave categories for which workers are usually paid at their full hourly rates (for example, annual, sick, and holiday), hospitals typically pay employees on military or jury duty only a fraction of their normal pay. The amount that the hospital pays is intended to only supplement the earnings that the employee receives from the government so that, while performing military or civic duties, the employee can continue to be paid the same salary level, as if he or she were still working at the hospital.

The hospitals and associations believe that including lower pay rates associated with employees' military and jury duty leave unfairly decreases a hospital's average hourly wage and, therefore, its wage index value. Therefore, we proposed to exclude from the wage index the paid hours associated with military and jury duty leave, beginning with the FY 2005 wage index. We also proposed that the associated salaries would continue to be reported on Worksheet S-3, Part II, Line 1 of the Medicare cost report.

Comment: A few commenters agreed that paid lunch hours and hours associated with military and jury duty leave should be removed from the wage index. Many more commenters, including some national and state hospital associations and Medicare fiscal intermediaries, opposed or expressed concern about whether excluding paid lunch hours and hours associated with military and jury duty leave would result in a more accurate wage index.

Those commenters who opposed the proposal to exclude paid lunch hours and hours associated with military and jury duty leave expressed concern that these changes would further complicate the wage index and that the additional data collection effort for providers might outweigh any benefits achieved through these changes. Further, the commenters believed that paid lunch hours, military, and jury leave affect all providers in the same way, so the changes would likely be immaterial. One commenter also expressed concern that excluding paid hours could cause hospitals to rewrite existing contracts to raise their wage index. In addition, some commenters cautioned that excluding these paid hours would be difficult for intermediaries to apply consistently; excluding these hours would require estimations because most payroll systems do not capture this data. Many commenters indicated that CMS had not published data to provide support that these changes are warranted.

One commenter suggested that, if CMS excludes paid lunch hours, CMS should set a standard for hospitals to qualify for excluding the hours, such as the Fair Labor Standards Act requirements for payment. Another suggested that the determination of excluding paid lunch hours should be based on whether lunch is included for the purpose of computing the hourly wage rate used to pay for overtime. If paid lunch hours are included in the overtime payment computation, and excluding them would result in an hourly rate that is higher than what is usually used for overtime, the paid lunch hours should be excluded. If the paid lunch hours are not included in computing the hourly wage for overtime, and excluding them would result in the correct hourly wage rate that should be used for overtime, the lunch hours should be excluded. Two commenters recommended that the wage index should also exclude time associated with paid breaks from the wage index, but acknowledged that paid breaks are not usually tracked in payroll systems. One commenter recommended that CMS allow all hospitals in an area to include paid hours on a standard basis in order to eliminate differences that are more a matter of how hours are reported rather than actual difference in wages.

Those commenters who opposed the exclusion of paid lunch hours were generally concerned that hospitals do not currently track paid lunch hours. They indicated that it would be a major burden for hospitals to change their systems to accommodate reporting the hours and the benefits are likely to be minimum.

A few commenters suggested that, if a hospital pays its employees at the full rate for military and jury duty leave, the full associated hours should be included. However, they added that if a hospital pays its employees at a reduced rate for these leave categories, the hospital should exclude hours based on the fraction of the salary that is not paid. If the hospital does not pay for any military or jury duty leave, all of the associated hours should be excluded. The commenters believed that this treatment would be consistent with our longstanding policy to include hours associated with paid time off, while a hospital's average hourly rate would not be negatively impacted by the reduced rates that some hospitals pay for military and jury duty leave. One commenter recommended that CMS permit hospitals to exclude the hours, but not require such reporting.

Several commenters opposed excluding paid hours associated with military and jury duty because they believe that military and jury duty leave affect all providers in the same way. Therefore, they believed that any changes in the wage index would likely be immaterial. Two commenters expressed concern that, if paid hours are excluded and wages are not, the wage index would be overstated. The commenters recommended that, if CMS excludes paid hours associated with military and jury duty leave, for consistency, CMS should also exclude the related wages. Alternatively, the commenters recommended that CMS collect data on the wages and hours associated with military and jury duty first, so that the impact of excluding the hours can be determined before the policy is implemented. One commenter believed that CMS should only include in the wage index, hours associated with regular hours, overtime, and sick leave, because these paid leave or paid time off categories are consistently offered among hospitals. The commenter also believed other paid leave or paid time off categories such as vacation hours, maternity leave, bereavement leave, and vacation hours should be excluded because they are not consistently offered among hospitals. In addition, the commenter believed that when hospitals are competing for employees in the labor market, if offered, these paid leave or paid time off hours could vary from hospital to hospital. For example, hospital A will only pay 2 weeks for paid vacation leave, while hospital B will pay 4 weeks for paid vacation leave. Therefore, the commenter believed these other paid leave or paid time off leave hours should be excluded from the wage index.

Response: As we stated above and in the proposed rule, it has been our longstanding policy to include paid hours in the calculation of the wage index because they more appropriately reflect a hospital's total wage costs. We solicited comments on the possible exclusion of paid lunch hours and proposed to exclude the paid hours associated with military and jury duty hours because of our concern that there were significant issues with the consistent treatment of these issues across hospitals that may impact the validity of the wage index. However, the comments indicate to us there is substantial disagreement with respect to whether either category of paid hours should be excluded from the wage index calculation. Therefore, we are not proceeding with either change at this time. We intend to explore a more comprehensive assessment of the use of paid hours in a future rule. For the FY 2005 final wage index, we are including paid lunch hours, and hours associated with military leave and jury duty.

D. Verification of Wage Data From the Medicare Cost Reports

The data file used to construct the wage index includes FY 2000 data submitted to us as of June 27, 2003. As in past years, we performed an intensive review of the wage data, mostly through the use of edits designed to identify aberrant data.

We constructed the proposed FY 2004 wage index based on the wage data for facilities that were IPPS hospitals in FY 2000, even for those facilities that have terminated their participation in the program as hospitals or have since been designated as a critical access hospital (CAH), as long as those data do not fail any of our edits for reasonableness. We stated that including the wage data for these hospitals is, in general, appropriate to reflect the economic conditions in the various labor market areas during the relevant past period.

Prior to the proposed rule, we had received correspondence suggesting that the wage data for hospitals that have subsequently been redesignated as CAHs should be removed from the wage index calculation because CAHs are a separate provider type and are unique compared to other short-term, acute care hospitals. CAHs are limited to only 15 acute care beds. An additional 10 beds may be designated as swing-beds, but only 15 beds can be used at one time to serve acute care patients. CAHs tend to be located in isolated, rural areas. In the May 19, 2003 proposed rule, we solicited comments on whether we should exclude wage data from such hospitals from the wage index calculation. However, we included the data for current CAHs in the proposed FY 2004 wage index if the CAH was paid under the IPPS during FY 2000 as an acute care hospital.

Comment: Commenters, including national hospital associations, generally supported the removal of CAH wage data from the wage index. One commenter agreed that CAHs are dissimilar to IPPS hospitals and described a situation in which including a CAH has a negative impact on the other hospitals' wage index. One commenter agreed that CMS should exclude the costs, but expressed concern about the immediate financial impact that excluding CAHs might have on all hospitals in FY 2004. The commenter recommended that CMS examine the impact of removing CAH wage data from the wage index and make this analysis available for public comment. Another commenter recommended that CMS establish a date prior to the release of the wage index public use file that the facility must be certified as a CAH to be excluded from the wage index calculation.

Several commenters opposed excluding CAH data from the wage index. Some commenters indicated that CMS does not exclude hospitals that converted to CAH status subsequent to the year used to derive DRG weights. Another commenter opposed excluding CAHs from the wage index because the commenter believed that the wage index should reflect conditions of a labor market at a specific point in time. The commenter believed that other conditions, such as closures, mergers, or expansions, are analogous circumstances and warned that excluding these hospitals would also distort the wage index. Another commenter recommended that CMS apply a hold-harmless policy.

Response: CAHs represent a substantial number of hospitals with significantly different labor costs in many labor market areas where they exist. Using data collected for the proposed FY 2004 wage index, we found that, in 89 percent of all labor market areas with hospitals that converted to CAH status some time after FY 2000, the average hourly wage for CAHs is lower than the average hourly wage for other short-term hospitals in the area. In 79 percent of the labor market areas with CAHs, the average hourly wage for CAHs is lower than the average hourly wage for other short-term hospitals by 5 percent or greater. These results suggest that the wage data for CAHs, in general, are significantly different from other short-term hospitals.

Further, we found that removing CAHs from the wage index would have a minimal redistributive effect on Medicare payments to hospitals. The majority of the labor market areas would decrease by only 0.30 percent in their wage index value. The actual payment impact would be even smaller because the wage index is applied to only the labor-related portion of the average standardized amount. Only 10 areas would experience a decrease in their wage index values greater than 0.30 percent. The greatest negative impact is 9.57 percent. Meanwhile, positive impacts occur in 48 areas, 30 of which are in rural areas. Overall, removing CAHs from the wage index would have a minimal redistributive effect on Medicare payments to hospitals.

We believe that removing CAHs from the wage index is prudent policy, given the substantial negative impact these hospitals have on the wage indexes in the areas where they are located and the minimal impact they have on the wage indexes of other areas. We note that we would continue to include the wage data for other terminating or converting hospitals for the period preceding their change in Medicare provider status, as long as those data do not fail any of our edits for reasonableness. This is because we continue to believe that the wage data for these hospitals, unlike CAHs, are not necessarily unique compared to other short-term hospitals, and these terminating or converting hospitals provide an accurate reflection of the labor market area during the relevant past period.

Therefore, beginning with the FY 2004 wage index, we are excluding from the wage index the wages and hours for all hospitals that are currently designated as a CAH, even if the hospital was paid under the IPPS during the cost reporting period used in calculating the wage index. We believe that this change improves the overall equity of the wage index. Therefore, it is important to proceed with this change for FY 2004. Consistent with our general approach to wage index changes, we are not holding other hospitals' payments harmless for this change.

As recommended, any hospital that is designated as a CAH by 7 days prior to the publication of the preliminary wage index public use file are excluded from the calculation of the wage index. Hospitals receiving designation after this date will remain in the wage index calculation.

We asked our fiscal intermediaries to revise or verify data elements that resulted in specific edit failures. The unresolved data elements that were included in the calculation of the proposed FY 2004 wage index have been resolved and are reflected in the calculation of the final FY 2004 wage index. For the final FY 2004 wage index in this final rule, we removed data for 23 hospitals that failed edits. For 9 of these hospitals, we were unable to obtain sufficient documentation to verify or revise the data because the hospitals are no longer participating in the Medicare program, are under new ownership, or are in bankruptcy status, and supporting documentation is no longer available. We identified 14 hospitals with incomplete or inaccurate data resulting in zero or negative, or otherwise aberrant, average hourly wages. Therefore, these hospitals were removed from the calculation. As a result, the final FY 2004 wage index is calculated based on FY 2000 wage data for 4,087 hospitals.

E. Computation of the FY 2004 Wage Index

The method used to compute the FY 2004 wage index follows:

Step 1—As noted above, we based the FY 2004 wage index on wage data reported on the FY 2000 Medicare cost reports. We gathered data from each of the non-Federal, short-term, acute care hospitals for which data were reported on the Worksheet S-3, Parts II and III of the Medicare cost report for the hospital's cost reporting period beginning on or after October 1, 1999 and before October 1, 2000. In addition, we included data from some hospitals that had cost reporting periods beginning before October 1999 and reported a cost reporting period covering all of FY 2000. These data were included because no other data from these hospitals are available for the cost reporting period described above, and because particular labor market areas might be affected due to the omission of these hospitals. However, we generally describe these wage data as FY 2000 data. We note that, if a hospital had more than one cost reporting period beginning during FY 2000 (for example, a hospital had two short cost reporting periods beginning on or after October 1, 1999 and before October 1, 2000), we included wage data from only one of the cost reporting periods, the longer, in the wage index calculation. If there was more than one cost reporting period and the periods were equal in length, we included the wage data from the later period in the wage index calculation.

Step 2—Salaries—The method used to compute a hospital's average hourly wage excludes certain costs that are not paid under the IPPS. In calculating a hospital's average salaries plus wage-related costs, we subtracted from Line 1 (total salaries) the GME and CRNA costs reported on lines 2, 4.01, and 6, the Part B salaries reported on Lines 3, 5 and 5.01, home office salaries reported on Line 7, and excluded salaries reported on Lines 8 and 8.01 (that is, direct salaries attributable to SNF services, home health services, and other subprovider components not subject to the IPPS). We also subtracted from Line 1 the salaries for which no hours were reported. To determine total salaries plus wage-related costs, we added to the net hospital salaries the costs of contract labor for direct patient care, certain top management, pharmacy, laboratory, and nonteaching physician Part A services (Lines 9, 9.01, 9.02, and 10), home office salaries and wage-related costs reported by the hospital on Lines 11 and 12, and nonexcluded area wage-related costs (Lines 13, 14, and 18).

We note that contract labor and home office salaries for which no corresponding hours are reported were not included. In addition, wage-related costs for nonteaching physician Part A employees (Line 18) are excluded if no corresponding salaries are reported for those employees on Line 4.

Step 3—Hours—With the exception of wage-related costs, for which there are no associated hours, we computed total hours using the same methods as described for salaries in Step 2.

Step 4—For each hospital reporting both total overhead salaries and total overhead hours greater than zero, we then allocated overhead costs to areas of the hospital excluded from the wage index calculation. First, we determined the ratio of excluded area hours (sum of Lines 8 and 8.01 of Worksheet S-3, Part II) to revised total hours (Line 1 minus the sum of Part II, Lines 2, 3, 4.01, 5, 5.01, 6, 7, and Part III, Line 13 of Worksheet S-3). We then computed the amounts of overhead salaries and hours to be allocated to excluded areas by multiplying the above ratio by the total overhead salaries and hours reported on Line 13 of Worksheet S-3, Part III. Next, we computed the amounts of overhead wage-related costs to be allocated to excluded areas using three steps: (1) we determined the ratio of overhead hours (Part III, Line 13) to revised hours (Line 1 minus the sum of Lines 2, 3, 4.01, 5, 5.01, 6, and 7); (2) we computed overhead wage-related costs by multiplying the overhead hours ratio by wage-related costs reported on Part II, Lines 13, 14, and 18; and (3) we multiplied the computed overhead wage-related costs by the above excluded area hours ratio. Finally, we subtracted the computed overhead salaries, wage-related costs, and hours associated with excluded areas from the total salaries (plus wage-related costs) and hours derived in Steps 2 and 3.

Step 5—For each hospital, we adjusted the total salaries plus wage-related costs to a common period to determine total adjusted salaries plus wage-related costs. To make the wage adjustment, we estimated the percentage change in the employment cost index (ECI) for compensation for each 30-day increment from October 14, 1999 through April 15, 2001 for private industry hospital workers from the Bureau of Labor Statistics' Compensation and Working Conditions. We use the ECI because it reflects the price increase associated with total compensation (salaries plus fringes) rather than just the increase in salaries. In addition, the ECI includes managers as well as other hospital workers. This methodology to compute the monthly update factors uses actual quarterly ECI data and assures that the update factors match the actual quarterly and annual percent changes. The factors used to adjust the hospital's data were based on the midpoint of the cost reporting period, as indicated below.

For example, the midpoint of a cost reporting period beginning January 1, 2000 and ending December 31, 2000 is June 30, 2000. An adjustment factor of 1.03786 would be applied to the wages of a hospital with such a cost reporting period. In addition, for the data for any cost reporting period that began in FY 2000 and covered a period of less than 360 days or more than 370 days, we annualized the data to reflect a 1-year cost report. Annualization is accomplished by dividing the data by the number of days in the cost report and then multiplying the results by 365.

Step 6—Each hospital was assigned to its appropriate urban or rural labor market area before any reclassifications under section 1886(d)(8)(B) or section 1886(d)(10) of the Act. Within each urban or rural labor market area, we added the total adjusted salaries plus wage-related costs obtained in Step 5 for all hospitals in that area to determine the total adjusted salaries plus wage-related costs for the labor market area.

Step 7—We divided the total adjusted salaries plus wage-related costs obtained under both methods in Step 6 by the sum of the corresponding total hours (from Step 4) for all hospitals in each labor market area to determine an average hourly wage for the area.

Step 8—We added the total adjusted salaries plus wage-related costs obtained in Step 5 for all hospitals in the nation and then divided the sum by the national sum of total hours from Step 4 to arrive at a national average hourly wage. Using the data as described above, the national average hourly wage is $24.8076.

Step 9—For each urban or rural labor market area, we calculated the hospital wage index value by dividing the area average hourly wage obtained in Step 7 by the national average hourly wage computed in Step 8.

Step 10—Following the process set forth above, we developed a separate Puerto Rico-specific wage index for purposes of adjusting the Puerto Rico standardized amounts. (The national Puerto Rico standardized amount is adjusted by a wage index calculated for all Puerto Rico labor market areas based on the national average hourly wage as described above.) We added the total adjusted salaries plus wage-related costs (as calculated in Step 5) for all hospitals in Puerto Rico and divided the sum by the total hours for Puerto Rico (as calculated in Step 4) to arrive at an overall average hourly wage of $11.5905 for Puerto Rico. For each labor market area in Puerto Rico, we calculated the Puerto Rico-specific wage index value by dividing the area average hourly wage (as calculated in Step 7) by the overall Puerto Rico average hourly wage.

Step 11—Section 4410 of Public Law 105-33 provides that, for discharges on or after October 1, 1997, the area wage index applicable to any hospital that is located in an urban area of a State may not be less than the area wage index applicable to hospitals located in rural areas in that State. Furthermore, this wage index floor is to be implemented in such a manner as to ensure that aggregate IPPS payments are not greater or less than those that would have been made in the year if this section did not apply. For FY 2004, this change affects 150 hospitals in 49 MSAs. The MSAs affected by this provision are identified by a footnote in Table 4A in the Addendum of this final rule.

Comment: One commenter indicated that there are serious deficiencies in the payment rates to Iowa hospitals under IPPS because of the development and application of the wage index, and, accordingly, CMS must make revisions to the wage index in this final rule. The comment suggested that CMS should: reduce the labor-related portion of the standardized amount to which the wage index is applied; adjust the wage index upward to account for low Medicare payments; or utilize a wage index floor or compress the wage index.

Response: We appreciate the concerns expressed by this commenter about the impact of the wage index upon Iowa's hospitals. We strive each year to ensure the wage index accurately reflects the relative wage differences across labor market areas. Further, the methodology we use to compute the wage index values is the same for all urban and rural hospitals. Therefore, the wage index values we include in the proposed and final rules for Iowa hospitals reflect the actual wage costs that are reported by these hospitals relative to those reported by hospitals across the nation.

With respect to the commenter's specific recommendations, we address comments related to the labor-related portion of the standardized amounts in section VII. of the preamble of this final rule. With respect to the other recommendations raised, these were not proposed and, therefore, we do not wish to implement them in this final rule. We are willing to explore these and other options in the future and to work with the commenter to address the concerns expressed.

Comment: One commenter indicated that we failed to address the problem associated with the exclusion of indirect patient care contract labor in the proposed rule. The commenter indicated that we recognized this problem in the FY 2002 final rule (67 FR 50022), but failed to carry out our commitment to address it.

Response: We indicated last year it would be necessary to revise the cost report form and instructions in order to collect the data necessary to separately identify the costs and hours associated with the following contracted overhead services: administrative and general; housekeeping; and dietary. In Transmittal Number 10 of the Medicare cost report, we revised Worksheet S-3, Part II to collect contract labor costs associated with these services, effective with cost reporting periods beginning on or after October 1, 2003.

We also indicated our final decision on whether to include contract indirect patient care labor costs in our calculation of the wage index will depend on the outcome of our analyses of the data collected and public comments.

F. Revisions to the Wage Index Based on Hospital Redesignation

1. General

Under section 1886(d)(10) of the Act, the Medicare Geographic Classification Review Board (MGCRB) considers applications by hospitals for geographic reclassification for purposes of payment under the IPPS. Hospitals can elect to reclassify for the wage index or the standardized amount, or both, and as individual hospitals or as rural groups. Generally, hospitals must be proximate to the labor market area to which they are seeking reclassification and must demonstrate characteristics similar to hospitals located in that area. Hospitals must apply for reclassification to the MGCRB. The MGCRB issues its decisions by the end of February for reclassification to become effective for the following fiscal year (beginning October 1). The regulations applicable to reclassifications by the MGCRB are located in §§ 412.230 through 412.280.

Section 1886(d)(10)(D)(v) of the Act provides that, beginning with FY 2001, a MGCRB decision on a hospital reclassification for purposes of the wage index is effective for 3 fiscal years, unless the hospital elects to terminate the reclassification. Section 1886(d)(10)(D)(vi) of the Act provides that the MGCRB must use the 3 most recent years' average hourly wage data in evaluating a hospital's reclassification application for FY 2003 and any succeeding fiscal year.

Section 304(b) of Pub. L. 106-554 provides that the Secretary must establish a mechanism under which a statewide entity may apply to have all of the geographic areas in the State treated as a single geographic area for purposes of computing and applying a single wage index, for reclassifications beginning in FY 2003. The implementing regulations for this provision are located at § 412.235.

Section 1886(d)(8)(B) of the Act permits a hospital located in a rural county adjacent to one or more urban areas to be designated as being located in the MSA to which the greatest number of workers in the county commute (1) if the rural county would otherwise be considered part of an urban area under the standards published in the Federal Register for designating MSAs (and for designating NECMAs), and (2) if the commuting rates used in determining outlying counties (or, for New England, similar recognized area) were determined on the basis of the aggregate number of resident workers who commute to (and, if applicable under the standards, from) the central county or counties of all contiguous MSAs (or NECMAs). Hospitals that meet these criteria are deemed urban for purposes of the standardized amounts and for purposes of assigning the wage index.

Revised MSA standards were published in the December 27, 2000 Federal Register (65 FR 82228). We are working with the Census Bureau to compile a list of hospitals that meet the new standards based on the 2000 census data; however, that work was not yet complete at the time of publication of the proposed rule.

As noted above, OMB announced the new Metropolitan and Micropolitan Statistical Area designations and definitions on June 6, 2003. These new designations have extensively revised the construct of many of the existing Metropolitan Areas and created many new designated areas. In order to implement these changes, we need to carefully evaluate the implications of these changes for each county and hospital nationwide. As a result, we are unable to incorporate these new standards for redesignating hospitals and, therefore, we are not implementing the new standards for purposes of redesignation for FY 2004 under section 1886(d)(8)(B) of the Act. As a result, to qualify for redesignation under this section in FY 2004, hospitals must be located in counties that meet the 1990 standards.

2. Effects of Reclassification

The methodology for determining the wage index values for redesignated hospitals is applied jointly to the hospitals located in those rural counties that were deemed urban under section 1886(d)(8)(B) of the Act and those hospitals that were reclassified as a result of the MGCRB decisions under section 1886(d)(10) of the Act. Section 1886(d)(8)(C) of the Act provides that the application of the wage index to redesignated hospitals is dependent on the hypothetical impact that the wage data from these hospitals would have on the wage index value for the area to which they have been redesignated. Therefore, as provided in section 1886(d)(8)(C) of the Act, the wage index values were determined by considering the following:

• If including the wage data for the redesignated hospitals would reduce the age index value for the area to which the hospitals are redesignated by 1 percentage point or less, the area wage index value determined exclusive of the wage data for the redesignated hospitals applies to the redesignated hospitals.
• If including the wage data for the redesignated hospitals reduces the wage index value for the area to which the hospitals are redesignated by more than 1 percentage point, the area wage index determined inclusive of the wage data for the redesignated hospitals (the combined wage index value) applies to the redesignated hospitals.
• If including the wage data for the redesignated hospitals increases the wage index value for the urban area to which the hospitals are redesignated, both the area and the redesignated hospitals receive the combined wage index value. Otherwise, the hospitals located in the urban area receive a wage index excluding the wage data of hospitals redesignated into the area.
• The wage data for a reclassified urban hospital is included in both the wage index calculation of the area to which the hospital is reclassified (subject to the rules described above) and the wage index calculation of the urban area where the hospital is physically located.
• Rural areas whose wage index values would be reduced by excluding the wage data for hospitals that have been redesignated to another area continue to have their wage index values calculated as if no redesignation had occurred (otherwise, redesignated rural hospitals are excluded from the calculation of the rural wage index).
• The wage index value for a redesignated rural hospital cannot be reduced below the wage index value for the rural areas of the State in which the hospital is located.

The wage index values for FY 2004 are shown in Tables 4A, 4B, 4C, and 4F in the Addendum to this final rule. Hospitals that are redesignated must use the wage index values shown in Table 4C. Areas in Table 4C may have more than one wage index value because the wage index value for a redesignated urban or rural hospital cannot be reduced below the wage index value for the rural areas of the State in which the hospital is located. Therefore, those areas with more than one wage index shown have hospitals from more than one State reclassified into them, and the rural wage index for a State in which at least one hospital is physically located is higher than the wage index for the area to which the hospital is reclassified.

Tables 3A and 3B in the Addendum of this final rule list the 3-year average hourly wage for each labor market area before the redesignation of hospitals, based on FYs 1998, 1999, and 2000 cost reporting periods. Table 3A lists these data for urban areas and Table 3B lists these data for rural areas. In addition, Table 2 in the Addendum to this final rule includes the adjusted average hourly wage for each hospital from the FY 1998 and FY 1999 cost reporting periods, as well as the FY 2000 period used to calculate the final FY 2004 wage index. The 3-year averages are calculated by dividing the sum of the dollars (adjusted to a common reporting period using the method described previously) across all 3 years, by the sum of the hours. If a hospital is missing data for any of the previous years, its average hourly wage for the 3-year period is calculated based on the data available during that period.

Table 9 in the Addendum of this final rule shows hospitals that have been reclassified under either section 1886(d)(8) or section 1886(d)(10)(D) of the Act. This table includes hospitals reclassified for FY 2004 by the MGCRB (68 for wage index, 31 for the standardized amount, and 34 for both the wage index and the standardized amount), as well as hospitals that were reclassified for the wage index in either FY 2002 (451) or FY 2003 (55) and are, therefore, in either the second or third year of their 3-year reclassification. In addition, it includes rural hospitals redesignated to an urban area under section 1886(d)(8)(B) of the Act for purposes of the standardized amount and the wage index (42). Since publication of the May 19 proposed rule, the number of reclassifications has changed because some MGCRB decisions were still under review by the Administrator and because some hospitals decided to withdraw their requests for reclassification.

Changes to the wage index that result from withdrawals of requests for reclassification, wage index corrections, appeals, and the Administrator's review process have been incorporated into the wage index values published in this final rule. The changes may affect not only the wage index value for specific geographic areas, but also the wage index value redesignated hospitals receive; that is, whether they receive the wage index value that includes the data for both the hospitals already in the area and the redesignated hospitals. Further, the wage index value for the area from which the hospitals are redesignated may be affected.

Applications for FY 2005 reclassifications are due to the MCGRB by September 2, 2003. We note that this is also the deadline for canceling a previous wage index reclassification withdrawal or termination under § 412.273(d). Applications and other information about MCGRB reclassifications may be obtained via the CMS Internet Web site at http://cms.hhs.gov/providers/prrb/mgcinfo.asp , or by calling the MCGRB at (410) 786-1174. The mailing address of the MGCRB is: 2520 Lord Baltimore Drive, Suite L, Baltimore, MD 21244-2670.

As noted previously, OMB announced its new Metropolitan and Micropolitan Statistical Area definitions on June 6, 2003. However, as noted previously as well as in the proposed rule, in order to implement these changes for the IPPS, it is necessary to identify the new area designations for each county and hospital in the country. This is not possible by the September 2, 2003 deadline for reclassification by the MCGRB for FY 2005. Therefore, hospitals submitting applications for reclassification by the MCGRB for FY 2005 should base those applications on the current MSAs. We plan to move deliberately in determining the implications the new definitions will have on hospitals' reclassification requests, and we are considering addressing these implications in the FY 2005 proposed rule.

G. Requests for Wage Data Corrections

In the May 19, 2003 proposed rule, we described the process for hospitals to review and revise their FY 2000 wage data. The preliminary wage data file was made available on January 10, 2003 (and subsequently on February 4, 2003), through the Internet on CMS's Web site at: http://www.cms.hhs.gov/providers/hipps/default.asp. At that time, we also made available, at the same Internet address, a file showing each MSA's and rural areas's FY 2004 average hourly wage based on data then available compared to its FY 2003 average hourly wage. In a memorandum dated December 31, 2002, we instructed all Medicare fiscal intermediaries to inform the IPPS hospitals they service of the availability of the wage data file and the process and timeframe for requesting revisions (including the specific deadlines listed below). We also instructed the fiscal intermediaries to advise hospitals that these data are made available directly through their representative hospital organizations.

If a hospital wished to request a change to its data as shown in that wage data file, the hospital was to submit corrections along with complete, detailed supporting documentation to its intermediary by February 17, 2003 (this deadline was initially announced as February 10, 2003, but was changed due to the need to repost some of the data). Hospitals were notified of this deadline and of all other possible deadlines and requirements, including the requirement to review and verify their data as posted on the preliminary wage data file on the Internet, through the December 31, 2002 memorandum referenced above.

After reviewing requested changes submitted by hospitals, fiscal intermediaries transmitted any revised cost reports to CMS and forwarded a copy of the revised Worksheet S-3, Parts II and III to the hospitals by April 4, 2003. In addition, fiscal intermediaries were to notify hospitals of the changes or the reasons that changes were not accepted. These deadlines were necessary to allow sufficient time to review and process the data so that the final wage index calculation could be completed for the development of the final FY 2004 prospective payment rates to be published by August 1, 2003.

If a hospital disagreed with the fiscal intermediary's resolution of a policy issue (for example, whether a general category of cost is allowable in the wage data), the hospital could have contacted CMS in an effort to resolve the issue. We note that the April 4, 2003 deadline also applied to these requests. Requests were required to be sent to CMS at the address below (with a copy to the hospital's fiscal intermediary). The request must have fully documented all attempts by the hospital to resolve the dispute through the process described above, including copies of relevant correspondence between the hospital and the fiscal intermediary. During review, we do not consider issues such as the adequacy of a hospital's supporting documentation, as we believe that fiscal intermediaries are generally in the best position to make evaluations regarding the appropriateness of these types of issues (which should have been resolved earlier in the process).

The final wage data public use file was released in May 2003. Hospitals had an opportunity to examine both Table 2 of the proposed rule and the May 2003 final public use wage data file (which reflected revisions to the data used to calculate the values in Table 2) to verify the data CMS used to calculate the wage index.

As with the file made available in January 2003, we made the final wage data released in May 2003 available to hospital associations and the public on the internet. However, the May 2003 public use file was made available solely for the limited purpose of identifying any potential errors made by CMS or the fiscal intermediary in the entry of the final wage data that result from the correction process described above (with the February 2003 deadline). Hospitals were encouraged to review their hospital wage data promptly after the release of the May 2003 file. Data presented at that time could not be used by hospitals to initiate new wage data correction requests.

If, after reviewing the May 2003 final file, a hospital believed that its wage data were incorrect due to a fiscal intermediary or CMS error in the entry or tabulation of the final wage data, it was provided an opportunity to send a letter to both its fiscal intermediary and CMS that outlined why the hospital believed an error existed and provided all supporting information, including relevant dates (for example, when it first became aware of the error). These requests had to be received by CMS and the fiscal intermediaries no later than June 6, 2003.

Changes to the hospital wage data were only made in those very limited situations involving an error by the intermediary or CMS that the hospital could not have known about before its review of the final wage data file. Specifically, at this stage of the process, neither the intermediary nor CMS accepted the following types of requests:

• Requests for wage data corrections that were submitted too late to be included in the data transmitted to CMS by fiscal intermediaries on or before April 4, 2003.
• Requests for correction of errors that were not, but could have been, identified during the hospital's review of the January 2003 wage data file.
• Requests to revisit factual determinations or policy interpretations made by the intermediary or CMS during the wage data correction process.

Verified corrections to the wage index received timely (that is, by June 6, 2003) are incorporated into the final wage index in the final rule to be published by August 1, 2003, and to be effective October 1, 2003.

We have created the process described above to resolve all substantive wage data correction disputes before we finalize the wage data for the FY 2004 payment rates. Accordingly, hospitals that did not meet the procedural deadlines set forth above will not be afforded a later opportunity to submit wage data corrections or to dispute the intermediary's decision with respect to requested changes. Specifically, our policy is that hospitals that do not meet the procedural deadlines set forth above will not be permitted to challenge later, before the Provider Reimbursement Review Board, the failure of CMS to make a requested data revision (See W. A. Foote Memorial Hospital v. Shalala, No. 99-CV-75202-DT (E.D. Mich. 2001), also Palisades General Hospital v. Thompson, No. 99-1230 (D.D.C. 2003)).

Again, we believe the wage data correction process described above provides hospitals with sufficient opportunity to bring errors in their wage data to the fiscal intermediaries' attention. Moreover, because hospitals had access to the final wage data by early May 2003, they had the opportunity to detect any data entry or tabulation errors made by the fiscal intermediary or CMS before the development and publication of the FY 2004 wage index in this final rule, and the implementation of the FY 2004 wage index on October 1, 2003. If hospitals avail themselves of this opportunity, the wage index implemented on October 1 should be accurate. Nevertheless, in the event that errors are identified after publication in the final rule, we retain the right to make midyear changes to the wage index under very limited circumstances.

Specifically, in accordance with § 412.63(x)(2) of our existing regulations, we make midyear corrections to the wage index only in those limited circumstances in which a requesting hospital can show: (1) that the intermediary or CMS made an error in tabulating its data; and (2) that the requesting hospital could not have known about the error or did not have an opportunity to correct the error, before the beginning of FY 2004 (that is, by the June 6, 2003 deadline.) This provision is not available to a hospital seeking to revise another hospital's data that may be affecting the requesting hospital's wage index. As indicated earlier, since a hospital had the opportunity to verify its data, and the fiscal intermediary notified the hospital of any changes, we do not expect that midyear corrections would be necessary. However, if the correction of a data error changes the wage index value for an area, the revised wage index value will be effective prospectively from the date the correction is approved.

Comment: One commenter requested that CMS release all of the assumptions used in developing the MSA average hourly wage file posted on the Internet, including the midpoint of cost reporting period adjustment factors. The commenter also requested that CMS release a file with the average hourly wage by hospital prior to the proposed rule. The commenter believed that this information would facilitate a hospital's review of its wage data.

Response: We agree that providing all of the assumptions used in calculating the wage index would be useful for hospitals and other interested parties. This year, we added to our Web site a spreadsheet that can be used to calculate a hospital's average hourly wage. Beginning with the release of the FY 2005 wage index, we will also publish on our Web site the midpoint of cost reporting period adjustment factors and a file that includes the average hourly wage for each hospital.

Comment: One commenter recommended that CMS establish a wage index list server similar to those available for the various open door forums. The list server would allow CMS to e-mail interested parties when items, such as the wage index PUF and program memoranda, are released.

Response: We currently notify all hospitals, through the fiscal intermediaries, regarding all public use files and program memorandum releases pertaining to the wage index. We also post this information on the IPPS Web site ( http://cms.hhs.gov/providers/hipps/ippswage.asp ). In addition, we make announcements regarding the wage index at the hospital open door forums. To supplement these efforts, we will also begin announcing the availability of wage index files and new program memoranda on the hospital open door forum Web site, at http://www.cms.hhs.gov/opendoor/. Those registered with the hospital open door forum list server will be automatically notified when there are announcements at this site pertaining to the wage index. Information on registering with the hospital open door forum list server is located at the open door forum Web site.

Comment: One commenter expressed concern regarding the average hourly wage calculator available on the Internet, stating that they were unable to replicate the average hourly wage published in the proposed rule for its area hospitals using the May public use file data and the online calculator.

Response: The average hourly wage values printed in the proposed rule, published on May 19, 2003 in the Federal Register, reflect the data saved in our database as of February 17, 2003. Alternatively, the May public use file was updated based on data collected through May 5, 2003. Therefore, calculating an average hourly wage using the May data could yield discrepancies between the value published in the proposed rule and the number generated by the online calculator.

H. Modification of the Process and Timetable for Updating the Wage Index

In the May 19, 2003 proposed rule, we stated that although the wage data correction process described in section III.G. of the preamble of this final rule has proven successful in the past for ensuring that the wage data used each year to calculate the wage indexes are generally reliable and accurate, we continue to be concerned about the growing volume of wage data revisions initiated by hospitals after the release of the first public use file in February. This issue has been discussed previously in the FY 1998 IPPS proposed rule (62 FR 29918) and in the FY 2002 IPPS proposed rule (66 FR 22682). In each discussion, we described the increasing number of revisions to wage data between the proposed rule and the final rule.

Currently, the fiscal intermediaries are required to conduct initial desk reviews on or before November 15 in advance of the preparation of the preliminary wage data public use file in early January (see Program Memorandum A-02-94, October 4, 2002). Furthermore, fiscal intermediaries are required to explain and attempt to resolve items that fall outside the established thresholds. This may involve further review of the supplementary documentation or contacting the hospital for additional documentation. In addition, fiscal intermediaries are required to notify State hospital associations regarding hospitals that fail to respond to issues raised during the desk review. These actions are to be completed in advance of sending the data to CMS to prepare the preliminary wage data public use file in early January. However, as we have indicated in prior Federal Register s, nearly 30 percent of hospitals subsequently request revisions to their data after the preliminary wage data file is made available.

This high volume of revisions results in an additional workload for the fiscal intermediaries. In particular, much of a fiscal intermediary's efforts prior to submitting the data to prepare the preliminary public use file may be in vain if the hospital subsequently revises all of its data prior to the early February deadline (which is the hospital's right at that point). Therefore, in the May 19 proposed rule, we proposed to modify the process to release the preliminary wage data file prior to requiring the fiscal intermediaries to conduct their initial desk reviews on the data. We proposed that this unaudited data would be available on the Internet by early October rather than early January. Hospitals would review this file to ensure it contains their correct data as submitted on their cost reports and request any changes by early November. At that time, the fiscal intermediaries would review the revised requests and conduct desk reviews of the data including all approved changes.

Under the proposed revised timetable, the fiscal intermediaries would notify the hospitals in early February of any changes to the wage data as a result of the desk reviews and the resolution of the hospitals' early November change requests. The fiscal intermediaries would also submit the revisions to CMS in early February. Hospitals would then have until early March to submit requests to the fiscal intermediaries for reconsideration of adjustments made by the fiscal intermediaries as a result of the desk review. Other than requesting reconsideration of desk review adjustments, hospitals would not be able to submit new requests for additional changes that were not submitted by early November. By early April, the fiscal intermediaries would notify all hospitals of their decisions regarding the hospitals' requests to reconsider desk review adjustments and submit all of the revised wage data to CMS. From this point (early April) until the publication of the final rule, the process would be identical to the current timetable. Similar to the current timetable, hospitals would also have the opportunity in early April to request CMS consideration of policy disputes.

Therefore, we proposed to revise the schedule to improve the quality of the wage index by initiating hospitals' review of their data sooner and allowing the fiscal intermediaries to focus their reviews on the final data submitted by hospitals to be included in the wage index. In addition, we would receive the revised data in time to incorporate them into the wage indexes published in the proposed rule, resulting in fewer changes from the proposed rule to the final rule. This will improve the ability of hospitals to assess whether they should request a withdrawal from a MGCRB reclassification. Because the decision of whether to withdraw a wage index reclassification must be made prior to publication of the final rule, the proposed schedule should decrease the likelihood that the final wage index will be dramatically different from the proposed wage index.

Comment: Commenters stated their appreciation of the desire to expedite the process and reduce the workload of its fiscal intermediaries, but some were concerned about the additional workload these timeframes would place on hospitals.

Some commenters were concerned about the 30-day review period for the hospitals, stating it would not be enough time to conduct a thorough and complete review of the detailed data, adding that a 45-day comment period should be the minimum review time for providers. Commenters also stated their concerns about adjusting to a new timetable while also collecting and submitting occupational mix data, and the possible adoption of the new MSA definitions for the FY 2005 wage index. They believe any changes to the timeline should be postponed until the FY 2006 wage index.

Other commenters were concerned about the additional workloads for hospitals whose fiscal year ends on June 30. These hospitals would most likely be preparing cost reports for the fiscal year just ended and this would be an additional burden. Another commenter expressed concern that the proposed rule did not mention the State hospital association notification for hospitals failing desk review edits and that the new deadlines would not afford hospitals any recourse to ensure accurate data. One commenter cited the major role its fiscal intermediary played in the delay of revisions to its wage index.

Several other commenters generally supported the proposal to modify the wage index timetable, but with some modification. The commenters asked that hospitals have 75 days from the proposed October release of the public use file to submit revised data to the fiscal intermediaries and that CMS finalize the timetable in June rather than waiting until the final rule is published. The commenters believed this would allow virtually all hospitals the time they need to do a thorough and complete review to determine the accuracy of the detail data needed to compute an accurate wage index. Commenters also believed this would give fiscal intermediaries time to respond to hospital issues raised during the desk review period.

Finally, other commenters expressed support for the timetable changes. These commenters believed the hospitals will have more time to review their wage data and there will be less of an administrative burden on fiscal intermediaries. Another commenter believed auditors' and hospitals' resources will be better utilized and this could help eliminate the problem of reauditing wage index data after revisions are submitted. Another commenter added that hospitals would be able to better determine how they compare to other hospitals and whether a reclassification would be appropriate using much more accurate data. Also, aberrant data would become more apparent earlier in the process.

Response: Although hospitals will be required to review the data sooner, they are not being asked to perform any more reviews or work than currently. Therefore, we do not believe this change will be burdensome to hospitals. Hospitals will still have sufficient time to complete a thorough review of the data, because the data for the FY 2005 wage index values will be taken from cost reporting periods beginning during FY 2001. These cost reports should have already been thoroughly reviewed before being submitted to their fiscal intermediary and sent to CMS earlier this year.

Further, since the ultimate goal is improvement of the wage index, we believe this will be achieved with a more streamlined process in which fiscal intermediary work is not duplicated and is instead focused on the final data submitted by hospitals instead of preliminary data, of which nearly 40 percent ends up being revised under the current timetable. As noted above, these revisions under the current process often nullify the desk reviews performed by the fiscal intermediary.

We recognize the commenters' concern with respect to the interaction of this process with the collection of occupational mix data and the potential adoption of OMB's new MSA definitions. As we proceed with developing the details of the occupational mix data collection for the FY 2005 wage index, we intend to schedule that collection effort in a way that accommodates this revised timetable. The details of that schedule will be forthcoming shortly.

Finally, as previously discussed, the ability of hospitals to assess whether they should request a withdrawal from a MGCRB reclassification will also be improved, thereby decreasing the likelihood that the final wage index will be dramatically different from the proposed wage index. For these reasons, we are adopting as final the proposed revisions to the wage data development timeline and will use the revised timeline for the development of the FY 2005 wage index.

However, in order to address commenter concerns about the 30-day review period being too short, we are modifying the timetable to have the preliminary public use file on the CMS Web site in mid-September, thereby giving hospitals approximately 45 days instead of 30 days to review the preliminary wage data. Further instructions and a detailed timeline will be released in the form of a Program Memorandum.

The following table illustrates the timetable that will be applicable for the development of the FY 2005 wage index:

IV. Other Decisions and Changes to the IPPS for Operating Costs and GME Costs

A. Transfer Payment Policy (§ 412.4)

Existing regulations at § 412.4(a) define discharges under the IPPS as situations in which a patient is formally released from an acute care hospital or dies in the hospital. Section 412.4(b) defines transfers from one acute care hospital to another, and § 412.4(c) defines transfers to certain postacute care providers. Our policy provides that, in transfer situations, full payment is made to the final discharging hospital and each transferring hospital is paid a per diem rate for each day of the stay, not to exceed the full DRG payment that would have been made if the patient had been discharged without being transferred.

The per diem rate paid to a transferring hospital is calculated by dividing the full DRG payment by the geometric mean length of stay for the DRG. Based on an analysis that showed that the first day of hospitalization is the most expensive (60 FR 45804), our policy provides for payment that is double the per diem amount for the first day (§ 412.4(f)(1)). Transfer cases are also eligible for outlier payments. The outlier threshold for transfer cases is equal to the fixed-loss outlier threshold for nontransfer cases, divided by the geometric mean length of stay for the DRG, multiplied by the length of stay for the case, plus one day.

1. Transfers to Another Acute Care Hospital (§ 412.4(b))

Medicare adopted its IPPS transfer policy because, if we were to pay the full DRG payment regardless of whether a patient is transferred or discharged, there would be a strong incentive for hospitals to transfer patients to another IPPS hospital early in their stay in order to minimize costs while still receiving the full DRG payment. The transfer policy adjusts the payments to approximate the reduced costs of transfer cases.

Currently, when a patient chooses to depart from a hospital against the medical opinion of treating physicians, the case is treated as a left against medical advice (LAMA) discharge and coded as discharge status “07-Left Against Medical Advice (LAMA)” on the inpatient billing claim form. Because, by definition, LAMA discharges are assumed not to involve the active participation of the hospital administration, our policy has been to treat LAMA cases as discharges. This policy applies even if the patient is admitted to another hospital on the date of the LAMA discharge. Consequently, we currently make a full DRG payment for any discharge coded as a LAMA case.

However, we are concerned that some hospitals may be incorrectly coding transfers as LAMA cases. The Office of Inspector General (OIG) issued a report in March 2002 (A-06-99-00045), asserting that of the approximately 60,000 LAMA discharges annually, 1,500 patients were subsequently admitted to another IPPS hospital the same day. The OIG performed a detailed review of the medical records at selected hospitals and found evidence that the hospitals actively participated in transferring the patients to a different IPPS hospital, yet the hospital coded the claim as a LAMA. OIG cited several examples of these cases:

“In the first example, the transferring hospital did not have an inpatient room available for the patient, who had been in the emergency room for 24 hours. The medical record showed that the treating physician contacted another PPS hospital to determine whether the hospital could accept the patient. Specifically, the medical record stated: ‘Upon request of the patient, [hospital name] was contacted since there is a good possibility of transferring patient to [name of hospital]. At present, he has been in emergency room for 24 hours waiting for a bed.’ ”

In this example, despite the overt participation of the physician in securing the admission to the other IPPS hospital and the fact that the transferring hospital did not have an inpatient room available for the patient, the claim was submitted as a LAMA discharge, rather than as a transfer to another IPPS hospital.

“In the second example, the patient was brought to the first hospital by ambulance. Subsequently, the patient's family indicated that they wanted a neurologist at another hospital to render the treatment needed by the patient. The attending physician contacted the neurologist in order to determine if the neurologist would accept, admit, and treat the patient. The medical record contained ample evidence of knowledge and participation of the transferring hospital, and the discharge should have been reported as a PPS transfer. Specifically, the medical record stated: ‘Patient's family wanted to sign the patient out against medical advice and take her to [name of hospital]. The physician spoke with the neurologist at [name of hospital], who agreed to accept the patient. The patient's family signed the patient discharged against medical advice. All the risks of self-discharge were explained.’ ”

In this case, although the medical record indicated the patient wanted to leave against medical advice, there is also evidence that the patient's attending physician at the hospital participated in the transfer to another IPPS hospital. While we do not wish to discourage such participation and cooperation in cases where a transfer occurs, this situation would seem almost indistinguishable from other transfer situations. For instance, we have long recognized situations where patients are transferred from a rural hospital to an urban hospital for a surgical procedure, then back to the rural hospital to complete the recuperative care, as appropriate transfer situations as long as the transfers are medically appropriate. In such a case, the rural hospital would receive a payment under the transfer policy for the first portion of the stay, the urban hospital would also receive payment under the transfer policy for the care it provided, and the rural hospital would receive a full DRG payment as the discharging hospital for the recuperative care it provided upon the patient's return from the urban hospital. In such situations, each portion of the stay may be assigned a different DRG.

Therefore, in the May 19, 2003 proposed rule, we proposed to expand our definition of a transfer under § 412.4(b) to include all patients who are admitted to another IPPS hospital on the same day that the patient is discharged from an IPPS hospital, unless the first (transferring) hospital can demonstrate that the patient's treatment was completed at the time of discharge from that hospital. In other words, unless the same-day readmission is to treat a condition that is unrelated to the condition treated during the original admission (for example, the beneficiary is in a car accident later that day), any situation where the beneficiary is admitted to another IPPS hospital on the same date that he or she is discharged from an IPPS hospital would be considered a transfer, even if the patient left against medical advice from the first hospital.

Although we considered proposing a policy that would be based on whether the hospital actively participated in the transfer, and exempting from the transfer definition cases where the hospital had absolutely no knowledge that the patient intended to go to another hospital, we did not propose such a policy for two reasons. First, it would be difficult to administer equitably a policy that required a determination as to whether the hospital or the physician had knowledge of the patient's intentions. Such a policy would require fiscal intermediaries to make a difficult judgment call in many cases. Second, if we were to base the determination of whether a case is a transfer on the level of involvement of the hospital and the physician caring for the patient, we would be creating a financial disincentive to hospitals for ensuring an efficient and cooperative transfer once a decision has been made by the patient or the patient's family to leave the hospital.

We recognize that, in some cases, a hospital cannot know the patient will go to another hospital. However, we note the claims processing system can identify cases coded as discharges where the date of discharge matches the admission date at another hospital. In these cases, the fiscal intermediary will notify the hospital of the need to submit an adjustment claim. However, if the hospital can present documentation showing that the patient's care associated with the admission to the hospital was completed before discharge, consistent with our current policy, the transfer policy will not be applied.

Comment: Commenters opposed the proposed expansion of the transfer policy to include all patients who are admitted to another IPPS hospital on the same day that the patient is discharged from an IPPS hospital. They argued that situations in which a limited number of hospitals are abusing the payment rules should be handled by review of those hospitals' claims, and not through a policy change that will place additional burdens on all hospitals.

Response: We disagree that this policy expansion would create an additional burden on all hospitals. We note that it is our current policy to consider patients discharged from one IPPS hospital and admitted to another IPPS hospital on the same day as a transfer in all situations except LAMA situations, unless the original discharging hospital can document that the discharge was appropriate and unrelated to the subsequent same-day admission. We understand from the OIG that these situations are extremely rare, and in the vast majority of cases, same-day readmissions to another hospital are, in fact, transfers.

Our proposal would merely extend this current policy to LAMA situations. As is the case under our present policy, we believe it will be exceedingly rare that a patient leaves one hospital in LAMA status, and is readmitted to a second hospital on the same day for an unrelated purpose. Because the need for a hospital to supply documentation would only arise in these rare situations, we do not believe this policy change creates an additional burden for hospitals.

In relation to the appropriateness of a general policy expansion as opposed to a review and adjustment of individual hospital's claims, we believe a general policy expansion is necessary in this circumstance. As described in the proposed rule and above in this final rule, we considered proposing a policy that would be based on whether the hospital actively participated in the transfer and that would exempt from the transfer definition cases in which the hospital had absolutely no knowledge that the patient intended to go to another hospital. However, we did not propose such a policy because it would require a determination as to whether the hospital or the physician had knowledge of the patient's intentions. We believed that if we adopted such a policy, we would be creating a financial disincentive to hospitals for ensuring an efficient and cooperative transfer once a decision has been made by the patient or the patient's family to leave the hospital.

Comment: Several commenters wrote that CMS was overreacting to anecdotal examples and that the proposed policy was “not sustainable under any application of reasonableness.” They suggested that, rather than put the burden on all hospitals due to the abuse from these isolated incidents, hospitals should be evaluated from the frequency of LAMA discharges. Those that fall outside of the “norm” could be investigated, similar to the outlier studies.

Response: We agree that the problems uncovered in the OIG's report on transfers reported as LAMAs are relatively small within the overall scope of the IPPS. In fact, we made the point to OIG in our comments on a draft of its report that their findings equated with one inappropriate LAMA discharge per hospital per year. However, the OIG found this problem was not spread equally across all hospitals, but occurred disproportionately in a small number of hospitals.

We believe we are establishing clear and unequivocal policies for handling those situations that do occur and that this policy change will have a minimal impact on the majority of hospitals nationwide. Consequently, we are finalizing the change to our regulations to expand our definition of a transfer under § 412.4(b) to include all patients who are admitted to another IPPS hospital on the same day that the patient is discharged from an IPPS hospital, unless the first (transferring) hospital can demonstrate that the patient's treatment was completed at the time of discharge from that hospital, effective for discharges occurring on or after October 1, 2003.

Comment: Commenters stated that the proposed expanded definition of a transfer provides no guidance to hospitals as to what would be acceptable documentation that the patient's treatment was completed at the time of discharge. Some commenters asked whether an exact match of the principal diagnoses codes for the two admissions would be used to determine that the same-day readmission was related to the prior discharge. One commenter suggested that it would be more appropriate for the fiscal intermediary to request medical documentation from both hospitals involved in the transfer in order to determine whether a transfer payment should be made to the transferring hospital, rather than solely requesting documentation from the transferring hospital.

Another commenter asserted that CMS is placing the burden of correcting this situation on all hospitals rather than directing fiscal intermediaries to develop screens to identify these cases. In addition, they noted possible conflicts of sharing information between hospitals regarding patient care due to new HIPAA requirements.

Response: We anticipate the documentation necessary to establish that the readmission was unrelated to the prior, same-day discharge would be similar to the type of documentation relied upon by fiscal intermediaries and Quality Improvement Organizations (QIOs) to evaluate whether patients were discharged prematurely. (For example, section 4135 of the Peer Review Manual discusses discharge review.) That is, there are existing practices for determining that patients were medically unstable at discharge or the discharge was inconsistent with the patient's need for continued acute inpatient hospitalization. Therefore, there should be no breach in HIPAA disclosure requirements.

We are developing claims processing systems edits to more accurately identify transfers that are inappropriately coded as discharges. These edits identify claims that are entered with inappropriate discharge codes and will prevent payment to the second hospital if there is already a discharge from another hospital in the system for the same beneficiary on the same day. If this situation occurs, the claim from the first hospital is sent back to the hospital for correction, and the second claim is paid. We expect a similar edit that identifies same-day readmissions following a LAMA discharge would be added to the claims processing system edits.

Comment: One commenter requested clarification as to the appropriate discharge destination code in those situations when a patient left the first hospital against medical advice and the fiscal intermediary notifies this hospital of a subsequent same-day admission to another hospital.

Response: This situation is similar to those situations in which a hospital believes and intends to discharge a patient to home, but is subsequently notified that the discharge qualifies under the postacute care transfer policy because the patient received qualifying postacute care. The hospital would submit an amended bill coded to reflect the fact that the hospital now has information that the patient received subsequent care.

2. Technical Correction

Section 412.4(b)(2) defines a discharge from one inpatient area of the hospital to another area of the hospital as a transfer. Although this situation may be viewed as an intrahospital transfer, it does not implicate the transfer policy under the IPPS. In the May 19, 2003 proposed rule, to avoid confusion and to be consistent with the changes to § 412.4(b) described at section IV.A.3. of this preamble, we proposed to delete existing § 412.4(b)(2) from the definition of a transfer. We did not receive any comments on this proposal. Therefore, we are deleting existing § 412.4(b)(2) from the definition of a transfer.

3. Expanding the Postacute Care Transfer Policy to Additional DRGs (§§ 412.4(c) and (d))

Under section 1886(d)(5)(J) of the Act, a “qualified discharge” from one of 10 DRGs selected by the Secretary, to a postacute care provider is treated as a transfer case beginning with discharges on or after October 1, 1998. This section requires the Secretary to define and pay as transfers all cases assigned to one of 10 DRGs selected by the Secretary, if the individuals are discharged to one of the following postacute care settings:

• A hospital or hospital unit that is not a subsection 1886(d) hospital. (Section 1886(d)(1)(B) of the Act identifies the hospitals and hospital units that are excluded from the term “subsection (d) hospital” as psychiatric hospitals and units, rehabilitation hospitals and units, children's hospitals, long-term care hospitals, and cancer hospitals.)
• A SNF (as defined at section 1819(a) of the Act).
• Home health services provided by a home health agency, if the services relate to the condition or diagnosis for which the individual received inpatient hospital services, and if the home health services are provided within an appropriate period (as determined by the Secretary).

In the July 31, 1998 IPPS final rule (63 FR 40975 through 40976), we specified the appropriate time period during which we would consider a discharge to postacute home health services to constitute a transfer as within 3 days after the date of discharge. Also, in the July 31, 1998 final rule, we did not include in the definition of postacute care transfer cases patients transferred to a swing-bed for skilled nursing care (63 FR 40977).

Section 1886(d)(5)(J) of the Act directed the Secretary to select 10 DRGs based upon a high volume of discharges to postacute care and a disproportionate use of postacute care services. As discussed in the July 31, 1998 final rule, these 10 DRGs were selected in 1998 based on the MedPAR data from FY 1996. Using that information, we identified and selected the first 20 DRGs that had the largest proportion of discharges to postacute care (and at least 14,000 such transfer cases). In order to select 10 DRGs from the 20 DRGs on our list, we considered the volume and percentage of discharges to postacute care that occurred before the mean length of stay and whether the discharges occurring early in the stay were more likely to receive postacute care. We identified the following DRGs to be subject to the special 10 DRG transfer rule:

• DRG 14 (Intracranial Hemorrhage and Stroke with Infarction (formerly “Specific Cerebrovascular Disorders Except Transient Ischemic Attack”));
• DRG 113 (Amputation for Circulatory System Disorders Except Upper Limb and Toe);
• DRG 209 (Major Joint Limb Reattachment Procedures of Lower Extremity);
• DRG 210 (Hip and Femur Procedures Except Major Joint Procedures Age >17 With CC);
• DRG 211 (Hip and Femur Procedures Except Major Joint Procedures Age >17 Without CC);
• DRG 236 (Fractures of Hip and Pelvis);
• DRG 263 (Skin Graft and/or Debridement for Skin Ulcer or Cellulitis With CC);
• DRG 264 (Skin Graft and/or Debridement for Skin Ulcer or Cellulitis Without CC);
• DRG 429 (Organic Disturbances and Mental Retardation); and
• DRG 483 (Tracheostomy With Mechanical Ventiliation 96 + Hours or Principal Diagnosis Except Face, Mouth, and Neck Diagnoses (formerly “Tracheostomy Except for Face, Mouth, and Neck Diagnoses”)).

Similar to the policy for transfers between two acute care hospitals, the transferring hospital in a postacute care transfer for 7 of the 10 DRGs receives twice the per diem rate the first day and the per diem rate for each following day of the stay before the transfer, up to the full DRG payment. However, 3 of the 10 DRGs exhibit a disproportionate share of costs very early in the hospital stay in postacute care transfer situations. For these 3 DRGs, hospitals receive 50 percent of the full DRG payment plus the single per diem (rather than double the per diem) for the first day of the stay and 50 percent of the per diem for the remaining days of the stay, up to the full DRG payment. This is consistent with section 1886(d)(5)(J)(i) of the Act, which recognizes that in some cases “a substantial portion of the costs of care are incurred in the early days of the inpatient stay.”

Section 1886(d)(5)(J)(iv) of the Act authorizes the Secretary to expand the postacute care transfer policy beyond 10 DRGs. In the May 9, 2002 IPPS proposed rule, we discussed the possibility of expanding this policy to either all DRGs or a subset of additional DRGs (we identified 13 additional DRGs in that proposed rule) (67 FR 31455). However, as discussed further in the August 1, 2002 final rule (65 FR 50048), we did not expand the postacute care transfer provision to additional DRGs for FY 2003. The commenters on the options in the May 9, 2002 proposed rule raised many issues regarding the impact of expanding this policy that we needed to consider further before proceeding. In particular, due to the limited time between the close of the comment period and the required publication date of August 1, we were unable to completely analyze and respond to all of the points that were raised. We indicated that we would continue to conduct research to assess whether further expansion of this policy may be warranted and, if so, how to design any such refinements.

Many commenters on the May 9, 2002 proposed rule argued that, in a system based on averages, expansion of the postacute care transfer policy negatively influences, and in fact penalizes, hospitals for efficient care. They claimed that this policy indiscriminately penalizes hospitals for efficient treatment and for ensuring that patients receive the right care at the right time in the right place. They believed that the postacute care transfer provision creates an inappropriate incentive for hospitals to keep patients longer.

Commenters also expressed concern that the expansion of the transfer provision violates the fundamental principle of the IPPS. The DRG system is based on payments that will, on average, be adequate. These commenters argued that expansion of the postacute care transfer policy would give the IPPS a per-diem focus and would mean that hospitals would be paid less for shorter than average lengths of stay, although they would not be paid more for the cases that are longer than average (except for outlier cases).

We agree that the transfer policy should not hamper the provision of effective patient care. We also agree that any future expansion must consider both the need to reduce payments to reflect cost-shifting out of the acute care setting due to reductions in length of stay attributable to early transfers to postacute care and the need to ensure that payments, on average, remain adequate to ensure effective patient care. Therefore, we have assessed the extent to which the current postacute care transfer policy balances these objectives.

The table below displays the results of our analysis. We first examined whether the 10 DRGs included in the policy continue to exhibit a relatively high percentage of cases transferred to postacute care settings, particularly among cases with lengths of stay shorter than the geometric mean for the DRG (these cases would be affected by the reduced payments for transfers). The table shows that these DRGs continue to contain high percentages of cases transferred to postacute care settings similar to those we reported in the FY 1999 final rule (63 FR 40975). These results would appear to demonstrate that the postacute care transfer policy has not greatly altered hospitals' treatment patterns for these cases.

This similarity in treatment patterns is further evidenced by the fact that, for 6 of the 10 DRGs, the geometric mean length of stay has continued to decline in the 5 years since the policy was implemented. Accordingly, hospitals have continued to transfer many patients in these DRGs before the mean length of stay, despite the transfer policy. As we stated in the July 31, 1998 final rule, the transfer provision adjusts payments to hospitals to reflect the reduced lengths of stay arising from the shift of patient care from the acute care setting to the postacute care setting (63 FR 40977). This policy does not require a change in physician clinical decisionmaking nor in the manner in which physicians and hospitals practice medicine: It simply addresses the appropriate level of payments once those decisions have been made.

With respect to whether this policy alters the fundamental averaging principles of the IPPS, we believe the current policy, which targets specific DRGs where evidence shows hospitals have aggressively moved care to postacute care settings, does not alter the averaging principles of the system. In fact, it could be said to enhance those principles because a transfer case is counted as only a fraction of a case toward DRG recalibration based on the ratio of its transfer payment to the full DRG payment for nontransfer cases. This methodology ensures the DRG weight calculation is consistent with the payment policy for transfer cases. The last column of the table below indicates that all but three of these DRGs have experienced increases in DRG weights since the policy was implemented. By reducing the contribution of transfer cases to the calculation of the DRG average charge, the relative weights (the result of dividing the DRG average charge by the national average charge per case) are higher than they would otherwise be. This is because transfers, particularly short-stay transfers, have lower total charges, on average.

We indicated in the proposed rule that we believe the current 10 DRG postacute care transfer policy appears to be appropriately balancing the objectives to reduce payments to reflect cost-shifting due to reductions in length of stay attributable to early postacute care transfers and to ensure that payments, on average, remain adequate to ensure effective patient care. Therefore, we once again undertook the analysis to identify additional DRGs to which the policy might be expanded.

However, we did not propose to expand the policy to all DRGs. Although we indicated that expanding the postacute care transfer policy to all DRGs might be the most equitable approach because a policy that is limited to certain DRGs may result in disparate payment treatment across hospitals, at this time, we believe an incremental expansion is appropriate. That is, we believe further analysis is necessary to assess whether it would be appropriate to apply a reduced payment for postacute care transfers across all DRGs. In particular, it is important to attempt to distinguish between DRGs where the care is increasingly being shifted to postacute care sites versus DRGs where some patients have always been discharged to postacute care early in the stay. It may not be appropriate to reduce payment for these latter DRGs if the base payment already reflects a similar postacute care utilization rate (for example, in these cases there would be no cost shifting).

As described below, we proposed an additional 19 DRGs, based on declining mean lengths of stay and high percentages of postacute transfers, for which an expansion of the current policy appeared warranted.

We also noted that MedPAC has conducted analysis on the current postacute care transfer policy. Most recently, in its March 2003 Report to Congress, MedPAC recommended adding 13 additional DRGs to the 10 DRGs covered under the current policy (page 46). The 13 DRGs were the same DRGs included in one of our proposals to expand the postacute care transfer policy in last year's IPPS proposed rule. MedPAC did not recommend expanding the policy to include all DRGs at this time, noting that this expansion might reduce payments to some hospitals by as much as 4 percent. Rather, it suggested evaluating the impact of a limited expansion before extending the policy to more DRGs.

MedPAC's report cites several reasons for expanding the postacute care transfer policy beyond the current 10 DRGs. First, it notes the continuing shifts in services from the acute care setting to the postacute care setting. Second, the report points to different postacute care utilization for different hospitals, particularly based on geographic location. Third, the report states: “the expanded transfer policy provides a better set of incentives to protect beneficiaries from potential premature discharge to postacute care.” Fourth, MedPAC notes that the policy improves payment equity across hospitals by: reducing payments to hospitals that transfer patients to postacute care while making full payments to hospitals that provide all of the acute inpatient services in an acute care setting; and maintaining more accurate DRG weights that reflect the true resource utilization required to provide the full course of acute inpatient care, as distinguished from the partial services provided to patients who are transferred to postacute care.

Since the publication of last year's rule, we have conducted an extensive analysis to identify the best method by which to expand the postacute care transfer policy. Similar to the analysis used to identify the current 10 DRGs, in the May 19, 2003 proposed rule, we proposed to identify DRGs with high postacute care transfer rates and at least 14,000 transfer cases. However, rather than ranking DRGs on the basis of the percentage of all postacute care transfers, we proposed to rank DRGs on the basis of the percentage of postacute care transfers occurring before the DRG geometric mean length of stay. This is because only transfers that occur before the geometric mean length of stay, minus one day due to the policy that hospitals receive double the per diem for the first day, are impacted by the transfer policy. In order to focus on those DRGs where this policy would have the most impact, we proposed to include only DRGs where at least 10 percent of all cases were transferred to postacute care before the geometric mean length of stay. (We note that preceding sentence was stated incorrectly in the proposed rule. The criterion should have read “at least 10 percent of all cases that were transferred to postacute care were transferred before the geometric mean length of stay.”) The next proposed criterion is to identify DRGs with at least a 7-percent decline in length of stay over the past 5 years (from FY 1998 to FY 2003). This criterion would focus on those DRGs for which hospitals have been most aggressively discharging patients sooner into postacute care settings. Finally, we proposed to include only DRGs with a geometric mean length of stay of at least 3 days because the full payment is reached on the second day for a DRG with a 3-day length of stay.

Using these criteria, we proposed 19 additional DRGs to include in the postacute care transfer policy. However, some of the 13 DRGs proposed last year (and included in MedPAC's proposed expansion) were not included in the May 19, 2003 proposed rule. For example, DRGs 79 and 80 (Respiratory Infections and Inflammations Age >17 With and Without CC, respectively) were included in last year's proposed expansion but were not included in the proposed rule for FY 2004. DRGs 79 and 80 were excluded from the proposed rule because they did not exhibit a decline in length of stay of at least 7 percent over the past 5 years.

We noted that 7 of the proposed 19 DRGs are paired DRGs (that is, they contain a CC and no-CC split). Because these DRGs are paired DRGs (that is, the only difference in the cases assigned to DRG 130, for example, as opposed to DRG 131 is that the patient has a complicating or comorbid condition), we proposed to include both DRGs under this expanded policy. If we were to include only DRG 130 in the transfer policy, we believed there would be an incentive for hospitals not to include any code that would identify a complicating or comorbid condition, so that a transfer case would be assigned to DRG 131 instead of DRG 130.

Using the selection criteria described above, we proposed the following 19 DRGs to include under the postacute care transfer policy (in addition to the 10 DRGs already subject to the policy).

We proposed to revise § 412.4(d) to incorporate these additional 19 DRGs as qualifying DRGs for transfer payments and to make a conforming change to § 412.4(c).

We also examined whether any of these DRGs would qualify for the alternative payment methodology of 50 percent of the full DRG payment plus the per diem for the first day of the stay, and 50 percent of the per diem for the remaining days of the stay, up to the full DRG payment specified in existing regulations under § 412.4(f). To identify the DRGs that might qualify, we compared the average charges for all cases with a length of stay of 1 day to the average charges of all cases in a particular DRG. To qualify for the alternative methodology, we indicated that the average charges of 1-day discharge cases must be at least 50 percent of the average charges for all cases in the DRG.

Based on this analysis, we determined that 5 out of the proposed 19 DRGs would qualify for this payment method (DRGs 25, 122, 131, 297, and 321). However, the fact that the average charges of 1-day stays equal at least 50 percent of the average charges for all cases in these DRGs is due to the very short lengths of stay for these DRGs. Therefore, we did not propose to include them in the alternative payment methodology. For example, for a DRG with a 3-day geometric mean length of stay, full DRG payment will be made on the second day of the stay, regardless of which payment methodology is used. Therefore, in the May 19, 2003 proposed rule, we proposed that none of the 19 additional DRGs that we were proposing to add to the postacute care transfer policy would be paid under the alternative payment methodology.

We also analyzed the 10 DRGs that are currently subject to the postacute care transfer policy. Of the three DRGs that are receiving payments under the special payment (transfers after 1 day incur charges equal to at least 50 percent of the average charges for all cases). Unlike the five DRGs that would otherwise meet this criterion, the geometric mean length of stay of both DRG 209 and 211 is over 4 days. In addition, DRG 210 is currently paid under the special payment methodology, but our current analysis indicates average charges for 1-day stays are less than 50 percent of the average charges for all cases in the DRG. Nonetheless, DRG 210 is paired with DRG 211, which meets the criteria. Therefore, we proposed that DRG 210 would continue to be paid under the special payment methodology. Similar to our rationale for including both paired DRGs when one qualifies for inclusion in the postacute care transfer policy, we proposed to include both DRGs in this pair under the special payment methodology. Accordingly, we proposed that only DRGs 209, 210, and 211 that are currently paid under the alternative transfer payment methodology would continue to be paid under this methodology.

Finally, we noted that the OIG has prepared several reports that examined hospitals' compliance with proper coding of patients' discharge status as transferred under our guidelines, and has found substantial noncompliance leading to excessive payments. Specifically, the OIG found hospitals submitting claims indicating the patient had been discharged when, in fact, the patient was transferred to a postacute care setting. As we indicated in the May 8, 1998 Federal Register (63 FR 25593), hospitals found to be intentionally engaging in such practices may be investigated for fraudulent or abusive billing practices. We intend to work with the OIG to develop the most appropriate response to ensure all hospitals are compliant with our guidelines.

Comment: Many commenters argued that any expansion of the postacute care transfer policy, and even the policy itself, undermines clinical decisionmaking and penalizes hospitals for providing the right care at the right time and in the right setting. Commenters further argued that the policy itself violates the original premise of the IPPS, because it makes it difficult or impossible for hospitals to break-even on patients who receive postacute care after discharge. One commenter argued that hospitals lose if patients are discharged prior to the mean length of stay, and they lose if patients are discharged after the mean length of stay.

Commenters also argued the postacute care transfer policy is not good policy because it may create a perverse incentive for hospitals to increase patients' lengths of stay. One commenter expressed concern that longer lengths of stay would result from a shift in focus from per-case cost control to per-day cost control. The commenter suggested that this policy sends a conflicting message to hospital administrators who have taken steps recently to reduce their hospitals' average lengths of stay.

Some commenters pointed out that the postacute care transfer policy fails to acknowledge or recognize that, for many patients, postacute care is already reflected in the IPPS base payment rate for many DRGs. In particular, hospitals in certain regions of the country have historically had lower average lengths of stay, and therefore, these hospitals are disproportionately impacted by this policy.

Other commenters suggested the DRG relative weights are self-adjusting, and as patients spend less time in the acute care setting and costs decrease, the DRG relative weights will begin to fall. Therefore, there is no need for a postacute care transfer policy.

Commenters also noted the increasing costs of dealing with these higher cost cases, and that transfer payments do not adequately cover the costs of the newer and better treatment that is resulting in shorter lengths of stay. Commenters objected to the expansion of the policy due to the current financial pressure that many hospitals are currently under because of nursing shortages, inadequate Medicare payment for services they provide, and increasing costs associated with malpractice and insurance costs and increasing costs of pharmaceuticals and equipment. They also noted the financial burden in preparing to treat the aging “baby boomer” generation and costs associated with emergency management preparation.

Commenters argued that many hospitals are suffering as a result of not receiving the full market basket update (accounting for inflation each year), and further expansion of the postacute care transfer policy will further limit their resources. In addition, they argued, Congress already addresses the issues of shorter lengths of stay when it determines the market basket update each year. In effect, they claimed, hospitals whose lengths of stay decline significantly are not praised, but penalized—twice—for their efforts to provide better care. One commenter wrote to “respectfully submit that to deal with fraudulent providers in this sweeping manner is inconsistent and inappropriate.”

Response: We disagree that the postacute care transfer policy is contrary to the fundamental theory of the IPPS. Concern that hospitals would shift a portion of the acute care services to other providers in response to the incentives of the IPPS has been an ongoing concern. In fact, in response to a comment during the first year of the IPPS on the hospital-to-hospital transfer policy, we stated that “(t)he rationale for per diem payments as part of our transfer policy is that the transferring hospital generally provides only a limited amount of treatment. Therefore, payment of the full prospective payment rate would be unwarranted” (49 FR 244). We also note that in its earliest update recommendations, the Prospective Payment Assessment Commission (a predecessor to MedPAC) included what it called a site-of-service substitution adjustment to account for the shifting of portions of inpatient care to other settings.

We disagree that the postacute care transfer policy creates a perverse incentive to keep patients in the hospital longer than necessary. Our view is the policy simply responds to changing medical practice and addresses the appropriate level of payment once clinical decisions about the most appropriate care in the most appropriate setting have been made. The validity of this position is substantiated by the finding that the geometric mean length of stay for 6 of the 10 DRGs currently included in the policy have continued to fall since the policy was implemented.

In regard to the comment that the policy fails to recognize that the DRG base payments reflect some degree of postacute care, we note that the policy is intended to recognize that, since the implementation of the IPPS, the use of postacute care has generally increased. For many DRGs, the use of postacute care continues to increase at a high rate. However, an increase in the frequency of the use of postacute care does not, by itself, necessitate a policy response. If patients continue to receive the full course of acute care in the IPPS setting prior to transfer, a full DRG payment is warranted. However, if patients begin to be transferred to postacute care settings to receive care that, during the IPPS base period, was provided in the IPPS setting, paying a full DRG would not be appropriate because some of the care on which the full DRG payment is based is now being provided in the postacute care setting.

This shift in the setting where care is provided is not accounted for through DRG recalibration. During recalibration, reductions in the relative weights of certain DRGs result in increases in the weights of other DRGs. Therefore, there is no net reduction in the IPPS payments to hospitals, even though some of the care that used to be provided in the acute inpatient setting is now provided elsewhere.

Comment: Commenters took issue with our evaluation of the impact of the postacute care transfer policy on the averaging aspects of the IPPS if the policy were expanded. Pointing to our statement in the August 1, 2002 Federal Register that we intended to undertake a more comprehensive analysis of this issue, some commenters stated that we did not provide such a comprehensive analysis or include a discussion of the topic in the proposed rule.

However, other commenters expressed appreciation for our analysis of the impacts of the existing policy in the proposed rule. One commenter noted that we had made some interesting and potentially valid points that an expanded transfer policy would eliminate or reduce some of the problems caused by making national average payments to all hospitals, regardless of treatment patterns and patient-mix within specific DRGs (although this commenter suggested that we address the payment inequities caused by expensive short-stay cases, or “inliers”).

Several commenters noted that the recalculation of weights in the affected DRGs is unfair because, in the system of averages, transfers are accounted for as only partial cases but the remaining cases are not adjusted upward. The commenter wrote: “[i]f a DRG's length of stay is declining, doesn't that suggest recalibration of the relative weight?” The commenter believed inclusion of reduction in length of stay criteria “begs the question of what is the true average length of stay for these particular DRGs. If these DRGs are experiencing a large percentage of cases transferred prior to the average length of stay, it logically follows that the average length of stay would be less.”

Response: We regret that commenters perceived that we neglected to address this important issue. Our point in evaluating the DRG relative weights for the 10 DRGs that are currently included in the policy was to make the point that reducing the contribution of transfer cases in the DRG relative weight recalibration enhances the averaging mechanism for these DRGs. By treating transfer cases as less than a full discharge (reducing the denominator), we effectively inflate the charges (the numerator) to reflect the higher charges that would have occurred if the patient had been transferred. This increases, rather than decreases, the average charges (and thus the relative weights) for the affected DRGs.

For example, the DRG weights for each of these 10 DRGs declined over the 5-year period (FYs 1993 through 1998) immediately preceding the implementation of this policy. However, as shown in the table above, the DRG weights for all but three of these DRGs have increased during the 5-years since implementation of this policy. Payments for all cases in these DRGs were declining as the number of cases being transferred to postacute care increased and the average length of the inpatient acute stay decreased. However, since implementation of the policy, payments for the cases that are not implicated under this policy are rising in most of the 10 DRGs. In those DRGs where the relative weight has declined in over the 5-year period since implementation of this policy, the geometric mean length of stay has continued to decline.

As discussed above, the premise of the postacute care transfer policy is that hospitals have shifted some of the acute care formerly provided in the hospital into the postacute care setting. This distorts the averaging principle of the IPPS because the average case is now less expensive without a corresponding adjustment to the base rate. However, a high percentage of postacute care utilization by cases in a particular DRG does not, by itself, create a distortion, if the high postacute care utilization was also reflected in the calculation of the base rate.

Therefore, to ensure that any proposed expansion of the postacute care transfer policy did not improperly distort the averaging principles of the IPPS, we evaluated the change in the mean lengths of stay for the DRGs we proposed to add to the policy to identify those in which the high postacute care utilization is resulting in shorter lengths of stay and lower costs. These shorter stays represent a shift in the site (and costs) of care relative to the base period, and, thus, a distortion in the averaging principle of the IPPS.

Comment: Several commenters argued that the postacute care transfer policy is no longer necessary, as lengths of stay have stabilized and Medicare spending on postacute care has slowed. In particular, commenters pointed to the transition of postacute care provider types to prospective payment systems, which reduces the incentives for postacute care providers to agree to admit very sick patients from an acute care hospital. One commenter argued that the concept of duplicate payment for the same care is a misconception when both the acute and the postacute care providers are paid under a prospective payment system.

Commenters claimed the policy puts an undue burden on them to be required to track patients after they are discharged to another setting. They claimed this creates an “unworkable” situation for them by making hospitals track patients and requiring frequent payment and claim readjustments. They noted the relatively small payment impact for all hospitals (only 0.2 percent) compared to the administrative burden hospitals will incur to administer the expansion of the policy.

Response: We agree that postacute care providers are likely to be less willing to admit very sick patients under prospective payment systems than they were under cost reimbursement payment methodologies. However, the incentives for acute care hospitals to reduce costs by transferring patients to a postacute care setting remain as strong as ever. Furthermore, duplicate payments would still exist if the acute care hospital is shifting costs for which it is paid under the IPPS to a postacute care provider; that is, receiving payment for the care under a prospective payment system (potentially at a rate even higher than its costs). Therefore, we believe there is still a need for the postacute care transfer policy, despite the adoption of prospective payment systems for most postacute care providers under Medicare. Similarly, it is appropriate to evaluate the need to expand the policy.

Comment: Commenters suggested that, under our proposed criterion for selecting additional DRGs to cover under the policy, we should apply the same criteria to the existing postacute care transfer DRGs as to the new proposed DRGs. These commenters pointed out that 7 of the 10 DRGs would not qualify under these criteria, and should no longer be included in the policy.

One commenter argued that DRG 209 should be removed from the current list of DRGs subject to the postacute care transfer policy because the rate of decline in the average length of stay for this DRG had fallen dramatically since its inclusion in the postacute care transfer policy.

In addition, one commenter applied the proposed criteria to more recent data and determined some of the DRGs proposed to be included in the policy no longer met all the criteria. Specifically, the commenter found that 11 of the 19 DRGs proposed to be included in the transfer policy fail to meet the criterion that at least 10 percent of the postacute care transfer cases occur prior to the geometric mean length of stay.

Several commenters also noted that it appears our analysis identifying the 19 DRGs that were proposed to be added to the list included transfers from IPPS-exempt units. The commenters added that these units are not subject to the postacute care transfer policy and should not have been included in the analysis. The commenters pointed out that DRG 462 (Rehabilitation) only qualifies as a result of the inclusion of transfers from IPPS-exempt units in the analysis.

Response: We do not believe it is necessary to evaluate whether the lengths of stay for the DRGs currently included in the policy are declining. One would expect that, to the extent patients were being transferred early in the episode of care to a postacute care setting in order to minimize costs to the acute care hospital (as opposed to a general shift in the clinical care for particular cases, which is more likely to result in a continued drop in the length of stay despite the inclusion of the DRG in the transfer policy), inclusion of a particular DRG in the postacute care transfer policy would be likely to stabilize the mean length of stay for the DRG. Therefore, we did not evaluate the current DRGs included in the policy to the 7-percent decline in the length of stay criterion.

We also note that included in the commenter's list of 11 DRGs that it claim did not meet the new criteria, 6 of these DRGs are paired DRGs and were not selected based on meeting the criteria, but rather were included due to the paired nature of the DRG.

We have analyzed the remaining 5 DRGs the commenter identified as having not met the criteria that at least 10 percent of all postacute care transfer cases occur before the geometric mean length of stay. However, it appears the commenter divided the total number of transfer cases by the total number of cases in the DRG, rather than dividing by the number of postacute care transfer cases. Using the data the commenter provided to us, we found that all but l DRG met the 10 percent short-stay transfer definition we had proposed, with one DRG being a pair to another DRG that does meet the criterion.

However, we do agree with the notion that, to be included in the postacute care transfer policy, DRGs currently included in the policy should continue to meet all of the other applicable criteria. In addition, concerns from the commenters encouraged us evaluate whether the variation from year to year might also needs to be accounted for in our new criteria. Therefore, in order to improve the year-to-year stability of all the DRGs included in the policy, in this final rule, we are adding the requirement that the criteria must be met during both of the 2 most recent years for which data are available. That is, to be included in the policy, a DRG must have, for both of the 2 most recent years for which data are available:

• At least 14,000 cases postacute care transfer cases;
• At least 10 percent of its postacute care transfers occurring before the geometric mean length of stay;
• A geometric mean length of stay of at least 3 days; and
• If a DRG is not already included in the policy, a decline in its geometric mean length of stay during the most recent 5 year period of at least 7 percent.

Applying these criteria, we determined that DRG 263 no longer qualifies (there were only 13,588 postacute care transfer cases in this DRG during FY 2002). In addition, this is a paired DRG with DRG 264. Therefore, for FY 2004, we are no longer including DRGs 263 and 264 in the postacute care transfer policy.

We also corrected the programming error noted by the commenters that allowed IPPS-exempt units to be included in the analysis. Removing these units from the analysis resulted in the exclusion of some DRGs that were proposed to be included in the policy, and the inclusion of some new DRGs. The table below displays all the DRGs that met the criteria during both of the 2 most recent years available (FYs 2001 and 2002), as well as their paired-DRG if one of the DRGs meeting the criteria includes a CC/no-CC split.

Transfers to postacute care from the DRGs listed in the above table will be included under this policy, effective for discharges occurring on or after October 1, 2003. As a result of our analysis in which we applied the new qualifying criteria, we removed DRG 263 and DRG 264 from the current list of 10 DRGs, and we removed DRG 243 and DRG 462 from the proposed list of additional 19 DRGs. However, we added four new DRGs (that were not included in our proposal) to the policy based on this analysis: DRG 88 (Chronic Obstructive Pulmonary Disease); DRG 127 (Heart Failure and Shock); DRG 294 (Diabetes Age >35); and DRG 395 (Red Blood Cell Disorders, Age >17). We will review and update this list periodically to assess whether additional DRGs should be added or existing DRGs should be removed.

Comment: One commenter contested the automatic inclusion of both DRGs in a paired-DRG combination. The commenter believed any incentive for hospitals not to include a code that would identify a complicating or comorbid condition would be very limited and would have negligible effect on hospital behavior. However, the commenter asserted that if CMS is going to include both DRGs in a paired-DRG combination, CMS must combine the data for the two DRGs when applying the selection criteria.

Response: We include both DRGs from a paired-DRG combination because if we were to include only the “with CC” DRG from a “with/without CC” DRG combination in the transfer policy, there would be an incentive for hospitals not to include any code that would identify a complicating or comorbid condition. We believe our approach of identifying either DRG from a paired-DRG combination individually for inclusion in the policy is appropriate.

Comment: One commenter argued that DRG 468 should not be included in the policy because of the variation in the types of cases included in this DRG. The commenter pointed out that the cases in the DRG are, by definition, atypical, and the average lengths of stay for procedures included in this DRG vary widely. The commenter noted that “every year CMS makes changes to the list of procedures that are assigned to this DRG. Therefore, a comparison of length of stay over time is not valid because the types of cases in the DRG change every year. The criterion that length of stay must have decreased by 7 percent compared to 1998 cannot be applied to DRG 468.” The commenter added that application of a per diem payment based on a mean length of stay to a DRG that contains such a wide variety of different types of cases will result in extreme inequities.

One commenter argued for the exclusion of DRG 483 from the policy. The commenter argued that due to the large variation of lengths of stay for treatments in this DRG, the transfer policy has a very significant impact on payment for these cases that is unrelated to the use of postacute care.

Response: We disagree that DRG 468 should be excluded from the policy because of the variation in the types of cases within this DRG. Over 40 percent of transfers to postacute care within this DRG occurred before the geometric mean length of stay. Although it is true the nature of this DRG makes it difficult to assess whether there is a trend to shift care out of the acute care setting into the postacute care setting or there is just a different mix of cases being assigned to this DRG, we believe it is equitable to adjust payments for short-stay cases transferred to postacute care within this DRG. As noted above, application of this policy in the DRG recalibration process results in an overall increase in the payments for other cases in the DRG. Given the heterogeneous nature of this DRG, we believe this is appropriate.

We have addressed similar concerns in the past with respect to the inclusion of DRG 483 in this policy.

Comment: One comment noted that DRGs 121 and 122 should be included in the special payment provision due to the fact that “these cases receive the most resource intensive services within the first day of the stay due to the acute nature of a myocardial infarction * * * [including care in] intensive care units, costly IV drug infusions, and multiple tests and monitoring.”

Response: Based on the revised list of DRGs that meet the criteria as described above, we analyzed which of these DRGs qualified for the special payment methodology. The only DRGs that had charges for short-stay transfer cases on the first day of stay that were greater than 50 percent of the average charges of all cases across the DRG were DRGs 209 and 211 (71 percent and 57 percent, respectively). Because DRG 211 is paired with DRG 210, we included DRG 210 in the payment policy as well (our analysis showed that short-stay transfer cases had 40 percent of costs on the first day of the stay compared to costs for all cases across the DRG). However, DRGs 121 and 122 did not meet the 50 percent threshold.

Comment: Commenters again noted their objection to the expansion of the policy to all DRGs, even though we did not propose to expand the policy to all DRGs at this time. They refer to the language in section 1886(d)(J) of the Act that states that only those DRGs that have a “high volume of discharges” and “disproportionate use of post discharge services” could be included in an expanded postacute care transfer policy. Since this language would not apply to many DRGs, it makes this possibility “implausible.”

Commenters also argue that, since we admit we need to do further analysis before expanding the policy to all DRGs, it is unclear why we do not need to conduct further analysis to make an incremental expansion.

Response: As noted previously, we did not propose to expand this policy to all DRGs because, for some DRGs, it may not be appropriate to reduce payment for these DRGs if the base payment already reflects a similar postacute care utilization rate. For the 29 DRGs included in the policy effective October 1, 2003, we have determined the data indicate there is substantial utilization of postacute care early in the stay, leading to decreasing lengths of stay.

Comment: Other commenters noted that, if we were focusing our efforts on analyzing lengths of stay in this manner, we should redirect our focus instead on a more thorough analysis of length of stay in particular regions to determine if changes are being adequately reflected in the yearly updates.

Response: We recognize that lengths of stay have tended to vary by region, and that regions with shorter lengths of stay tend to also have lower average costs due to the fewer number of days that patient spend in the hospitals. One of the reasons for the variation is the greater reliance on postacute care earlier in the stay in those areas with lower average lengths of stay.

We do not believe it would be appropriate to base the transfer payment methodology on regional average lengths of stay. The national standardized amounts, which apply across all regions, reflect costs and lengths of stay across all regions. To the extent hospitals in one area of the country are transferring patients early in the course of their treatment while hospitals in another part of the country are providing the entire treatment in the acute care hospital, adjusting payments for those hospitals transferring patients early in the stay and reflecting this in the process of recalibration maintains full DRG payments for hospitals in areas of the country providing the full course of treatment in the acute care hospital.

B. Rural Referral Centers (§ 412.96)

Under the authority of section 1886(d)(5)(C)(i) of the Act, the regulations at § 412.96 set forth the criteria that a hospital must meet in order to qualify under the IPPS as a rural referral center. For discharges occurring before October 1, 1994, rural referral centers received the benefit of payment based on the other urban amount rather than the rural standardized amount. Although the other urban and rural standardized amounts are the same for discharges beginning with that date, rural referral centers continue to receive special treatment under both the DSH payment adjustment and the criteria for geographic reclassification.

Rural referral centers with a disproportionate share percentage of at least 30 percent are not subject to the 5.25 percent cap on DSH payments that is applicable to other rural hospitals (with the exception of rural hospitals with 500 or more beds). Rural referral centers are not subject to the proximity criteria when applying for geographic reclassification, and they do not have to meet the requirement that a hospital's average hourly wage must exceed 106 percent of the average hourly wage of the labor market area where the hospital is located.

As discussed in Federal Register documents at 62 FR 45999 and 63 FR 26325, under section 4202 of Pub. L. 105-33, a hospital that was classified as a rural referral center for FY 1991 is to be considered as a rural referral center for FY 1998 and later years so long as that hospital continues to be located in a rural area and does not voluntarily terminate its rural referral center status. Effective October 1, 2000, if a hospital located in what is now an urban area was ever a rural referral center, it is reinstated to rural referral center status (65 FR 47089). Otherwise, a hospital seeking rural referral center status must satisfy the applicable criteria.

One of the criteria under which a hospital may qualify as a rural referral center is to have 275 or more beds available for use (§ 412.96(b)(1)(ii)). A rural hospital that does not meet the bed size requirement can qualify as a rural referral center if the hospital meets two mandatory prerequisites (a minimum case-mix index and a minimum number of discharges) and at least one of three optional criteria (relating to specialty composition of medical staff, source of inpatients, or referral volume) (§ 412.96(c)(1) through (c)(5)). (See also the September 30, 1988 Federal Register (53 FR 38513).) With respect to the two mandatory prerequisites, a hospital may be classified as a rural referral center if—

• The hospital's case-mix index is at least equal to the lower of the median case-mix index for urban hospitals in its census region, excluding hospitals with approved teaching programs, or the median case-mix index for all urban hospitals nationally; and
• The hospital's number of discharges is at least 5,000 per year, or, if fewer, the median number of discharges for urban hospitals in the census region in which the hospital is located. (The number of discharges criterion for an osteopathic hospital is at least 3,000 discharges per year, as specified in section 1886(d)(5)(C)(i) of the Act.)

1. Case-Mix Index

Section 412.96(c)(1) provides that CMS will establish updated national and regional case-mix index values in each year's annual notice of prospective payment rates for purposes of determining rural referral center status. The methodology we use to determine the proposed national and regional case-mix index values is set forth in regulations at § 412.96(c)(1)(ii). The proposed national mean case-mix index value for FY 2004 in the May 19, 2003 proposed rule included all urban hospitals nationwide, and the proposed regional values for FY 2004 were the median values of urban hospitals within each census region, excluding those hospitals with approved teaching programs (that is, those hospitals receiving indirect medical education payments as provided in § 412.105). These proposed values were based on discharges occurring during FY 2002 (October 1, 2001 through September 30, 2002) and included bills posted to CMS' records through December 2002.

In the May 19, 2003 proposed rule, we proposed that, in addition to meeting other criteria, if they are to qualify for initial rural referral center status for cost reporting periods beginning on or after October 1, 2003, rural hospitals with fewer than 275 beds must have a case-mix index value for FY 2002 that is at least—

• 1.3374; or
• The median case-mix index value (not transfer-adjusted) for urban hospitals (excluding hospitals with approved teaching programs as identified in § 412.105) calculated by CMS for the census region in which the hospital is located. (See the table set forth in the May 19, 2003 proposed rule at 68 FR 27201.)

Based on the latest data available (FY 2002 bills received through March 2003), in addition to meeting other criteria, hospitals with fewer than 275 beds, if they are to qualify for initial rural referral center status for cost reporting periods beginning on or after October 1, 2003, must have a case-mix index value for FY 2003 that is at least—

• 1.3373; or
• The median case-mix index value (not transfer-adjusted) for urban hospitals (excluding hospitals with approved teaching programs as identified in § 412.105) calculated by CMS for the census region in which the hospital is located. The final median case-mix index values by region are set forth in the following table:

Hospitals seeking to qualify as rural referral centers or those wishing to know how their case-mix index value compares to the criteria should obtain hospital-specific case-mix index values (not transfer-adjusted) from their fiscal intermediaries. Data are available on the Provider Statistical and Reimbursement (PS&R) System. In keeping with our policy on discharges, these case-mix index values are computed based on all Medicare patient discharges subject to DRG-based payment.

2. Discharges

Section 412.96(c)(2)(i) provides that CMS will set forth the national and regional numbers of discharges in each year's annual notice of prospective payment rates for purposes of determining rural referral center status. As specified in section 1886(d)(5)(C)(ii) of the Act, the national standard is set at 5,000 discharges. In the May 19, 2003 proposed rule, we proposed to update the regional standards based on discharges for urban hospitals' cost reporting periods that began during FY 2002 (that is, October 1, 2001 through September 30, 2002).

Therefore, in the May 19, 2003 proposed rule, we proposed that, in addition to meeting other criteria, a hospital, if it is to qualify for initial rural referral center status for cost reporting periods beginning on or after October 1, 2003, must have as the number of discharges for its cost reporting period that began during FY 2002 a figure that is at least—

• 5,000 (3,000 for an osteopathic hospital); or
• The median number of discharges for urban hospitals in the census region in which the hospital is located. (See the table set forth in the May 19, 2003 proposed rule at 68 FR 27201.)

Based on the latest discharge data available for FY 2002, the final median number of discharges for urban hospitals by census region area are as follows:

We reiterate that if an osteopathic hospital is to qualify for rural referral center status for cost reporting periods beginning on or after October 1, 2003, the hospital must have at least 3,000 discharges for its cost reporting period that began during FY 2002.

We did not receive any comments on the criteria for rural referral centers.

C. Indirect Medical Education (IME) Adjustment (§ 412.105) and Disproportionate Share Hospital (DSH) Adjustment (§ 412.105)

1. Available Beds and Patient Days: Background (§ 412.105(b) and § 412.106(a)(1)(ii))

Section 1886(d)(5)(B) of the Act provides that subsection (d) hospitals that have residents in approved graduate medical education (GME) programs receive an additional payment for each discharge of Medicare beneficiaries to reflect the higher indirect patient care costs of teaching hospitals relative to nonteaching hospitals. The existing regulations regarding the calculation of this additional payment, known as the indirect medical education (IME) adjustment, are located at § 412.105. The additional payment is based on the IME adjustment factor, calculated using hospitals' ratios of residents to beds. The determination of the number of beds, based on available bed days, is specified at § 412.105(b). This determination of the number of available beds is also applicable for other purposes, including the level of the disproportionate share hospital (DSH) adjustment payments under § 412.106(a)(1)(i).

Section 1886(d)(5)(F) of the Act specifies two methods for a hospital to qualify for the Medicare DSH adjustment. The primary method, which is a subject of this final rule, is for a hospital to qualify based on a complex statutory formula under which payment adjustments are based on the level of the DSH patient percentage. The first computation includes the number of patient days that are furnished to patients who were entitled to both Medicare Part A and Supplemental Security Income (SSI) benefits. This number is divided by the total number of patient days that are associated with patients entitled to benefits under Medicare Part A. The second computation includes hospital patient days that are furnished to patients who, for those days, were eligible for Medicaid but were not entitled to benefits under Medicare Part A. This number is divided by the number of total hospital inpatient days in the same period.

Hospitals whose DSH patient percentage exceeds 15 percent are eligible for a DSH payment adjustment (prior to April 1, 2001, the qualifying DSH patient percentage varied, in part, by the number of beds (66 FR 39882)). The DSH payment adjustment may vary based on the DSH patient percentage and the type of hospital: the statute provides for different adjustments for urban hospitals with 100 or more beds and rural hospitals with 500 or more beds, hospitals that qualify as rural referral centers or SCHs, and other hospitals.

As described in the May 19, 2003 proposed rule, we are combining in this final rule our discussion of changes to the policies for counting beds and patient days, in relation to the calculations at §§ 412.105(b) and 412.106(a)(1) because the underlying concepts are similar, and we believe they should generally be interpreted in a consistent manner for both purposes. Specifically, we proposed to clarify that beds and patient days that are counted for these purposes should be limited to beds or patient days in hospital units or wards that would be directly included in determining the allowable costs of inpatient hospital care payable under the IPPS on the Medicare cost reports. As a preliminary matter, beds, and patient days associated with these beds, that are located in units or wards that are excluded from the IPPS (for example, psychiatric or rehabilitation units), and thus from the determination of allowable costs of inpatient hospital care under the IPPS on the Medicare cost report, are not to be counted for purposes of §§ 412.105(b) and 412.106(a)(1).

The remainder of this discussion pertains to beds and patient days in units or wards that are not excluded from the IPPS and for which costs are included in determining the allowable costs of inpatient hospital care under the IPPS on the Medicare cost report. For example, neonatal intensive care unit beds are included in the determination of available beds because the costs and patient days associated with these beds are directly included in the determination of the allowable costs of inpatient hospital care under the IPPS. In contrast, beds, and patient days associated with the beds, that are located in excluded distinct-part psychiatric or rehabilitation units would not be counted for purposes of §§ 412.105(b) and 412.106(a)(1) under any circumstances, because the costs associated with those units or wards are excluded from the determination of the costs of allowable inpatient care under IPPS.

This policy has been upheld in the past by various courts. (See, for example, Little Co. of Mary Hospital and Health Care Centers v. Shalala, 165 F.3d 1162 (7th Cir. 1999; Grant Medical Center v. Shalala, 905 F. Supp. 460 (S.D. Ohio 1995); Sioux Valley Hospital v. Shalala, No. 93-3741SD, 1994 U.S. App. LEXIS 17759 (8th Cir. July 20, 1996) (unpublished table decision); Amisub v. Shalala, No. 94-1883 (TFH) (D.D.C. December 4, 1995) (mem.).) In these cases, the courts agreed with the Secretary's position distinguishing between the treatment of neonatal intensive care unit beds and well-baby nursery beds based on the longstanding policy of CMS that neonatal intensive care unit days are considered intensive care days (part of inpatient routine care) rather than nursery days.

Our policies on counting beds are applied consistently for both IME and DSH although the incentives for hospitals can be different for IME and DSH. For purposes of IME, teaching hospitals have an incentive to minimize their number of available beds in order to increase the resident-to-bed ratio and maximize the IME adjustment. On the other hand, for DSH purposes, urban hospitals with under 100 beds and rural hospitals with under 500 beds may have an incentive to increase their bed count in order to qualify for the higher DSH payments for urban hospitals with over 100 beds or rural hospitals with over 500 beds.

However, some courts have applied our current rules in a manner that is inconsistent with our current policy and that would result in inconsistent treatment of beds, patient days, and costs. For example, in Clark Regional Medical Center v. United States Department of Health & Human Services, 314 F.3d 241 (6th Cir. 2002), the court upheld the district court's ruling that all bed types not specifically excluded from the definition of available bed days in the regulations must be included in the count of available bed days. Similarly, in a recent decision in the Ninth Circuit Court of Appeals (Alhambra v. Thompson, 259 F.3d 1071 (Ninth Cir. 2001), the court ruled that days attributable to groups of beds that are not separately certified as distinct part beds (that is, nonacute care beds in which care provided is at a level below the level of routine inpatient acute care) but are adjacent to or in an acute care “area” are included in the “areas of the hospital that are subject to the prospective payment system” and should be counted in calculating the Medicare DSH patient percentage.

These courts considered subregulatory guidance (program instructions) in formulating their decisions. Although this final rule clarifies the underlying principles for our bed and patient days counting policies and amends the relevant regulations to be consistent with these clarifications, we recognize the need to revise some of our program instructions to make them fully consistent with these clarifications and will act to do so as soon as possible.

While some of the topics discussed below pertain only to counting available beds (unoccupied beds) and some only to counting patient days (section 1115 waiver days, dual-eligible days, and Medicare+Choice days), several important topics are applicable to both bed-counting and day-counting policies (nonacute care beds and days, observation beds and days, and swing-beds and days). Therefore, for ease of discussion, we have combined all topics pertaining to counting available beds and patient days together in the following discussion.

Comment: One commenter expressed concern about our policy to use the same definition of beds for IME and DSH. The commenter argued that Congress used different terminology to define the types of beds that should be used for these two payment adjustments. Section 1886(d)(5)(B)(vi)(I) of the Act indicates the IME adjustment is to be based on “the hospital's available beds (as defined by the Secretary).” For purposes of the DSH adjustment, section 1886(d)(5)(F)(v) of the Act simply refers to the number of “beds” in the hospital. The commenter believed that, because the Act does not narrow the bed count for DSH purposes to those that are available, it is unlawful and inappropriate for CMS to use the available bed definition for DSH purposes.

Response: We believe both statutory references cited by the commenter provide the Secretary with administrative discretion to define beds, one explicitly and one implicitly. In light of this discretion, we strongly believe it is important to apply a consistent definition for purposes of both IME and DSH adjustments, particularly because many hospitals receive both types of adjustments. We note that we have used available beds for purposes of determining whether hospitals qualify for DSH payments Congress directed us to make this adjustment in 1988. Since that time, Congress has amended the DSH provisions in the Act on numerous occasions, and certainly could have made clear its intention that we not use available beds for DSH purposes if that was its intent. Therefore, we disagree with this comment.

2. Unoccupied Beds

We are still reviewing the large number of comments on our proposal on unoccupied beds in the May 19, 2003 proposed rule. Due to the number and nature of the comments we received on our proposed policy, we are addressing the public comments in a separate document. We refer individuals who are interested in reviewing the background information and discussion of the proposed policy to the May 19, 2003 proposed rule (68 FR 37202 through 37204).

3. Nonacute Care Beds and Days

As noted above, our policies for counting beds are generally consistent with the method of reporting patient days for the purpose of calculating the costs of hospital inpatient care in individual cost centers on the Medicare cost report. Furthermore, since the IME and DSH adjustments are part of the IPPS, we read the statutory references to beds and days to apply only to inpatient beds and days.

Under the existing provisions of § 412.105(b), the regulations specifically exclude beds or bassinets in the healthy newborn nursery, custodial care beds, or beds in excluded distinct part hospital units as types of beds excluded from the count of available beds.

Existing regulations at § 412.106(a)(1)(ii) state that the number of patient days used in the DSH percentage calculation includes only those days attributable to areas of the hospital that are subject to the IPPS and excludes all others. This regulation was added after being proposed in the March 22, 1988 Federal Register (53 FR 9339), and made final in the September 30, 1988 Federal Register (53 FR 38479). At that time, we indicated that, “based on a reading of the language in section 1886(d)(5)(F) of the Act, which implements the disproportionate share provision, we are in fact required to consider only those inpatient days to which the prospective payment system applies in determining a prospective payment hospital's eligibility for a disproportionate share adjustment.” Using this reasoning, we stated that the DSH patient percentage calculation should only include patient days associated with the types of services paid under the IPPS.

As noted previously, a recent decision in the Ninth Circuit Court of Appeals (Alhambra v. Thompson) ruled that days attributable to groups of beds that are not separately certified as distinct part beds (that is, nonacute care beds in which care provided is generally at a level below the level of routine inpatient acute care), but are adjacent to or in an acute care “area,” are included in the “areas of the hospital that are subject to the prospective payment system” and should be counted in calculating the Medicare DSH patient percentage.

In light of the Ninth Circuit decision that our rules were not sufficiently clear to permit exclusion of bed days based on the area where the care is provided, in the May 19, 2003 proposed rule, we proposed to revise our regulations to be more specific. Therefore, we proposed to clarify that beds and patient days are excluded from the calculations at § 412.105(b) and § 412.106(a)(1)(ii) if the nature of the care provided in the unit or ward is inconsistent with what is typically furnished to acute care patients, regardless of whether these units or wards are separately certified or are located in the same general area of the hospital as a unit or ward used to provide an acute level of care. Although the intensity of care may vary within a particular unit, such that some patients may be acute patients while others are nonacute, believe that a patient-by-patient, day-by-day review of whether the care received would be paid under the IPPS would be unduly burdensome. Therefore, we believe it is more practical to apply this principle (that is, that we should consider only the inpatient days to which the IPPS applies) by using a proxy measure that is based upon the location at which the services were furnished.

In particular, we proposed to revise our regulations to clarify that the beds and patient days attributable to a nonacute care unit or ward should not be included in the calculations at § 412.105(b) and § 412.106(a)(1)(ii), even if the unit is not separately certified by Medicare as a distinct-part unit and even if the unit or ward is within the same general location of the hospital as areas that are subject to the IPPS (that is, a unit that provides an IPPS level of care is on the same floor of the hospital as a subacute care unit that does not provide an IPPS level of care).

Exceptions to this policy to use the level of care generally provided in a unit or ward as proxy for the level of care provided to a particular patient on a particular day are outpatient observation bed days and swing-bed days, which are excluded from the count of available bed days even if the care is provided in an acute care unit. Our policies pertaining to these beds and days are discussed further below. Another exception is healthy newborn nursery days. The costs, days, and beds associated with a healthy newborn nursery are excluded from inpatient calculations for Medicare purposes. Meanwhile, for the purpose of computing the Medicaid patient share computation of the DSH patient percentages, these days are included both as Medicaid patient days and as total patient days. Newborn nursery costs, days, and beds are treated this way because the costs are not directly included in calculating Medicare hospital inpatient care costs because Medicare does not generally cover services for infants. However, Medicaid does offer extensive coverage to infants, and nursery costs would be directly included in calculating Medicaid hospital inpatient care costs. Therefore, these costs, days, and beds are excluded for Medicare purposes, but included for determining the Medicaid DSH percentage. (This policy was previously communicated through a memorandum to CMS Regional Offices on February 27, 1997.)

Generally, as discussed previously, if the nature of the care provided in the unit or ward is consistent with what is typically furnished to acute care patients, and, therefore, would be characteristic of services paid under the IPPS, the patient days, beds, and costs of that unit or ward would be classified as inpatient acute care (except for observation bed days and swing bed days, as discussed later in this preamble). Conversely, if the intensity and type of care provided in the unit or ward are not typical of a service that would be paid under the IPPS (for example, nonacute care), we proposed that the beds and patient days attributable to a nonacute care unit or ward should not be included in the calculations of beds and patient days at § 412.105(b) and § 412.106(a)(1)(ii).

The proposed policy is not intended to focus on the level or type of care provided to individual patients in a unit, but rather on the level and type of care provided in the unit as a whole. For example, the bed days for a patient participating in an experimental procedure that is not covered under the IPPS should be counted as long as the patient is treated in a unit of the hospital that generally provides acute inpatient care normally payable under the IPPS. The expectation is that a patient located in an acute care unit or ward of the hospital is receiving a level of care that is consistent with what would be payable under the IPPS.

There are instances where services that are provided in units excluded from the IPPS (such as rehabilitation and psychiatric distinct-part units) are also consistent with the level of care that would qualify for payment under the IPPS. However, §§ 412.105(b) and 412.106(a)(1)(ii) specifically exclude the beds and patient days associated with these excluded units. That exclusion is because the costs of care provided in these units are paid outside the IPPS, even though some of the care provided may be of a type that would be payable under the IPPS if the care was provided in an IPPS unit.

We proposed to revise § 412.105(b) to clarify that beds in units or wards established or used to provide a level of care that is not consistent with care that would be payable under the IPPS cannot be counted. We also proposed to revise the DSH regulations at § 412.106(a)(1)(ii) to clarify that the number of patient days includes only those attributable to patients that receive care in units or wards that generally furnish a level of care that would generally be payable under the IPPS.

We note the proposed revisions were clarifications of our regulations to reflect our longstanding interpretation of the statutory intent, especially relating to the calculation of the Medicare DSH patient percentage.

Comment: Several commenters objected to our proposal and indicated that we were attempting to codify the Secretary's litigation position in Alhambra and administratively overrule the Ninth Circuit's decision in that case. Commenters asserted that the flaw in the proposal is that it is inconsistent with the Act to base the Medicaid days calculation of the DSH patient percentage on whether or not Medicare pays for the services that are generally provided within a unit. Specifically, commenters believed the proposal would restrict the definition of patient days in a way that is not authorized by the Act.

Response: We disagree that our proposed clarification is inconsistent with the statute. First, the clarification is merely a codification of the Secretary's longstanding policy. In addition, we believe that interpreting the statute as we have historically done is reasonable and permissible. Section 1886(d)(5)(F)(vi)(II) of the Act governs the portion of the disproportionate share percentage made up of the percentage of patient days used by patients eligible for medical assistance under a title XIX State plan. Specifically, section 1886(d)(5)(F)(vi)(II) of the Act states that the numerator of such fraction equals the “number of the hospital's patient days for such period which consist of patients who (for such days) were eligible for medical assistance under a State plan approved under title XIX, but who were not entitled to benefits under part A of this title.” The statute does not define the term “hospital's patient days.” Thus, the statute is ambiguous, and the Secretary has the authority to reasonably interpret that term.

We note that although the calculation performed under section 1886(d)(5)(F)(vi)(II) of the Act includes a count of patient days used by Medicaid-eligible individuals, the calculation actually is used to determine how much additional payment the hospital should receive under Medicare for the higher Medicare costs associated with treating a disproportionate share of low-income individuals. This point is demonstrated in the rationale for establishing the DSH adjustment as described in the Committee Report accompanying Pub. L. 99-272: “Hospitals that serve a disproportionate share of low-income patients have higher Medicare costs per case” (H. Rept. No. 99-242(I), 99th Cong., 2d Sess., (1985), p. 16).

Furthermore, we view section 1886(d)(5)(F)(vi)(II) of the Act as purely a Medicare, inpatient hospital provision, given that there already exists a distinct formula for computing DSH payments under title XIX—the Medicaid title. Because the DSH formula in title XVIII of the Act is intended to provide an add-on payment to inpatient hospitals for additional amounts they incur in treating low-income, Medicare patients, we believe it is reasonable to count only those days spent in wards or units that would generally provide an acute level of care.

We believe it is reasonable to interpret the phrase, “hospital's patient days,” to mean only the hospital's inpatient days at a level of care that would be covered under the IPPS as a means to determine an IPPS payment adjustment. Further, we believe that it is administratively inefficient and impracticable to calculate a hospital's inpatient days based on a determination, on a day-by-day basis, of whether a particular patient in a particular inpatient bed is receiving a level of care that would be covered under the IPPS. Therefore, we proposed to use, as a proxy, the level of care that is generally provided in particular units or wards, and to exclude patient days attributable to units or wards in which care delivered is not generally of a type that would be covered under the IPPS.

We also do not believe that by placing our longstanding interpretation of our rules in regulations we are unlawfully overruling or nullifying the decision by the Ninth Circuit in Alhambra Hospital v. Thompson, 259 F.3d 1071 (9th Cir. 2001). The Ninth Circuit decision focused on an interpretation of CMS’ previous regulation at § 412.106(a)(1)(ii)—not on an interpretation of the statute. (For example, when the court stated the “Standard of Review” it would use to decide the case, it referred only to “[o]ur review of an agency's interpretation of its own regulations.” Alhambra at 1074). Although we respectfully disagree with the Ninth Circuits interpretation of the existing regulations, we are nonetheless amending them, through notice and comment rulemaking to ensure that going forward the regulations clearly reflect our longstanding position. Therefore, we do not agree with the commenter's assertion that our proposed policy is an illegal attempt to administratively overrule the Ninth Circuit's decision in Alhambra. Therefore, going forward, we plan to apply the clarified regulation to hospitals in all U.S. jurisdictions, including hospitals in the Ninth Circuit.

4. Observation Beds and Swing-Beds

Observation services are those services furnished by a hospital on the hospital's premises that include use of a bed and periodic monitoring by a hospital's nursing or other staff in order to evaluate an outpatient's condition or to determine the need for a possible admission to the hospital as an inpatient. When a hospital places a patient under observation but has not formally admitted him or her as an inpatient, the patient initially is treated as an outpatient. Consequently, the observation bed days are not recognized under the IPPS as part of the inpatient operating costs of the hospital.

Observation services may be provided in a distinct observation bed area, but they may also be provided in a routine inpatient care unit or ward. In either case, our policy is the bed days attributable to beds used for observation services are excluded from the counts of available bed days and patient days at §§ 412.105(b) and 412.106(a)(1)(ii). This policy was clarified in a memorandum that was sent to all CMS Regional Offices (for distribution to fiscal intermediaries) dated February 27, 1997, which stated that if a hospital provides observation services in beds that are generally used to provide hospital inpatient services, the days that those beds are used for observation services should be excluded from the available bed day count (even if the patient is ultimately admitted as an acute inpatient).

A swing-bed is a bed that is otherwise available for use to provide acute inpatient care and is also occasionally used to provide SNF-level care. The criteria for a hospital to meet the requirements to be granted an approval from CMS to provide posthospital extended care services are located under § 482.66, and for a swing-bed CAH under § 485.645. Under § 413.114(a)(1), payment for posthospital SNF care furnished in swing-beds is in accordance with the provisions of the prospective payment system for SNF care (effective for services furnished in cost reporting periods beginning on and after July 1, 2002). Similar to observation beds and patient days, swing-beds and patient days are excluded from the counts of available bed days and patient days at §§ 412.105(b) and 412.106(a)(1)(ii) when the swing-bed is used to furnish SNF care.

Observation beds and swing-beds are both special, frequently temporary, alternative uses of acute inpatient care beds. That is, only the days an acute inpatient care unit or ward bed is used to provide outpatient observation services are to be deducted from the available bed count under § 412.105(b). Otherwise, the bed is considered available for acute care services (as long as it otherwise meets the criteria to be considered available). This same policy applies for swing-beds. The policies to exclude observation bed days and swing-bed days as described above stem from the fact that these days are not payable under the IPPS.

Some hospitals have contested our policy excluding swing-beds and patient days and observation beds and patient days under existing §§ 412.105(b) and 412.106(a)(1)(ii). For example, in Clark Regional Medical Center v. United States Department of Health & Human Services, 314 F.3d 241 (6th Cir. 2002), the court upheld the district court's ruling that all bed types not specifically excluded from the definition of available bed days in the regulations must be included in the count of available bed days. The hospitals involved in this decision wanted to include observation and swing-bed days in their bed count calculation in order to qualify for higher DSH payments as available to hospitals with more than 100 beds. The Court found that “the listing of beds to be excluded from the count restricts the class of excluded beds only to those specifically listed.” Because observation beds and swing-beds are not currently specifically mentioned in § 412.105(b) as being excluded from the bed count, the Court ruled that these beds must be included in the count.

The list of the types of beds excluded from the count under existing § 412.105(b) was never intended to be an exhaustive list of all of the types of beds to be excluded from the bed count under this provision. In fact, over the years, specific bed types have been added to the list as clarifications of the types of beds to be excluded, not as new exclusions (see the September 1, 1994 Federal Register (59 FR 45373) and September 1, 1995 Federal Register (60 FR 45810), where we clarified exclusions under our policy that were not previously separately identified in the regulation text).

Although the Court in Clark found that Congress had not explicitly “addressed the question of whether swing and observation beds should be included in the count of beds in determining whether a hospital qualifies for the DSH adjustment,” Clark, 314 F.3d at 245, the Court found that observation and swing-bed days were included under the “plain meaning” of the regulation text at § 412.106(a)(1)(ii), which reads: “The number of patient days includes only those days attributable to areas of the hospital that are subject to the prospective payment system and excludes all others.” However, the preamble language of the rule that promulgated the regulatory provision at § 412.106(a)(1)(ii) clarified its meaning (53 FR 38480, September 30, 1988):

“Although previously the Medicare regulations did not specifically define the inpatient days for use in the computation of a hospital's disproportionate share patient percentage, we believe that, based on a reading of the language in section 1886(d)(5)(F) of the Act, which implements the disproportionate share provision, we are in fact required to consider only those inpatient days to which the prospective payment system applies in determining a prospective payment hospital's eligibility for a disproportionate share adjustment.”

Our policy excluding outpatient observation and swing-bed days is consistent with this regulatory interpretation of days to be counted under § 412.106(a)(1)(ii). That is, the services provided in these beds are not payable under the IPPS (unless the patient is admitted, in the case of observation bed days).

As outlined previously, our consistent and longstanding policy, which has been reviewed and upheld previously by several courts, including the United States District Court for the District of Columbia in Amisub v. Shalala, is based on the principle of counting beds in generally the same manner as the patient days and costs are counted. Our policy to exclude observation and swing-bed days under the regulations at § 412.105(b) and § 412.106(a)(1) stems from this policy.

In the May 19, 2003 proposed rule, although we reiterated our longstanding policy that observation beds and swing bed days generally are excluded, we proposed to amend our policy with respect to observation bed days of patients who ultimately are admitted. We are still in the process of reviewing the comments and defer action until a later rule with respect this issue—for example, patients in observation beds who are ultimately admitted to the hospital.

Comment: Some commenters objected to the exclusion of observation bed days from the available bed days count on the grounds that it is a flawed premise that the size of a hospital's bed complement should be impacted by the payment policy classification of the services provided to the patient. That is, a bed should not be excluded from the available bed day count because it is used to provide services not payable under the IPPS on a particular day.

Response: When the application of IPPS payment policy is dependent on a determination of a hospital's number of beds, it seems reasonable to base that determination on the portion of the hospital that generates the costs that relate to those IPPS payments. As stated above, our bed counting policies start with the premise that the treatment of beds should be consistent with the treatment of the patient days and the costs of those days on the Medicare cost report. Therefore, we continue to believe it is appropriate to exclude outpatient observation bed days, even when the beds used to provide that service is located in a routine inpatient care unit or ward.

5. Labor, Delivery, and Postpartum Beds and Days

Prior to December 1991, Medicare's policy on counting days for maternity patients was to count an inpatient day for an admitted maternity patient in the labor/delivery room at the census taking hour. This is consistent with Medicare policy for counting days for admitted patients in any other ancillary department at the census-taking hour. However, based on decisions adverse to the government regarding this policy in a number of Federal courts of appeal, including the United States Court of Appeals for the District of Columbia Circuit, the policy regarding the counting of inpatient days for maternity patients was revised to reflect our current policy.

Our current policy regarding the treatment of labor and delivery bed days is described in Section 2205.2 of the PRM, which states that a maternity inpatient in the labor/delivery room at midnight is not included in the census of inpatient routine care if the patient has not occupied an inpatient routine bed at some time since admission. For example, if a Medicaid patient is in the labor room at the census and has not yet occupied a routine inpatient bed, the bed day is not counted as a routine bed day of care in Medicaid or total days and, therefore, is not included in the counts under existing §§ 412.105(b) and 412.106(a)(1)(ii). If the patient is in the labor room at the census but had first occupied a routine bed, a routine inpatient bed day is counted, in Medicaid and total days, for DSH purposes and for apportioning the cost of routine care on the cost report (consistent with our longstanding policy to treat days, costs, and beds similarly).

Increasingly, hospitals are redesigning their maternity areas from separate labor and delivery rooms, and postpartum rooms, to single multipurpose labor, delivery, and postpartum (LDP) rooms. In order to appropriately track the days and costs associated with LDP rooms, it is necessary to apportion them between the labor and delivery cost center, which is an ancillary cost center and the routine adults and pediatrics cost center. This is done under our policy by determining the proportion of the patient's stay in the LDP room that the patient was receiving ancillary services (labor and delivery) as opposed to routine adult and pediatric services (postpartum).

An example of this would be if 25 percent of the patient's time in the LDP room was for labor/delivery services and 75 percent for routine care, over the course of a 4-day stay in the LDP room. In that case, 75 percent of the time the patient spent in the LDP room is applied to the routine inpatient bed days and costs (resulting in 3 routine adults and pediatrics bed days for this patient, 75 percent of 4 total days). For purposes of determining the hospital bed count, the time that the beds are unoccupied should be counted as available bed days using an average percentage (for example, 75 percent adults and pediatrics and 25 percent ancillary) based on all patients. In other words, in this example, 75 percent of the days the bed is unoccupied would be counted in the available bed count.

We realize that it may be burdensome for a hospital to determine for each patient in this type of room the amount of time spent in labor/delivery and the amount of time spent receiving routine care. Alternatively, the hospital could calculate an average percentage of time patients receive ancillary services, as opposed to routine inpatient care in the LDP room(s) during a typical month, and apply that percentage through the rest of the year.

Comment: Some commenters stated that the LDP days that patients spend in routine inpatient wards of hospitals prior to the day those patients give birth are in areas of the hospital where routine inpatient beds are located, and they are not excluded from the IPPS. Therefore, the commenters asserted that these days should be counted in the patient days and available bed days counts. Commenters also pointed out the LDP days are in licensed beds, and argued that these days should be counted in their entirety.

Other commenters supported our proposal to allow calculation of an average percentage of time LDP patients spend in labor/delivery compared to postpartum to be used to apportion LDP days. Commenters commended CMS for recognizing the cumbersome recordkeeping and reporting that would otherwise be required.

One commenter suggested that it is not necessary for our policy applicable to counting patient days for purposes of the DSH computation to comply with other Medicare cost reporting policies, such as the need to separately allocate the ancillary costs associated with LDP rooms. The commenter cited prior PRRB appeals in which CMS took this position.

Response: As we previously stated above and in the proposed rule, initially, Medicare's policy did count an inpatient day for an admitted maternity patient even if the patient was in the labor/delivery room at the census-taking hour. However, based on adverse court decisions, the policy was revised to state that the patient must first occupy an inpatient routine bed before being counted as an inpatient. With the development of LDP rooms, we found it necessary to apply this policy consistently in those settings, in order to appropriately apportion the costs between labor and delivery ancillary services and routine inpatient care.

Although we have not previously formally specified in guidance or regulations the methodology for applying this policy to LDP rooms, this is not a new policy. However, as suggested by the commenters, we believe this policy may not have been applied consistently. Therefore, we believe it is important to clarify the policy as part of our discussion of our policies pertaining to counting patient bed days.

We continue to believe the LDP apportionment described above is an appropriate policy and does not, in fact, impose a significant additional burden because hospitals are already required to allocate cost on the cost report between ancillary and routine costs. In addition, this allocation is already required to be consistent with our treatment of costs, days, and beds and is consistent with our other patient bed day policies. Therefore, this policy will be applied to all currently open and future cost reports. However, it is not necessary to reopen previously settled cost reports to apply this policy.

6. Days Associated With Demonstration Projects Under Section 1115 of the Act

Some States extend medical benefits to a given population that could not have been made eligible for Medicaid under a State plan amendment under section 1902(r)(2) or section 1931(b) of the Act under a section 1115(a)(2) demonstration project (also referred to as a section 1115 waiver). These populations are specific, finite populations identifiable in the award letters and special terms and conditions apply to the demonstrations.

On January 20, 2000, we issued an interim final rule with comment period (65 FR 3136), followed by a final rule issued on August 1, 2000 (65 FR 47086 through 47087), to allow hospitals to include the patient days of all populations that receive benefits under a section 1115 demonstration project in calculating the Medicare DSH adjustment. Previously, hospitals were to include only those days for populations under the section 1115 demonstration project who were, or could have been made, eligible under a State plan. Patient days of those expansion waiver groups who could not be made eligible for medical assistance under the State plan were not to be included for determining Medicaid patient days in calculating the Medicare DSH patient percentage. Under the January 20, 2000 interim final rule with comment period (65 FR 3137), hospitals could include in the numerator of the Medicaid fraction those patient days for individuals who receive benefits under a section 1115 expansion waiver demonstration project (effective with discharges occurring on or after January 20, 2000).

In the January 20, 2000 interim final rule with comment period, we explained that including the section 1115 expansion populations “in the Medicare DSH calculation is fully consistent with the Congressional goals of the Medicare DSH adjustment to recognize the higher costs to hospitals of treating low-income individuals covered under Medicaid.”

Since that revision, we have become aware that there are certain section 1115 demonstration projects that serve expansion populations with benefit packages so limited that the benefits are not similar to the medical assistance available under a Medicaid State plan. These section 1115 demonstration projects extend coverage only for specific services and do not include inpatient care in the hospital. Because of the limited nature of the coverage offered, the population involved may have a significantly higher income than traditional Medicaid beneficiaries.

In allowing hospitals to include patient days related to section 1115 expansion waiver populations, our intention was to include patient days of section 1115 expansion waiver populations who receive benefits under the demonstration project that are similar to those available to traditional Medicaid beneficiaries, including inpatient benefits. Because of the differences between expansion populations in these limited benefit demonstrations and traditional Medicaid beneficiaries, in the May 19, 2003 proposed rule, we proposed that the Medicare DSH calculation should exclude from treatment as Medicaid patient days those patient days attributable to limited benefit section 1115 expansion waiver populations (proposed § 412.106(b)(4)(i)).

For example, a State may extend a family planning benefit to an individual for 2 years after she has received the 60-day postpartum benefit under Medicaid, or a State may choose to provide a family planning benefit to all individuals below a certain income level, regardless of having previously received the Medicaid postpartum benefit. This is a limited, temporary benefit that is generally administered in a clinic setting (see section 1905(a)(4)(C) of the Act). Also, a number of States are developing demonstrations that are limited to providing beneficiaries an outpatient prescription drug benefit. Generally, these limited benefits under a demonstration project do not include inpatient benefits. If a hospital were to include the days attributable to patients receiving benefits under such a limited benefit, the hospital would be able to receive higher DSH payments, perhaps substantially, for patients who may otherwise be insured for inpatient care. For example, these limited demonstrations provide benefits that may be needed to supplement private insurance coverage for individuals who do not have incomes low enough to qualify for Medicaid under the State plan. We do not believe such patients should be counted in the DSH patient percentage as eligible for title XIX.

As we have noted previously, at the time the Congress enacted the Medicare DSH adjustment provision (which was added to the law by section 9105 of COBRA and was effective for discharges occurring on or after May 1, 1986), there were no approved section 1115 demonstration projects involving expansion populations and the statute does not address the treatment of these days. Although we did not initially include patient days for individuals who receive extended benefits only under a section 1115 demonstration project, we nevertheless expanded our policy in the January 20, 2000 revision to these rules to include such patient days. We now believe that this reading is warranted only to the extent that those individuals receive inpatient benefits under the section 1115 demonstration project.

Therefore, we proposed to revise § 412.106(b)(4)(i) to clarify that patients must be eligible for medical assistance inpatient hospital benefits under an approved State Medicaid plan (or similar benefits, including inpatient hospital benefits, under a section 1115 demonstration project) in order for their hospital inpatient days to be counted as Medicaid days in the calculation of a hospital's DSH patient percentage. Under the proposed clarification, hospital inpatient days attributed to patients who do not receive coverage for inpatient hospital benefits either under the approved State plan or through a section 1115 demonstration would not be counted in the calculation of Medicaid days for purposes of determining a hospital's DSH patient percentage.

Under this reading, in the examples given above, the days associated with a hospital inpatient who receives coverage of prescription drugs or family planning services on an outpatient basis, but no inpatient hospital coverage, through either a Medicaid State plan or a section 1115 demonstration, would not be counted as Medicaid days for purposes of determining the DSH patient percentage.

The proposed revision addressed an unintended potential consequence of our interpretation that hospitals may include in the DSH calculation patient days associated with section 1115 demonstration populations (65 FR 3136). As discussed above, that interpretation was based on our finding that individuals receiving a comprehensive benefit package under a section 1115 demonstration project could appropriately be included in the numerator of the Medicaid fraction (even though the statute does not require such an inclusion), but did not address individuals who were receiving limited benefit packages under a section 1115 demonstration project.

Comment: Some commenters questioned our authority to require a patient obtain to covered inpatient benefits under either a Medicaid State plan or a section 1115 demonstration, in order to be included in the numerator of the Medicaid ratio for the DSH computation. One commenter pointed out that there are many circumstances under which an individual may have income low enough to qualify for Medicaid but still not qualify due to other qualifying criteria, and requested that all patient days of such individuals be counted as Medicaid-eligible.

Response: As stated above and in the proposed rule, we do not believe patients covered under limited-benefit section 1115 demonstration projects that are so limited that they are not similar to the medical assistance available under a Medicaid State plan should not be included in the count of Medicaid-eligible patients.

Under a traditional State Medicaid program, States are required to offer inpatient benefits to all eligible beneficiaries (see section 1902(a)(10)(A) of the Act). However, under the 1115 demonstration authority, the Secretary has permitted coverage for a limited set of services, such as pharmaceuticals or family planning services, and thus inpatient hospital services may be excluded for expansion populations under some of the section 1115 demonstration programs.

Our intention in allowing hospitals to include patient days related to section 1115 expansion waiver populations was to include patient days of demonstration populations who receive benefits under the demonstration project that are similar to traditional Medicaid beneficiaries, including inpatient benefits.

Comment: One commenter requested that the effective date of the proposed change be delayed until January 1, 2004, to allow fiscal intermediaries to contact States and identify specific coverage for their various section 1115 waiver populations.

Response: Because the DSH adjustment is reconciled when hospitals' cost reports are settled, we do not believe it is necessary to delay the implementation of this policy until January 1, 2004. Furthermore, although we believe it would have been reasonable for hospitals or fiscal intermediaries to have applied this interpretation of our policy regarding the inclusion of section 1115 waiver days prior to this clarification, we recognize that there may be situations in which this policy was not already applied. Therefore, we are making this change and the regulation at § 412.106(b)(4)(i) will be effective for discharges occurring on or after October 1, 2003.

7. Dual-Eligible Patient Days

We are still reviewing the large number of comments received on the proposed provision relating to dual-eligible patient days in the May 19, 2003. Due to the number and nature of the comments we received on our proposed policies, we are addressing the public comments in a separate document. We refer individuals who are interested in reviewing the background information and discussions regarding this policy to the May 19, 2003 proposed rule (68 FR 27207-27208).

8. Medicare+Choice (M+C) Days

We are still reviewing the large number of comments we received on the proposed provision relating to the counting of Medicare+Choice days for purposes of the IME and DSH adjustments. Due to the number and nature of the comments we received on our proposed policies, we are addressing the public comments in a separate document. We refer individuals interested in reviewing the background information and the discussion regarding these policies to the May 19, 2003 proposed rule (68 FR 27208).

D. Medicare Geographic Classification Review Board (MGCRB) Reclassification Process (§ 412.230)

With the creation of the MGCRB, beginning in FY 1991, under section 1886(d)(10) of the Act, hospitals could request reclassification from one geographic location to another for the purpose of using the other area's standardized amount for inpatient operating costs or the wage index value, or both (September 6, 1990 interim final rule with comment period (55 FR 36754), June 4, 1991 final rule with comment period (56 FR 25458), and June 4, 1992 proposed rule (57 FR 23631)). Implementing regulations in Subpart L of Part 412 (§§ 412.230 et seq.) set forth criteria and conditions for redesignations for purposes of the wage index or the average standardized amount, or both, from rural to urban, rural to rural, or from an urban area to another urban area, with special rules for SCHs and rural referral centers.

Effective with reclassifications for FY 2003, section 1886(d)(10)(D)(vi)(II) of the Act provides that the MGCRB must use the average of the 3 years of hourly wage data from the most recently published data for the hospital when evaluating a hospital's request for reclassification. The regulations at § 412.230(e)(2)(ii) stipulate that the wage data are taken from the CMS hospital wage survey used to construct the wage index in effect for prospective payment purposes. To evaluate applications for wage index reclassifications for FY 2004, the MGCRB used the 3-year average hourly wages published in Table 2 of the August 1, 2002 IPPS final rule (67 FR 50135). These average hourly wages are taken from data used to calculate the wage indexes for FY 2001, FY 2002, and FY 2003, based on cost reporting periods beginning during FY 1997, FY 1998, and FY 1999, respectively.

Last year, we received a comment suggesting that we allow for the correction of inaccurate data from prior years as part of a hospital's bid for geographic reclassification (67 FR 50027). The commenter suggested that not to allow corrections to the data results in inequities in the calculation in the average hourly wage for purposes of reclassification. In the August 1, 2002 IPPS final rule, we responded:

“Hospitals have ample opportunity to verify the accuracy of the wage data used to calculate their wage index and to request revisions, but must do so within the prescribed timelines. We consistently instruct hospitals that they are responsible for reviewing their data and availing themselves to the opportunity to correct their wage data within the prescribed timeframes. Once the data are finalized and the wage indexes published in the final rule, they may not be revised, except through the mid-year correction process set forth in the regulations at § 412.63(x)(2). Accordingly, it has been our consistent policy that if a hospital does not request corrections within the prescribed timeframes for the development of the wage index, the hospital may not later seek to revise its data in an attempt to qualify for MGCRB reclassification.

“Allowing hospitals the opportunity to revise their data beyond the timelines required to finalize the data used to calculate the wage index each year would lessen the importance of complying with those deadlines. The likely result would be that the data used to compute the wage index would not be as carefully scrutinized because hospitals would know they may change it later, leading to inaccuracy in the data and less stability in the wage indexes from year to year.”

Since responding to this comment in the FY 2003 IPPS final rule, we have become aware of a situation in which a hospital does not meet the criteria to reclassify because its wage data were erroneous in prior years, and these data are now being used to evaluate its reclassification application. In addition, in this situation, the hospital's wage index was subject to the rural floor because the hospital was located in an urban area with an actual wage index below the statewide rural wage index for the State, and it was for a time period preceding the requirement for using 3 years of data. Therefore, the hospital contends, it had no incentive to ensure its wage data were completely accurate. (However, we would point out that hospitals are required to certify that their cost reports submitted to CMS are complete and accurate. Furthermore, inaccurate or incomplete reporting may have other payment implications beyond the wage index.)

We now more fully understand this particular hospital's situation and we have the administrative authority to establish a policy allowing corrections for this particular set of circumstances, in the proposed rule, we solicited comments on whether it may be appropriate to establish a policy whereby, for the limited purpose of qualifying for reclassification based on data from years preceding the establishment of the 3-year requirement (that is, cost reporting years beginning before FY 2000), a hospital in an urban area that was subject to the rural floor for the period during which the wage data the hospital wishes to revise were used to calculate the wage index, a hospital may request that its wage data be revised.

Comment: One commenter supported the proposed establishment of the exception. However, the commenter recommended that CMS consider allowing all hospitals to make corrections to the data that is used in reclassification determinations.

Response: We continue to believe that requiring wage data corrections by specified deadlines is essential to ensuring that wage data is finalized in an efficient manner. We also continue to believe that final wage data published in the annual IPPS final rules should be as complete and accurate as possible. However, we believe that, in the limited circumstances raised in our proposed rule where the hospital could not have foreseen that its wage data would later be used in a 3-year average, and the hospital was subject to the rural floor, it is feasible to permit a limited exception. Therefore, in this final rule, we are amending § 412.230(e)(2)(ii)(A) to allow, for the limited purpose of qualifying for geographic reclassification, hospitals demonstrating that they meet the limited circumstances described in the amended regulation be considered for reclassification after taking into account revisions subsequent to its use to construct the wage index for IPPS payment purposes. We are not adopting a broader exception, because we continue to believe it is important to ensure that final wage data published in the annual IPPS final rule are complete and accurate. Creating a broad exception to allow for corrections of prior years' data would affect the accuracy and stability in the wage indices from year to year. Therefore, we will continue to require hospitals—other than hospitals meeting the limited exception described in § 412.230(e)(2)(ii)(A)—to ensure that their wage data are correct by applicable deadlines and will not allow for wage data corrections after such deadlines.

Comment: Several hospitals who were interested in reclassifying, as a group, for purposes of the wage index, commented that their efforts to reclassify as an urban group have been unsuccessful primarily because they fail to meet the established requirement set forth in § 412.234(c)(2) that the requesting hospitals must demonstrate that their costs exceed their current payments by 75 percent of the additional payments they would receive through reclassification. The commenters submitted several recommendations for our consideration to clarify or improve our policies and regulations. They recommended that we consider:

• Allowing hospital groups to seek geographic reclassification for purposes of the wage index or standardized amount;
• Allowing hospital groups seeking geographic reclassification to areas where the reclassification would not result in a different standardized amount to seek reclassification for purposes of the wage index without having to satisfy the criteria applicable to hospitals seeking reclassification for purposes of the standardized amount;
• Allowing hospitals in NECMAs to seek reclassification to another MSA under the alternative criteria at § 412.236(c);
• Lowering the cost-to-payment threshold used to evaluate group reclassification applications; or
• In order to evaluate the interrelationship between the area where the hospitals are located and the target area in which they are seeking to reclassify, replacing the cost comparison criteria used to evaluate reclassification eligibility for purposes of the standardized amount with a better indicator of the connection such as, census commuting patterns.

Response: We appreciate the comments and recommendations presented by the hospitals and the importance of this issue. We note that, in developing the proposed rule, we did consider including a proposal to allow urban hospitals to reclassify as a group either for wage index or the standardized amount, or both. However, we did not go forward with the proposal because, upon further review, the criterion that hospitals demonstrate that their costs are in excess of their payments seemed appropriate. We will consider the commenters' recommendations in the future.

Comment: One commenter recommended that CMS consider lowering the applicable qualifying thresholds at § 412.230(c)(1)(iii) and (iv) for urban hospitals seeking reclassification for purposes of the wage index. The commenter specifically suggested that the threshold be lowered from 108 percent of the average hourly wage of hospitals in the area in which the hospital is located, and 84 percent of the average hourly wage of hospitals in the area to which the hospital seeks reclassification, to 106 percent and 82 percent, respectively, for urban hospitals. The commenter further recommended that, if the lower thresholds cannot be reduced for all urban hospitals, CMS consider implementing the lower thresholds for urban hospitals in areas where they are paid as if they are rural.

Response: As pointed out by the commenter, this issue was discussed, in detail, in the August 1, 2000 Federal Register (65 FR 47089 through 47090). While we will consider the recommendations for possible inclusion in a future proposed rule, we did not propose any changes or clarifications to the existing policy. Therefore, we are not adopting this comment.

E. Costs of Approved Nursing and Allied Health Education Activities (§ 413.85)

1. Background

Medicare has historically paid providers for the program's share of the costs that providers incur in connection with approved educational activities. The activities may be divided into the following three general categories to which different payment policies apply:

• Approved graduate medical education (GME) programs in medicine, osteopathy, dentistry, and podiatry. Medicare makes direct and indirect medical education payments to hospitals for residents training in these programs. Existing policy on direct GME payment is found at 42 CFR 413.86, and for indirect GME payment at 42 CFR 412.105.
• Approved nursing and allied health education programs operated by the provider. The costs of these programs are excluded from the definition of inpatient hospital operating costs and are not included in the calculation of payment rates for hospitals paid under the IPPS or in the calculation of payments to hospitals and hospital units excluded from the IPPS that are subject to the rate-of-increase ceiling. These costs are separately identified and “passed through” (that is, paid separately on a reasonable cost basis). Existing regulations on nursing and allied health education program costs are located at 42 CFR 413.85.
• All other costs that can be categorized as educational programs and activities are considered to be part of normal operating costs and are included in the per discharge amount for hospitals subject to the IPPS, or are included as reasonable costs that are subject to the rate-of-increase limits for hospitals and hospital units excluded from the IPPS.

In the May 19, 2003 proposed rule, we proposed to clarify our policy governing payments to hospitals for provider-operated nursing and allied health education programs. Under the regulations at § 413.85 (“Cost of approved nursing and allied health educational activities”), Medicare makes reasonable cost payment to hospitals for provider-operated nursing and allied health education programs. A program is considered to be provider-operated if the hospital meets the criteria specified in § 413.85(f), which means the hospital directly incurs the training costs, controls the curriculum and the administration of the program, employs the teaching staff, and provides and controls both clinical training and classroom instruction (where applicable) of a nursing or allied health education program.

In the January 12, 2001 Federal Register (66 FR 3358), we published a final rule that clarified the policy for payments for approved nursing and allied health education activities in response to section 6205(b)(2) of the Omnibus Budget Reconciliation Act of 1989 (Pub. L. 101-239) and sections 4004(b)(1) and (2) of the Omnibus Budget Reconciliation Act of 1990 (Pub. L. 101-508).

Section 6205(b)(2) of Pub. L. 101-239 directed the Secretary to publish regulations clarifying the rules governing allowable costs of approved educational activities. The Secretary was directed to publish regulations to specify the conditions under which those costs are eligible for pass-through, including the requirement that there be a relationship between the approved nursing or allied health education program and the hospital. Section 4004(b)(1) of Pub. L. 101-508 provides an exception to the requirement that programs be provider-operated to receive pass-through payments. The section provides that, effective for cost reporting periods beginning on or after October 1, 1990, if certain conditions are met, the costs incurred by a hospital (or by an educational institution related to the hospital by common ownership or control) for clinical training (as defined by the Secretary) conducted on the premises of the hospital under an approved nursing or allied health education program that is not operated by the hospital are treated as pass-through costs and paid on the basis of reasonable cost. Section 4004(b)(2) of Pub. L. 101-508 sets forth the conditions that a hospital must meet to receive payment on a reasonable cost basis under section 4004(b)(1).

2. Continuing Education Issue for Nursing and Allied Health Education

Since publication of the January 12, 2001 final rule on nursing and allied health education, we have encountered questions concerning the substantive difference between provider-operated continuing education programs for nursing and allied health education (which would not be reimbursable under Medicare on a reasonable cost basis) and provider-operated approved programs that are eligible to receive Medicare reasonable cost payment. In that final rule, we stated that Medicare would generally provide reasonable cost payment for “programs of long duration designed to develop trained practitioners in a nursing or allied health discipline, such as professional nursing or occupational therapy. This is contrasted with a continuing education program of a month to a year in duration in which a practitioner, such as a registered nurse, receives training in a specialized skill such as enterostomal therapy. While such training is undoubtedly valuable in enabling the nurse to treat patients with special needs and in improving the level of patient care in a provider, the nurse, upon completion of the program, continues to function as a registered nurse, albeit one with special skills. Further distinction can be drawn between this situation and one in which a registered nurse undergoes years of training to become a CRNA. For these reasons, the costs of continuing education training programs are not classified as costs of approved educational activities that are passed-through and paid on a reasonable cost basis. Rather, they are classified as normal operating costs covered by the prospective payment rate or, for providers excluded from the IPPS, as costs subject to the target rate-of-increase limits” (66 FR 3370).

Accordingly, upon publication of the final rule, we revised § 413.85(h)(3) to include continuing education programs in the same category as “educational seminars and workshops that increase the quality of medical care or operating efficiency of the provider.” Costs associated with continuing education programs, as stated above, are recognized as normal operating costs and are paid in accordance with applicable principles.

Prior to the issuance of the May 19, 2003 proposed rule, we received an inquiry requesting further clarification on what is meant by continuing education. It is our belief that provider-operated programs that do not lead to any specific certification in a specialty would be classified as continuing education. In the proposed rule (68 FR 27210), we stated that our use of the term “certification” does not mean certification in a specific skill, such as when an individual is certified to use a specific piece of machinery or perform a specific procedure. Rather, we stated that we believe certification means the ability to perform in the specialty as a whole.

Although, in the past, we believe we have allowed hospitals to be paid for operating a pharmacy “residency” program, in the May 19, 2003 proposed rule, we stated that it has come to our attention that those programs do not meet the criteria for approval as a certified program. Once individuals have finished their undergraduate degree in pharmacy, there are some individuals who go on to participate in 1-year hospital-operated postundergraduate programs. It is our understanding that many individuals complete the 1-year postundergraduate program practice pharmacy inside the hospital setting. However, we also understand that there are pharmacists who do not complete the 1-year postundergraduate program, but have received the undergraduate degree in pharmacy, who also practice pharmacy inside the hospital setting. Because pharmacy students need not complete the 1-year residency program to be eligible to practice pharmacy in the hospital setting, the 1-year programs that presently are operated by hospitals would be considered continuing education, and therefore, would be ineligible for pass-through reasonable cost payment.

We stated that we understood that all individuals who wish to be nurses practicing in a hospital must either complete a 4-year degree program in a university setting, a 2-year associate degree in a community or junior college setting, or a diploma program traditionally offered in a hospital setting. Since participants that complete a provider-operated diploma nursing program could not practice as nurses without that training, the diploma nursing programs are not continuing education programs and, therefore, may be eligible for pass-through treatment.

Because of the apparent confusion concerning the distinction between continuing education programs and approved education programs in the context of reasonable cost pass-through payments for nursing and allied health education activities, in the May 19, 2003 proposed rule, we proposed to revise § 413.85(h)(3) to state that educational seminars, workshops, and continuing education programs in which the employees participate that enhance the quality of medical care or operating efficiency of the provider and, effective October 1, 2003, do not lead to certification required to practice or begin employment in a nursing or allied health specialty, would be treated as educational activities that are part of normal operating costs. We also proposed to add a conforming definition of “certification” for purposes of nursing and allied health education under § 413.85(c) to mean “the ability to practice or begin employment in a specialty as a whole.”

Comment: A large number of commenters responded to our proposal to clarify that, effective October 1, 2003, activities that do not lead to certification required to practice or begin employment in a nursing or allied health specialty would be treated as educational activities (continuing education) that are part of normal operating costs, and not as approved programs that are eligible for reasonable cost reimbursement. Many commenters strongly disagreed with the section of the proposed rule that included clinical pastoral education (CPE) as continuing education and stated that CMS must have been badly misinformed when writing the proposed rule. The commenters argued that CPE is a rigorous and structured education program accredited by the Association for Clinical Pastoral Education, Inc. (ACPE). The commenters stressed that, in varying amounts, CPE is a requirement for graduation for the master of divinity degree and for professional certification by the Association of Professional Chaplains (APC) as a health care chaplain, or as a CPE supervisor. Many commenters also noted prior Provider Reimbursement Review Board (PRRB) rulings that recognized chaplaincy as an allied health discipline, and asserted that hospitals that receive Medicare reasonable cost pass-through payment for CPE do so for the purpose of their professional CPE programs, not as continuing education for individuals already qualified to practice in hospital chaplaincy. Many commenters mentioned that the Joint Commission on Accreditation of Healthcare Organizations also recognizes chaplains as allied health professionals and considers them “primary care providers.” Similarly, commenters referred to various studies that have shown the positive spiritual and therapeutic benefits of pastoral care. The commenters warned that removal of funding for CPE would represent a huge step backward for American health care. The commenters urged CMS to ensure continuing pass-through payments for CPE.

Response: In the May 19, 2003 proposed rule (68 FR 27210), we stated that we received an inquiry requesting further clarification of what is meant by continuing education. We proceeded to explain what constitutes “continuing education” for the purpose of determining whether a nursing or allied health education activity would or would not qualify for Medicare reasonable cost pass-through payments. We acknowledge that the definition of “continuing education” for Medicare payment purposes may differ from the academic view of what, in general, constitutes such activities. In the proposed rule, we stated that we believed that provider-operated programs that do not lead to any specific certification or the ability to perform in the specialty would be classified as “continuing education.”

Our intent is to ensure that Medicare pass-through payments are only provided for programs that enable an individual to be employed in a capacity that he or she could not have been employed without having first completed a particular education program. We believe that, for Medicare purposes, training that enhances an individual's competencies, but does not permit that individual to be employed in a new capacity in which he or she could not have been employed without completing the additional training, would not qualify for Medicare reasonable cost pass-through payment. Medicare provides payments for such educational activities, but only under the methodology applicable to payment of normal operating costs. Our intent was simply to provide clarification for the purpose of distinguishing between those educational programs that qualify for reasonable cost pass-through payment (that is, programs that enable an individual to begin employment in a specialty as a whole) and those programs that should be paid as normal operating costs (that is, activities that are intended to enhance the current skill set of an individual's profession or advance an individual's professional career).

Since publication of the proposed rule, we have learned from information provided by the ACPE and the APC that there are several levels of CPE. Specifically, the ACPE accredits three different levels of CPE. The first level of CPE is generally geared to interns and beginning residents. The second level of CPE is generally geared to residents doing specialization and preparation for chaplaincy certification. The third level is supervisory training, which is geared toward preparation for certification by the ACPE as a CPE supervisor.

We understand that, as a part of the requirements for a master of divinity degree, many theological schools and seminaries require or strongly recommend completion of an internship, or 1 unit of CPE for graduation. A unit of CPE is 400+ hours of supervised CPE in a health care or institutional setting. Students taking either 1 or 2 units of CPE are generally referred to as interns. In addition, many faith groups require, at their national or regional levels, that individuals complete at least 1 unit of CPE in order for them to be ordained into professional ministry. Theological schools that offer doctoral degrees (for example, a doctor of philosophy, a doctor of ministry, or a doctor of theology) with specialties in pastoral counseling and related fields also generally require some amount of CPE as a part of those degree programs. Upon completion of a CPE internship, the health care institution typically reports to the theological school in which the student is enrolled that the student has successfully completed the internship, and the theological school subsequently awards credit for the training. Based upon information received from the commenters, we understand that completion of only an internship, or 400+ hours of CPE, would not qualify an individual for employment as a chaplain in a hospital setting.

In contrast to CPE internships, CPE residents generally participate in a 1-year, or occasionally a 2-year, full-time CPE program. A 1-year residency typically consists of 4 units of postgraduate CPE (that is, 1,600+ hours of supervised CPE), in a health care or institutional setting. Generally, individuals who undertake 1,600 hours of CPE do so in order to become a board-certified chaplain. The ACPE has established 4 units, or 1,600 hours of supervised CPE, as the national minimum amount of CPE that is required to become a board-certified chaplain. However, some certifying boards or particular programs may require some additional hours of CPE for board certification. We note that, in instances where academic credit is granted for completion of 1 unit, or 400 hours, of CPE prior to receipt of a degree, an individual seeking to become a board-certified chaplain generally must complete an additional 1,600 hours of CPE training.

The board certification of chaplains is carried out by nationally recognized organizations that are part of the Commission on Ministry in Specialized Settings (COMISS), an umbrella network for pastoral care organizations that share the same standards of educational preparation and clinical training. These organizations include the Association of Professional Chaplains (APC), the National Association of Catholic Chaplains (NACC), the National Association of Jewish Chaplains (NAJC), and the Canadian Association for Pastoral Practice and Education (CAPPE). The ACPE accredits CPE training for all of these certifying organizations.

Based on information received from the commenters, we understand that most health care organizations that are accredited by the Joint Commission on Accreditation of Healthcare Organizations (JCAHO) advertise for and recruit only board-certified chaplains, which means that qualified applicants for employment as hospital chaplains will usually have completed at least 1,600 hours of CPE.

Individuals who seek to develop a health care chaplaincy specialization (for example, hospice, pediatrics, cardiology, rehabilitation, neurology) may undertake a second year of CPE residency. A second year of residency consists of an additional 4 units of CPE (or 1,600+ hours of supervised CPE). However, there is currently no established board certification process for residents completing a second year of CPE residency training.

To be eligible to apply for supervisory CPE training, an individual must have completed at least 4 units (1 year) of CPE training. Upon completion of supervisory training, an individual becomes certified by the ACPE as a CPE supervisor and is qualified to develop and conduct CPE training for all ACPE-accredited programs.

Based on information submitted by the commenters on the different levels of CPE training, two important points relative to Medicare reimbursement have become clear to us. First, in instances where internship training is completed as a prerequisite for a degree granted by an educational institution other than a hospital, such training is not provider-operated, and, therefore, does not qualify for Medicare reasonable cost pass-through payment under § 413.85. Under § 413.85(f), a program is considered to be provider-operated only if the hospital directly incurs the training costs, directly controls the curriculum and the administration of the program, employs the teaching staff, and provides and controls both clinical training and classroom instruction (where applicable). While a hospital may serve as the site for a CPE internship, such training is provided to satisfy curriculum requirements of a theological school, which grants the master degree upon completion of the internship. While the hospital might incur training costs and employ the supervising faculty, it would not ordinarily meet the other “provider-operated” criteria concerning controlling the curriculum and providing both the didactic and clinical training necessary for the degree. Thus, a CPE internship, or any other CPE training that is a requirement for a degree, whether it is undergraduate, graduate, or doctoral, is not eligible for Medicare reasonable cost pass-through payment.

Secondly, a CPE residency consisting of 1,600 hours of training could be a provider-operated program and could also lead to certification and the ability to be employed in a new or different capacity. Specifically, a CPE residency consisting of approximately 1,600 hours of training leads to board certification in chaplaincy, and, as we understand it, most JCAHO-accredited hospitals generally only employ board-certified chaplains. In consideration of these facts, the costs of CPE training programs that meet the requirements under § 413.85, including accreditation by a nationally recognized accrediting body, direct operation by a provider, and lead to certification that is generally a requirement for employment in a particular specialty, may be eligible for Medicare reasonable cost pass-through payment.

In the May 19, 2003 proposed rule (68 FR 27210), we proposed to revise the regulations at § 413.85(h)(3) to state that activities treated as normal operating costs include “Educational seminars, workshops, and continuing education programs in which the employees participate that enhance the quality of medical care or operating efficiency of the provider and, effective October 1, 2003, do not lead to certification required to practice or begin employment in a nursing or allied health specialty.” We proposed to add a conforming definition of “certification” for purposes of nursing and allied health education under § 413.85(c) to mean “the ability to practice or begin employment in a specialty as a whole.” However, it is apparent from the comments we received that our proposed definition of “certification” was not clear. Some commenters believed we intended, through the proposed definition, to allow pass-through payments for the costs of a program that would only enhance an individual's set of skills. However, that was not our intent. We believe it would have been more appropriate to use the word “and” instead of the word “or”, to further emphasize that pass-through payment would only apply to activities that enable an individual to practice and begin employment in a specialty, but would not apply to activities that serve to add to or to enhance an individual's current skill set.

In addition, based on the comments received, we understand that there may be several distinct levels of training in a given health profession, and each level of training may be a requirement in order for an individual to work in a new capacity or “specialty” in that profession, but not a requirement to practice or begin employment in the specialty “as a whole.” Since a second level of training is not required to begin practicing in a profession, under the proposed definition, we would not have been able to allow for pass-through payments for a second (or potentially a third) level of training. Therefore, we understand that inclusion of the words “as a whole” in the proposed definition of “certification” was misleading. Consequently, where a subsequent level of training is a requirement to practice in a new specialty in a given profession, pass-through payment may be made for the subsequent level of training.

Finally, we have concluded that it is not necessary to include a specific definition of “certification” at § 413.85. In this final rule, we are deleting the proposed definition of “certification” from § 413.85(c), and amending § 413.85(h)(3) by removing the words “certification required” and inserting the words “the ability.” We are also changing the word “or” to “and”. Specifically, we are amending the proposed regulations at § 413.85(h)(3) to state that activities treated as normal operating costs include “Educational seminars, workshops, and continuing education programs in which the employees participate that enhance the quality of medical care or operating efficiency of the provider and, effective October 1, 2003, do not lead to the ability to practice and begin employment in a nursing or allied health specialty.”

Our view of a “specialty” in the nursing and allied health education context is based on what the industry views as the standard of practice in a specific area within a profession. The training required to allow a person to serve in the “specialty” is tailored to the skill level and context that an individual is expected to use in that “specialty.”

Consistent with what we stated in the proposed rule, Medicare reasonable cost pass-through payments are only provided for programs that, according to industry norms, qualify an individual to be employed in a specialty in which the individual could not have been employed before completing a particular education program. Given the confusion expressed by commenters, we recognize the need to specify how we will determine whether completion of a particular education program enables an individual to be employed in a specialty. We will use “industry norms” as the standard to determine whether participation in a specialty enables an individual to be employed in a capacity that he or she could not have been employed without having first completed a particular education program. We are defining “industry norm” to mean that more than 50 percent of hospitals in a random, statistically valid sample require the completion of a particular training program before an individual may be employed in a specialty. (We understand that, in some instances, due to the unique staffing circumstances faced by many smaller hospitals, inclusion of small hospitals in the sample would introduce factors that are not typically representative of the industry as a whole and would skew the results inappropriately. In such a case, if appropriate, we would consider excluding hospitals with less than 100 beds, which would still retain over 75 percent of all hospitals in the universe).

Based on comments received, we believe that it is the “industry norm” to require a CPE residency and board certification for employment as a hospital chaplain. Since it is currently the “industry norm” for hospitals to employ only board-certified chaplains, and since completion of approximately 1,600 hours of CPE training is a requirement to practice and begin employment in hospital chaplaincy, we view hospital chaplaincy as a “specialty” of pastoral counseling. Consequently, a hospital that operates a CPE residency may be eligible for reasonable cost pass-through payment.

Specifically, assuming all requirements under § 413.85 are met, Medicare reasonable cost pass-through payments may only be made to hospitals for CPE hours that are not prerequisites for any academic degree, and are provided to students in order to obtain board certification in hospital chaplaincy. A hospital may not receive reasonable cost payment for any costs incurred in connection with providing CPE that is undertaken to meet the requirements of an academic degree. In addition, since generally a minimum of approximately 1,600 hours of CPE is required to become a board-certified chaplain, any costs incurred for an individual participating in CPE training that exceeds the minimum number of hours required to obtain board certification would not be eligible to be paid on a reasonable cost basis.

However, we note that we do not completely defer to the information provided by industry representatives in order to determine the “industry norm.” Rather, if at any time we obtain information that calls our view of industry norms into question, we may make our own determination based on a random sample of hospitals. Therefore, assuming all other requirements under § 413.85 are met, a hospital may receive reasonable cost pass-through payment for the hours of CPE for which academic credit is not granted (since those CPE hours are not generally provider-operated), and for the hours of CPE that may be used to satisfy training requirements for board certification. We will continue to allow reasonable cost payment for CPE that leads to board certification as long as we do not have evidence indicating that, based on a statistically valid, random sample, the “industry norm” is not to require board certification for chaplains that are employed by hospitals.

We also recognize that industry norms are susceptible to change over time. Therefore, although it may not currently be the “industry norm” to require completion of a particular nursing or allied health education program in order to practice and begin employment in a particular specialty, it may become the “industry norm” in the future. If we find that it has become the “industry norm,” we may allow the hospitals operating those programs (and meeting the requirements at § 413.85) to be paid for the costs of those programs on a reasonable cost basis.

In relation to the commenters' recommendation that reasonable cost reimbursement should be provided for CPE supervisory training, we understand that, essentially, the purpose of the supervisory training is to prepare a chaplain to develop CPE programs and to teach interns and residents. We believe that CPE supervisors are practicing in the teaching profession, not within a nursing or allied health discipline. Furthermore, we do not believe that Congress intended to provide for reasonable cost pass-through payments for programs that are intended to produce instructors or teachers. While we recognize that CPE supervisors are necessary to train and prepare individuals for hospital chaplaincy, we believe that it is appropriate for the costs of supervisory programs in general to be treated as normal operating costs and paid accordingly.

Comment: One commenter stated that our proposed definition of provider-operated programs intended to exclude programs “that do not lead to certification required to practice or begin employment in a nursing or allied health specialty * * *” is not appropriate in light of the growing number of skills that require intensive clinical experiences. Another commenter stated that this proposal will seriously hinder reversal of the nursing shortage across the nation and, as a result, will have an adverse impact on the quality and safety of care provided in hospitals. The commenters used the example of nurse residencies, which a number of hospitals across the country are hosting for registered nurses. The commenters explained that these residencies, which are postgraduate and typically last 1 year, are designed to equip the newly licensed nurse with the skills to care for patients who require the most complex and sophisticated diagnostic and therapeutic services, and to prepare the nurses for leadership roles earlier in their careers and give them the tools to improve the quality of care and reduce medical errors. The commenters claimed that the Federal Government has thus far provided minimal funding to help ameliorate the nursing shortage and, therefore, the proposed rule is particularly distressing. They urged CMS to include criteria in the final rule for pass-through payment of nurse residencies.

Response: First, we do not believe that nurse residencies, which are intended to help integrate newly licensed nurses into complex acute care environments by enhancing their competencies and skills, are programs that qualify these nurses to be employed in a new specialty. Accordingly, it is more appropriate to treat such activities as normal operating costs. As we stated above, Medicare reasonable cost pass-through payment will only be provided for programs that, according to industry norms, qualify an individual to be employed in a specialty in which the individual could not have been employed prior to completing a particular education program. Second, we note that nurse residencies do not qualify for reasonable cost payment because they also do not meet the requirement for accreditation by a national approving body under § 413.85(d)(1)(i)(A). Therefore, while we are sympathetic to the commenters' concerns, we do not believe that it is appropriate at the present time to allow for pass-through payment to be made under the Medicare program for nurse residencies.

Comment: Some commenters stated that CMS was “entirely correct” in identifying CPE as continuing education and concurred with our proposal to discontinue pass-through payments for CPE. One commenter contended that ACPE-accredited training is not primarily used to prepare students to be health care chaplains. Rather, CPE is primarily ministry training, and there are various ways that one can choose to use CPE. One commenter added that very few individuals who train in CPE, including those individuals in 1-year residencies, become employed as health care chaplains. The commenter further stated that CPE is “properly a funding responsibility of the church rather than the government”. The commenters argued that Medicare should not be supporting continuing education for religious care providers whose primary base and certifying group is their denomination or faith group.

Another commenter presented a similar argument concerning pharmacy residencies and questioned why Medicare (that is, taxpayers) should subsidize these residency programs. The commenter claimed that hospitals “use government monies in order to hire these ‘residents,' utilize them in ‘clinical' positions under the guise of postgraduate training, thereby bypassing having to use FTEs in the hospital pharmacy budget.” The commentator believed that if hospitals and pharmacists were truly concerned with improving patient care, hospital pharmacy departments would train their own staff pharmacists to perform the clinical aspects themselves, rather than having taxpayers provide the funding.

Response: We are sympathetic to the commenters' concerns. However, we understand that many CPE programs do occur in hospitals, and that, while there may be various kinds of CPE training, generally, completion of approximately 1,600 hours of CPE training is required for board certification and employment by a hospital. Therefore, we believe that CPE residencies that lead to board certification generally would not be considered continuing education.

In response to the commenters' concerns about the taxpayers, through the Medicare program, providing support for CPE and pharmacy residencies, we note Medicare payment for these and other similar programs are made in accordance with the Medicare statute. Under section 1861(v) of the Act, Congress provides for Medicare payments to be made in support of certain medical education activities. Currently, if a program meets the regulatory requirements under § 413.85, which were specified earlier in this preamble, a hospital operating that program may qualify for Medicare reasonable cost pass-through payment.

Comment: One commenter explained that a dietetic internship is a post-baccalaureate program that is one of the requirements for practicing as a registered dietitian. The commenter pointed out that the Commission on Accreditation of Dietetic Education (CADE) of the American Dietetic Association accredits these internships and the interns contribute directly to patient care in a hospital. The commenter urged us to continue to pay health care organizations for dietetic internships.

Response: We appreciate the comment and note that, as long as a dietetic internship meets the requirements under § 413.85 (and we do not find that it is not the industry norm to require this training to be employed as a registered dietitian), the hospital operating the internship may qualify for Medicare reasonable cost pass-through payment.

Comment: A large number of commenters responded to our proposal to clarify that, effective October 1, 2003, training that does not lead to certification required to practice or begin employment in a nursing or allied health specialty would be treated as educational activities (continuing education) that are part of normal operating costs, and not as approved programs that are eligible for reasonable cost pass-through payments. Many commenters strongly disagreed with our proposal that included pharmacy residencies in the type of training that is considered continuing education and claimed that the proposed rule reflected a fundamental misunderstanding of pharmacy education. The commenters stated that educational seminars, workshops, and continuing education programs are generally performed outside the provider setting, and in most instances do not exceed 40 hours per year, whereas a pharmacy residency is a full-time commitment that lasts for 1 year. The commenters emphasized that the pharmacy residencies are structured, intensive programs that incorporate direct patient care experience where residents work as part of a clinical team and are required to complete a comprehensive project. The commenters contended that residency experience provides focused, invaluable training that yields proven positive clinical and financial outcomes. The commenters also noted that, while residencies are not a requirement for all hospital pharmacy positions, they are a requirement for most clinical specialist positions. The commenters maintained that residencies would be a more universal hiring requirement were it not for the current shortage of pharmacists and residency programs. The commenters stressed the benefits of clinical pharmacist involvement in patient care and cautioned that CMS' attempt at short-term cost savings will result in significant long-term cost of care increases. The commenters urged CMS to ensure continuing reasonable cost pass-through payments for pharmacy residencies.

Response: As we stated above in response to the comments received from the clinical pastoral counseling community, in the May 19, 2003 proposed rule (68 FR 27210), we explained what constitutes “continuing education” for the purpose of determining whether a nursing or allied health education activity would or would not qualify for Medicare reasonable cost pass-through payments. We acknowledge that the definition of “continuing education” for Medicare payment purposes may differ from the academic view of what, in general, constitutes such activities. As we stated earlier, we believe that provider-operated programs that do not lead to any specific certification, or the ability to perform in the specialty, would be classified as “continuing education.”

Our intent is to ensure that Medicare reasonable cost pass-through payments are only provided for programs that enable an individual to be employed in a capacity that he or she could not have been employed without having first completed a particular education program. We believe that, for Medicare purposes, training that enhances an individual's competencies, but does not permit that individual to be employed in a new specialty in which he or she could not have been employed without completing the additional training, would not qualify for Medicare reasonable cost pass-through payment. Medicare provides payment for such educational activities, but only under the methodology applicable to payments for normal operating costs. Our intent was to provide clarification for the purpose of distinguishing between those educational programs that qualify for reasonable cost pass-through payment (that is, programs that enable an individual to begin employment in a specialty), and those programs that should be paid as normal operating costs (that is, activities that are intended to enhance the current skill set of an individual for a profession or advance an individual's professional career).

Since publication of the proposed rule, we have learned from information provided by the commenters that there are two categories of pharmacy residencies—pharmacy practice residencies and specialized pharmacy residencies, both of which are accredited by the American Society of Health-System Pharmacists (ASHP). If a pharmacist chooses to participate in residency training, he or she would generally do so after completion of an undergraduate bachelor of science degree or a doctor of pharmacy degree. (In some cases, residencies are offered as a part of a postgraduate degree (a master of science or a doctor of pharmacy). However, these programs would not meet our provider-operated criteria.) A pharmacy practice residency is typically a 1-year, organized, directed, postgraduate training program in a defined area of pharmacy practice that may take place in a variety of settings, including hospitals. For those seeking additional skills in a focused area of pharmacy practice (for example, oncology), an individual may choose to complete a second year of specialized pharmacy residency. Currently, ASHP, in partnerships with other professional organizations, accredits 17 second-year pharmacy residencies, in areas such as cardiology, geriatrics, infectious diseases, and oncology.

Of the 17 second-year pharmacy residencies, only 5 of these residencies currently lead to board certification. The Board of Pharmaceutical Specialties (BPS) is the organization that administers the certifying examinations after completion of each of these five residencies. Upon completion of a residency in 1 of the other 12 second-year residencies, the hospital in which the resident has trained issues a certificate to the pharmacist.

We understand that many employers, including hospitals, increasingly are requiring completion of an ASHP-accredited first year pharmacy practice residency as a condition for employment as a clinical (“on the floor”) or direct patient care pharmacist. While a licensed pharmacist who has not completed a pharmacy practice residency might be hired by a hospital as a staff or distribution pharmacist, a hospital typically would only hire an individual who has completed at least a 1-year pharmacy practice residency to fill a position that requires direct work with hospital patients. Some hospitals may even require their pharmacists to have completed a second-year specialized residency before allowing those pharmacists to specialize on a particular group or type of patients. For example, before a pharmacist may work exclusively to design, implement, and monitor a course of treatment for oncology patients, some hospitals require that the pharmacist complete a residency in oncology pharmacy. However, many hospitals may employ pharmacists who have only completed a pharmacy practice residency to treat these groups or types of patients, including oncology patients.

As we explained above in response to the comments on CPE, in the May 19, 2003 proposed rule (68 FR 27210), we proposed to revise the regulations at § 413.85(h)(3) to state that activities treated as normal operating costs include “Educational seminars, workshops, and continuing education programs in which the employees participate that enhance the quality of medical care or operating efficiency of the provider and, effective October 1, 2003, do not lead to certification required to practice or begin employment in a nursing or allied health specialty.” We proposed to add a conforming definition of “certification” for purposes of nursing and allied health education under § 413.85(c) to mean “the ability to practice or begin employment in a specialty as a whole.” However, it is apparent from the comments we received that our proposed definition of “certification” was not clear. Some commenters believed we intended, through the proposed definition, to allow pass-through payments for the costs of a program that would only enhance an individual's set of skills. However, that was not our intent. We believe it would have been more appropriate to use the word “and” instead of the word “or” to further emphasize that pass-through payment would only apply to activities that enable an individual to practice and begin employment in a specialty, but would not apply to activities that serve to add to or to enhance an individual's current skill set.

In addition, based on the comments received, we understand that there may be several distinct levels of training in a given health profession, and each level of training may be a requirement in order for an individual to work in a new capacity or “specialty” in that profession, but not a requirement to practice or begin employment in the specialty “as a whole.” Since a second level of training is not required to begin practicing in a profession, under the proposed definition, we would not have been able to allow for pass-through payments for a second (or potentially a third) level of training. Therefore, we understand that inclusion of the words “as a whole” in the proposed definition of “certification” was misleading. Consequently, where a subsequent level of training is a requirement to practice in a new specialty in a given profession, pass-through payment may be made for the subsequent level of training.

Finally, we have concluded that it is not necessary to include a specific definition of “certification” in the regulations at § 413.85. In this final rule, we are deleting the proposed definition of “certification” from § 413.85(c), and amending § 413.85(h)(3) by removing the words “certification required” and inserting the words “the ability.” We are also changing the word “or” to “and”. Specifically, we are amending the proposed § 413.85(h)(3) to state that activities treated as normal operating costs include “Educational seminars, workshops, and continuing education programs in which the employees participate that enhance the quality of medical care or operating efficiency of the provider and, effective October 1, 2003, do not lead to the ability to practice and begin employment in a nursing or allied health specialty.”

As we stated above in response to the comments concerning CPE, our view of a “specialty” in the nursing and allied health education context is based on what the health care industry views as the standard of practice in a specific area within a profession. We are defining “industry norm” to mean that more than 50 percent of hospitals in a random, statistically valid sample require the completion of a particular training program before an individual may be employed in a specialty. (We understand that, in some instances, due to the unique staffing circumstances faced by many smaller hospitals, inclusion of small hospitals in the sample would introduce factors that are not typically representative of the industry as a whole and would skew the results inappropriately. In such cases, we would consider excluding hospitals with less than 100 beds, which would still retain over 75 percent of all hospitals in the sample universe.)

Based on comments received, we believe that it is currently the “industry norm” for hospitals to generally hire only pharmacists who have completed a pharmacy practice residency to work directly in patient care. Specifically, without having completed a pharmacy practice residency, a pharmacist would typically be employed by a hospital as a staff or distribution pharmacist, but not as a clinical pharmacist who works directly with patients to develop treatment plans. Since completion of a pharmacy practice residency has become a requirement by hospitals to practice or begin employment in a position that involves direct patient care, we would view “hospital pharmacy” as a “specialty” of the pharmacy profession. Accordingly, pharmacy practice residency training programs that meet the requirements under § 413.85, including accreditation by a nationally recognized accrediting body, direct operation by a provider, and lead to certification that is a requirement for employment, may be eligible for Medicare reasonable cost pass-through payment.

However, it is apparent from the comments that it is not unusual for a hospital to employ a pharmacist that has only completed a pharmacy practice residency in an area in which an accredited second-year program exists (that is, geriatrics, cardiology, or oncology), without requiring the pharmacist to first complete that second-year residency program. For example, we would view further training in oncology pharmacy or cardiology pharmacy as specializations within the pharmacy field under the policy in this final rule. However, these second-year residencies would not qualify for reasonable cost pass-through payment because, based on information received from commenters, it is not currently the “industry norm” to require completion of these programs before beginning work in these specialties. If we find in the future that it has become the “industry norm” for hospitals to require second-year pharmacy residencies, we may allow the hospitals operating those programs to be reimbursed for the costs of those programs on a reasonable cost basis.

3. Programs Operated by Wholly-Owned Subsidiary Educational Institutions of Hospitals

Another matter that has come to our attention since publication of the January 12, 2001 final rule (66 FR 3363) on nursing and allied health education concerns the preamble language of the rule, which states:

“Concerning those hospitals that have established their own educational institution to meet accrediting standards, we believe that, in some cases, these providers can be eligible to receive payment for the classroom and clinical training of students in approved programs. If the provider demonstrates that the educational institution it has established is wholly within the provider's control and ownership and that the provider continues to incur the costs of both the classroom and clinical training portions of the program, the costs would continue to be paid on a reasonable cost basis. An independent college would not meet these criteria.

“An example of a program that could be considered provider-operated would be one in which the hospital is the sole corporate member of the college, elects the board of trustees, has board members in common, employs the faculty and pays the salaries, controls the administration of the program and the curriculum, and provides the site for the clinical and classroom training on the premises of the hospital. We believe that, in these situations, the community has not undertaken to finance the training of health professionals; the provider has merely restructured its provider-operated program to meet certain State or accrediting requirements. In most cases, providers have aligned themselves with already established educational institutions. We note that a program operated by an educational institution that is related to the provider through common ownership or control would not be considered to meet the criteria for provider operated.” (66 FR 3363)

We have received a question from a hospital that pertains to the cited preamble language in the narrow circumstance where the hospital previously received Medicare reasonable cost payment for direct operation of nursing or allied health education programs and then established its own wholly owned subsidiary college to operate the programs, in order to meet accreditation standards. The hospital has continued to receive Medicare payments after the hospital moved operation of the programs to the wholly owned subsidiary college. The hospital believes that, based on the cited preamble language regarding wholly owned subsidiary colleges and the lack of prior specific guidance on this particular organizational structure (as well as its continued receipt of pass-through payments) and because the hospital continues to pay all of the costs of the nursing and allied health education programs, the hospital is still the direct operator of the programs and should continue to receive pass-through treatment. However, we believe that once the hospital moved the direct operation of its nursing and allied health education programs to the college, the programs no longer met our provider-operated criteria at § 413.85(f). At the very least, it appears that the hospital did not hire the faculty for the program(s) and did not have direct control of the curriculum of the program(s) after operation was transferred to the wholly owned subsidiary college. As we stated in the preamble language quoted above: “a program operated by an educational institution that is related to the provider through common ownership or control would not be considered to meet the criteria for provider operated” (66 FR 3363).

However, we understand that some hospitals, including this hospital, may have interpreted the preamble language that stated, “if the provider demonstrates that the educational institution it has established is wholly within the provider's control and ownership and that the provider continues to incur the costs of both the classroom and clinical training portions of the program, the costs would continue to be paid on a reasonable cost basis” (Ibid.), to mean that hospitals that establish wholly owned subsidiary colleges or educational institutions would continue to receive Medicare reasonable cost payment if the hospitals incur the costs of the classroom instruction and clinical training. In the May 19, 2003 proposed rule, we proposed to clarify that transferring operation of previously provider-operated programs to educational institutions, even if the institutions are wholly owned by the hospital, does not necessarily mean that the programs continue to meet our provider-operated criteria under § 413.85(f). In order to remain provider operated, the hospital must have direct control of the program; the hospital itself must employ the teaching staff, have direct control of the program curriculum, and meet other requirements, as stated at § 413.85(f).

While we proposed to clarify that merely operating programs through a wholly owned subsidiary college does not constitute direct operation of nursing or allied health education programs unless the hospital itself meets the requirements of the regulations at § 413.85(f), we believe it would be unfair to recoup Medicare payments that have already been made to hospitals that meet this very narrow fact pattern. Therefore, we proposed that Medicare would not recoup reasonable cost payment from hospitals that have received pass-through payments for portions of cost reporting periods occurring before October 1, 2003 for the nursing or allied health education program(s) where the program(s) had originally been operated by the hospital, and then operation of the program(s) had been transferred by the hospital to a wholly owned subsidiary educational institution in order to meet accreditation standards prior to October 1, 2003, and where the hospital had continuously incurred the costs of both the classroom and clinical training portions of the programs at the educational institution.

In addition, we proposed that, for portions of cost reporting periods occurring on or after October 1, 2003, such a hospital would continue to receive reasonable cost payments for the clinical training costs incurred by the hospital for the program(s) described above that were previously provider operated. However, we further proposed that, with respect to classroom costs, only those classroom costs incurred by the hospital for the courses that were paid by Medicare on a reasonable cost basis and included in the hospital's provider-operated program(s) could continue to be reimbursed on a reasonable cost basis. That is, Medicare would pay on a reasonable cost basis for the classroom costs associated with the courses provided as part of the nursing and allied health education programs (for example, the courses relating to the theory and practice of the particular nursing and allied health discipline(s)) that were offered by the hospital when the hospital was the direct operator of the program(s).

We believe the proposed policy is appropriate since continued pass-through payment will allow these hospitals to maintain equal footing with other hospitals that receive pass-through payments and have maintained their provider-operated programs. In addition, it would not be equitable to discontinue longstanding Medicare pass-through payment to these hospitals (in fact, reasonable cost payment to at least one of these hospitals for nonprovider-operated programs preceded the publication of the January 12, 2001 final rule on nursing and allied health education payments by many years) that restructured operation of their nursing and allied health education program(s) as wholly owned subsidiaries in order to meet accreditation standards while relying on their understanding of CMS' prior expressions of provider-operated requirements and the recent preamble language. If these providers were now forced to restructure in order to meet the requirements of § 413.85(f), they would not be able to maintain their accreditation.

We note that Congress has specifically expressed its intent that providers that have restructured their programs to be operated by a wholly owned subsidiary educational institution in order to meet accreditation standards should continue to receive Medicare reasonable cost payment. In the conference report accompanying the Consolidated Appropriations Resolution for FY 2003, Congress stated:

“The conferees are particularly concerned about nursing and allied health educational programs that cannot meet the regulations set forth at 42 CFR 413.85(f) solely as a result of regional educational accrediting criteria. Given the shortage of nursing and allied health professionals, the conferees support the payment of costs on a reasonable cost basis for a hospital that has historically been the operator of nursing and allied health education programs(s) that qualified for Medicare payments under 42 CFR 413.85, but, solely in order to meet educational standards, subsequently relinquishes some control over the program(s) to an educational institution, which meets regional accrediting standards; is wholly owned by the provider; and is supported by the hospital, that is, the hospital is incurring the costs of both the classroom and clinical training of the program.” (H.R. Rep. No. 108-10, 108th Cong., 1st Sess., 1115 (2003).)

However, we note that the proposed policy would not allow these hospitals to be paid for additional classroom costs for courses that were not paid on a reasonable cost basis to the hospitals in conjunction with their provider-operated programs (for example, additional classes needed to meet degree requirements). We believe that to allow pass-through payment for those additional costs would provide these hospitals with an unfair advantage over other hospitals with provider-operated programs.

We note that any hospital that chooses to restructure its programs to be operated by a wholly-owned subsidiary educational institution on or after the effective date of this proposal when finalized (October 1, 2003) would not be eligible for pass-through payments under the proposed provision unless the hospital continues to meet the requirements of § 413.85(f). We believe it is appropriate to limit the proposed payments to hospitals that restructured before October 1, 2003 because our policy with respect to programs by a wholly-owned subsidiary of a hospital will have been clarified by that date (the date that this final rule is effective).

We proposed to revise § 413.85 by adding new paragraphs (d)(1)(iii) and (g)(3) to reflect the proposed payment policy.

Comment: Several comments supported our proposal. Specifically, the commenters believed that the proposed rule is consistent with the recent expressions of Congressional intent reflected in the conference report to the 2003 Consolidated Appropriations Resolution, which recognize that there is a shortage of nursing and allied health professionals, and that payments made for programs that are operated by wholly-owned subsidiary educational institutions of hospitals should not be retrospectively recouped and may continue in the future.

However, several commenters disagreed with the proposal under proposed § 413.85(g)(3)(iii) that, effective for portions of cost reporting periods occurring on or after October 1, 2003, eligible hospitals could receive payment for the clinical training costs and for the classroom costs, but only those classroom costs incurred by the hospital for the courses that were included in the program(s) that had originally been provider-operated before transfer of operation of the program(s) to a wholly owned subsidiary educational institution. One commenter stated that such criteria regarding reimbursement of classroom costs appears to presume that while a hospital was operating its own program before transferring the operation of the program to a wholly-owned subsidiary, the hospital must have offered fewer or different programs. The commenter believed that our example in the preamble of the proposed rule seems to suggest that “noncore” or nonnursing related classes would be excluded from reasonable cost reimbursement, effective October 1, 2003. The commenter contended that we have incorrectly assumed that diploma programs include only nursing courses because, in fact, such diploma programs typically included general courses for English, basic science, math, and similar subjects. The commenter asked that we revise the preamble to clarify that courses for which costs were historically reimbursed would continue to qualify for reasonable cost payment without regard to whether they are “core” or “noncore” nursing courses.

Other commenters argued that restricting reimbursement to courses originally offered by the provider-operated program would discourage providers from ensuring that training of health care professionals is kept up to date and would not allow providers to meet evolving requirements of accrediting organizations. One commenter noted that the conference report accompanying the Consolidated Appropriations Resolution for FY 2003 states that “* * * the conferees support the payment of costs on a reasonable cost basis for a hospital that has historically been the operator of nursing and allied health education program(s) * * *” (Emphasis added) (H.R. Rept. No. 108-10, 108th Cong., 1st Sess., 1115 (2003)). The commenter believed this language indicates that Congress intended that schools should be reimbursed, not particular courses.

In addition, commenters expressed concern that capping reimbursement for educational programs effective October 1, 2003, would further aggravate the existing shortage of appropriately trained healthcare workers. Finally, commenters suggested that the October 1, 2003 effective date be postponed because this date will cause hardship for institutions currently in the process of creating educational organizations for the purpose of transitioning their programs to those educational organizations.

Response: We acknowledge the commenters' general support of the proposed changes. In response to the commenters who disagreed with our proposal for limiting payment to certain classroom costs, as we stated in the preamble to the proposed rule (68 FR 27210), this proposed exception to the reasonable cost payment policy for programs operated by wholly-owned subsidiary educational institutions was based on a question that we received from a hospital pertaining to the language in the January 12, 2001 Federal Register (66 FR 3363) concerning hospitals that established their own educational institutions to meet accreditation standards. Specifically, the hospital that raised the issue previously received Medicare reasonable cost payment for the direct operation of nursing and allied health education programs and then established its own wholly-owned subsidiary college to operate the programs, in order to meet accreditation standards. The hospital in question has continued to receive Medicare payments after the hospital moved operation of the programs to the wholly-owned subsidiary college. The hospital believed that, based on the cited preamble language in the January 12, 2001 Federal Register regarding wholly owned subsidiary colleges and the lack of prior specific guidance on this particular organizational structure (as well as its continued receipt of pass-through payments) and because the hospital continues to pay all of the costs of the nursing and allied health education programs, that it is still the direct operator of the programs and should continue to receive pass-through treatment.

As we stated in the proposed rule, we believe that once the hospital moved the direct operation of its nursing and allied health education programs to the college, the programs no longer met our provider-operated criteria at § 413.85(f). As we stated in the preamble language quoted above: “a program operated by an educational institution that is related to the provider through common ownership or control would not be considered to meet the criteria for provider operated” (66 FR 3363).

We explained that we understood that some hospitals may have interpreted the preamble language that stated, “if the provider demonstrates that the educational institution it has established is wholly within the provider's control and ownership and that the provider continues to incur the costs of both the classroom and clinical training portions of the program, the costs would continue to be paid on a reasonable cost basis' (Ibid.), to mean that hospitals that establish wholly owned subsidiary colleges or educational institutions would continue to receive Medicare reasonable cost payment if the hospitals incur the costs of the classroom instruction and clinical training. Accordingly, although we proposed to clarify in the proposed rule that, in general transferring operation of previously provider-operated programs to educational institutions, even if the institutions are wholly owned by the hospital, does not necessarily mean that the programs continue to meet our provider-operated criteria under § 413.85(f), we believed it would be unfair to recoup Medicare payments that have already been made to such a hospital that meets this very narrow fact pattern. Therefore, we proposed to add a limited exception to § 413.85 to reflect the unique circumstances of such a hospital.

First, we proposed that, for portions of cost reporting periods occurring on or before October 1, 2003, Medicare would not recoup reasonable cost payment from such a hospital that has received pass-through payments for the nursing or allied health education program(s) where the program(s) had originally been operated by the hospital, and then operation of the program(s) had been transferred by the hospital to a wholly owned subsidiary educational institution in order to meet accreditation standards prior to October 1, 2003, and where the hospital had continuously incurred the costs of both the classroom and clinical training portions of the programs at the educational institution.

Second, since we believed that such a hospital's programs were no longer provider-operated, and therefore, should not continue in the future to receive full reasonable cost payments for the clinical and classroom costs of programs that are now operated by the wholly owned subsidiary educational institution, we proposed that, for portions of cost reporting periods occurring on or after October 1, 2003, such a hospital would continue to receive reasonable cost payments for the clinical training costs incurred by the hospital for the program(s) described above that were previously provider operated. However, we further proposed that, with respect to classroom costs, only those classroom costs incurred by the hospital for the courses that were paid by Medicare on a reasonable cost basis and were included in the hospital's provider-operated program(s) could continue to be reimbursed on a reasonable cost basis. That is, we proposed that Medicare would pay on a reasonable cost basis for the classroom costs associated with the courses provided as part of the nursing and allied health education programs (for example, the courses relating to the theory and practice of the particular nursing and allied health discipline(s)) that were offered by the hospital when the hospital was the direct operator of the program(s).

In proposing that, effective for portions of cost reporting periods occurring on or after October 1, 2003, we would only continue to pay on a reasonable cost basis for classroom costs associated with the courses that relate to the theory and practice of the particular nursing or allied health discipline(s) that were offered by the hospital when the hospital was the direct operator of the program(s), and not for additional classes needed to meet degree requirements provided as part of the nursing or allied health education programs, we did assume, as a commenter suggested, that diploma nursing programs typically only include courses related to the theory and practice of nursing. However, regardless of whether diploma programs include additional general courses other than “core” nursing courses, we continue to believe it is more appropriate to pay a hospital that meets the limited exception that allows continued payment for only those costs associated with courses included in the program(s) when the hospital was still the direct operator of the program(s). If, in fact, a hospital that meets the limited exception currently offers the same courses that it had offered when it was still the direct operator of the programs, we would continue to pay for the classroom costs associated with those courses, even if those courses do not relate directly to the theory and practice of the nursing or allied health program(s). However, if new courses, whether or not they are nursing-related or allied health-related course, have been added after the operation of the program(s) was transferred to a wholly owned subsidiary educational institution, we would not pay on a reasonable cost basis for the classroom costs associated with those new courses, effective October 1, 2003. If the courses offered currently are the same as the courses offered prior to transfer of the programs to the wholly owned subsidiary, but, for example, the names of the courses have changed, or there have been course substitutions, we would evaluate each course on an individual basis to determine whether we would continue to allow reasonable cost payment for those courses. All other things being equal (that is, after adjusting for inflation and changes in enrollment), our intent is not to pay more on a reasonable cost basis as of October 1, 2003, for classroom costs to such a hospital than we had paid to the hospital when the hospital was still the direct operator of the program(s).

In response to the comments we received that urged us not to restrict the number of courses for which we would provide reasonable cost reimbursement due to concerns about evolving accreditation requirements and the existing nursing shortage, we emphasize again that this proposal is not at all broad in scope. Rather, based on the information we currently have available to us, we believe this provision would have a limited application.. Therefore, we do not believe that our proposal will aggravate the nursing shortage or adversely affect hospitals that otherwise meet the requirements for reasonable cost payment under § 413.85 but add courses to their programs. Similarly, we do not believe that the effective date of October 1, 2003, will cause hardship to other providers that are currently in the process of transitioning their programs to educational organizations, since the proposed changes would only apply to a provider that had already created its own educational institution. We also note that, as indicated above, programs that transition in some respect to educational institutions created by providers could possibly be considered “provider-operated” under § 413.85(f) and, if all other requirements are met, could qualify to receive reasonable cost reimbursement.

Comment: One commenter disagreed with our statement in the proposed rule (68 FR 27211) that “* * * transferring operation of previously provider-operated programs to educational institutions, even if the institutions are wholly owned by the hospital, does not necessarily mean that the programs continue to meet our provider-operated criteria under § 413.85(f).” Rather, the commenter believed that programs that are wholly owned or wholly controlled by a hospital are provider-operated programs. The commenter asserted that CMS” distinction between provider-operated programs and wholly owned programs conflicts with CMS” regulations at § 413.17(c)(2) which state that “If the provider obtains items of services, facilities, or supplies from an organization, even though it is a separate legal entity, and the organization is owned or controlled by the owner(s) of the provider, in effect the items are obtained from itself.” The commenter also referenced § 412.2(c)(5)(i) concerning the DRG 3-day payment window that applies to services provided by a hospital or by an entity wholly owned or operated by the hospital, and asserted that there is “no rational basis” for treating wholly owned or wholly controlled affiliates differently for purposes of pass-through payment.

Response: The commenter is incorrect in stating that, in the proposed rule, we indicated that wholly owned (or wholly controlled) programs by definition cannot meet the provider-operated criteria and, therefore, would not qualify for reasonable cost pass-through payments. In fact, as we have stated in the January 12, 2001 final rule (66 FR 3363), and reiterated in the preamble to the proposed rule, if the hospital that wholly owns the educational institution meets the provider-operated criteria, the hospital would qualify to receive reasonable cost pass-through payment. Specifically, we stated in the proposed rule (68 FR 27210) that “Concerning those hospitals that have established their own educational institution to meet accrediting standards, we believe that, in some cases, these providers can be eligible to receive payment for the classroom and clinical training of students in approved programs. * * * An example of a program that could be considered provider-operated would be one in which the hospital is the sole corporate member of the college, elects the board of trustees, has board members in common, employs the faculty and pays the salaries, controls the administration of the program and the curriculum, and provides the site for the premises of the hospital (emphasis added). Thus, while we still believe that transferring operation of previously provider-operated programs to educational institutions, even if the institutions are wholly owned by the hospital, does not necessarily mean that the programs continue to meet our provider-operated criteria under § 413.85(f) (68 FR 27211), we reiterate that only in instances where the hospital continues to meet the provider-operated criteria under § 413.85(f) would the hospital continue to qualify for reasonable cost pass-through payments, as it did prior to transferring operation of a provider-operated program(s) to a wholly owned educational institution.

The commenter also mentioned the generally applicable “related-entity” rules, and suggested that a wholly owned school would be a related entity that should be treated as if it is the provider. Thus, a wholly owned educational institution would remain provider-operated. However, we note that, for purposes of nursing or allied health education payment under § 413.85, it is not sufficient for a program to be operated by a related entity. Rather, the “related entity” principles do not apply under the agency's nursing and allied health education payment policy because, as indicated in previous rulemakings, that policy requires that a program be directly operated by the provider itself. Requiring direct operation of a program by the provider ensures that, under § 413.85(c), costs borne by related organizations (that is, the community) are not redistributed to the hospital and claimed as a pass-through under the Medicare program.

Comment: Commenters requested clarification on whether the proposed change regarding providers that created wholly owned subsidiary educational institutions to meet accreditation requirements would have any effect on provider-operated nursing or allied health programs that have entered into written contracts with colleges or universities to award their degrees.

Response: As we have explained in response to a previous comment, the proposed change was extremely limited in scope and only relates to hospitals with a unique set of circumstances surrounding operation of their programs by a wholly owned subsidiary educational institution. Therefore, the proposed changes do not have any impact on existing policy related to hospitals that enter into contracts with academic institutions to award their degrees. However, we stress that, in the instance where an academic institution other than the hospital grants the final certificate or degree upon completion of the program, the burden of proof is on the hospital to demonstrate that it, in fact, meets the “provider-operated” criteria under § 413.85(f) before reasonable cost payment may be made to that hospital.

Comment: One commenter believed that it is inappropriate to use the term “wholly owned” in reference to entities that, in many cases, are nonprofit institutions because, technically, nonprofit organizations are public trusts. The commenter suggested that it would be more accurate to refer to “wholly owned” or “wholly controlled” educational institutions.

Response: We believe that, for purposes of payment under § 413.85, it is appropriate to use the term “wholly owned.” Although we recognize that nonprofit entities would not technically be “wholly owned” since they do not issue stock, we do not agree with the commenter that “wholly controlled” is an appropriate alternative because of the potential for confusion over issues relating to “control” and “provider operation.” Further, we believe that the term “wholly owned” is commonly used in the context of nonprofit entities, and implies the kind of relationship we intend—where there is a single founder or member. Therefore, we will continue to use the term “wholly owned subsidiary” in the context of payment under § 413.85.

We are finalizing the two proposals associated with programs operated by wholly owned subsidiary educational institutions of hospitals. Specifically, we are finalizing the proposal under new § 413.85(g)(3) that, effective for portions of cost reporting periods occurring on or after October 1, 2003, a provider that incurs costs for a nursing or allied health education program(s) where those program(s) had originally been provider-operated, and then operation of the programs) was transferred to a wholly owned subsidiary educational institution in order to meet accreditation standards prior to October 1, 2003, and where the provider has continuously incurred the costs of both the classroom and clinical training portions of the program(s) at the educational institution, may receive reasonable cost payment for such a program(s). Further, reasonable cost payment will be made if a provider received reasonable cost payment for those nursing and allied health education program(s) both prior and subsequent to the date the provider transferred operation of the program(s) to this wholly owned subsidiary educational institution (and ceased to be provider-operated program(s)). Such a provider would receive reasonable cost payments for: (a) The clinical training costs incurred for the program(s), and (b) classroom costs, but only those classroom costs incurred by the provider for the courses that were included in the programs that were originally provider-operated prior to the transfer to a wholly owned subsidiary educational institution. That is, Medicare would pay on a reasonable cost basis for the classroom costs associated with the courses provided as part of the nursing or allied health education programs that were offered by the hospital when the hospital was the direct operator of the program(s). We would not allow such a hospital to be paid for additional classroom costs for courses that were not paid on a reasonable cost basis to the hospital in conjunction with its provider-operated programs.

F. Payment for Direct Costs of Graduate Medical Education (§ 413.86)

1. Background

Under section 1886(h) of the Act, Medicare pays hospitals for the direct costs of graduate medical education (GME). The payments are based in part on the number of residents trained by the hospital. Section 1886(h)(4)(F) of the Act caps the number of allopathic and osteopathic residents that hospitals may count for direct GME.

Section 1886(h) of the Act, as added by section 9202 of the Consolidated Omnibus Budget Reconciliation Act (COBRA) of 1985 (Pub. L. 99-272) and implemented in regulations at § 413.86(e), establishes a methodology for determining payments to hospitals for the costs of approved GME programs. Section 1886(h)(2) of the Act, as added by COBRA, sets forth a payment methodology for the determination of a hospital-specific, base-period per resident amount (PRA) that is calculated by dividing a hospital's allowable costs of GME for a base period by its number of residents in the base period. The base period is, for most hospitals, the hospital's cost reporting period beginning in FY 1984 (that is, the period of October 1, 1983 through September 30, 1984). The PRA is multiplied by the weighted number of full-time equivalent (FTE) residents working in all areas of the hospital complex (or nonhospital sites, when applicable), and the hospital's Medicare share of total inpatient days to determine Medicare's direct GME payments.

Existing regulations at § 413.86(e)(4) specify the methodology for calculating each hospital's weighted average PRA and the steps for determining whether a hospital's PRA will be revised.

2. Prohibition Against Counting Residents Where Other Entities First Incur the Training Costs

a. General Background on Methodology for Determining FTE Resident Count

As we explain earlier in this preamble, Medicare makes both direct and indirect GME payments to hospitals for the training of residents. Direct GME payments are reimbursed in accordance with section 1886(h) of the Act, based generally on hospital-specific PRAs, the number of FTE residents a hospital trains, and the hospital's Medicare patient share. The indirect costs of GME are reimbursed in accordance with section 1886(d)(5)(B) of the Act, based generally on the ratio of the hospital's FTE residents to the number of hospital beds. It is well-established that the calculation of both direct GME and IME payments is affected by the number of FTE residents that a hospital is allowed to count; generally, the greater the number of FTE residents a hospital counts, the greater the amount of Medicare direct GME and IME payments the hospital will receive. In an attempt to end the implicit incentive for hospitals to increase the number of FTE residents, Congress instituted a cap on the number of allopathic and osteopathic residents a hospital is allowed to count for direct GME and IME purposes under the provisions of section 1886(h)(4)(F) (direct GME) and section 1886(d)(5)(B)(v) (IME) of the Act. Dental and podiatric residents were not included in this statutorily mandated cap.

With respect to reimbursement of direct GME costs, since July 1, 1987, hospitals have been allowed to count the time residents spend training in sites that are not part of the hospital (referred to as “nonprovider” or “nonhospital sites”) under certain conditions. Section 1886(h)(4)(E) of the Act requires that the Secretary's rules concerning computation of FTE residents for purposes of separate reimbursement of direct GME costs “provide that only time spent in activities relating to patient care shall be counted and that all the time so spent by a resident under an approved medical residency training program shall be counted towards the determination of full-time equivalency, without regard to the setting in which the activities are performed, if the hospital incurs all, or substantially all, of the costs for the training program in that setting.” (Section 1886(h)(4)(E) of the Act, as added by section of 9314 of the Omnibus Budget Reconciliation Act of 1986, Pub. L. 99-509.)

Regulations on time spent by residents training in nonhospital sites for purposes of direct GME payment were first implemented in the September 29, 1989 final rule (54 FR 40286). We stated in that rule (under § 413.86(f)(3)) that a hospital may count the time residents spend in nonprovider settings for purposes of direct GME payment if the residents spend their time in patient care activities and there is a written agreement between the hospital and the nonprovider entity stating that the hospital will incur all or substantially all of the costs of the program. The regulations at that time defined “all or substantially all” of the costs to include the residents' compensation for the time spent at the nonprovider setting.

Prior to October 1, 1997, for IME payment purposes, hospitals could only count the time residents spend training in areas subject to the IPPS and outpatient areas of the hospital. Section 4621(b)(2) of the Balanced Budget Act of 1997 (Pub. L. 105-33) revised section 1886(d)(5)(B) of the Act to allow providers to count time residents spend training in nonprovider sites for IME purposes, effective for discharges occurring on or after October 1, 1997. Specifically, section 1886(d)(5)(B)(iv) of the Act was amended to provide that “all the time spent by an intern or resident in patient care activities under an approved medical residency program at an entity in a non-hospital setting shall be counted towards the determination of full-time equivalency if the hospital incurs all, or substantially all, of the costs for the training program in that setting.”

In the regulations at §§ 412.105(f)(1)(ii)(C) and 413.86(f)(4) (as issued in the July 31, 1998 Federal Register), we specify the requirements a hospital must meet in order to include a resident training in a nonhospital site in its FTE count for Medicare reimbursement for portions of cost reporting periods occurring on or after January 1, 1999 for both direct GME and for IME payments. The regulations at § 413.86(b) redefine “all or substantially all of the costs for the training program in the nonhospital setting” as the residents' salaries and fringe benefits (including travel and lodging where applicable), and the portion of the cost of teaching physicians' salaries and fringe benefits attributable to direct GME. A written agreement between the hospital and the nonhospital site is required before the hospital may begin to count residents training at the nonhospital site; the agreement must provide that the hospital will incur the costs of the resident's salary and fringe benefits while the resident is training in the nonhospital site. The hospital must also provide reasonable compensation to the nonhospital site for supervisory teaching activities, and the written agreement must specify that compensation amount.

b. Inappropriate Counting of FTE Residents

As we stated above, dental residents, along with podiatric residents, are excepted from the statutory cap on the count of FTE residents for both direct GME and IME payment purposes. We have become aware of a practice pertaining to the counting of FTE residents at a nonhospital site, particularly dental residents, that we see as inappropriate under Medicare policy. Most often, the situation involves dental schools that, for a number of years, have been training dental residents in programs at the dental schools of universities affiliated with teaching hospitals, and the schools have been directly incurring the costs of the dental residents training at the dental schools (for example, the teaching faculty costs, the resident salary costs, the office space costs, and any overhead expenses of the programs). We also understand that there are dental clinics at these dental schools that treat patients (that is, are involved in “patient care activities”).

As a result of the provisions that Congress added to allow hospitals to count FTE residents and receive IME payment, as well as direct GME payment, if the hospital incurs “all or substantially all” the costs of training residents in nonhospital settings, a significant number of dental schools are shifting the resident training costs of the dental programs from the schools to the hospital, and thus to the Medicare program, when the hospitals count the FTE dental residents training in these dental schools (that is, “nonhospital sites”) under the regulations at § 413.86(f)(4). Furthermore, in the case of training dentists at dental school clinics, as a result of this cost-shifting and because dental residents are excepted from the cap, hospitals are receiving significant amounts of Medicare direct GME and IME payments when they have incurred relatively small costs of the residents training in a dental school.

The following actual situations are illustrative of the inappropriate application of Medicare direct GME and IME policy that we have found:

• An academic medical center hospital associated with a university has been training allopathic residents for at least 20 years. Prior to 1999, the university's affiliated dental school had always incurred the costs of dental residency programs at the dental school. Beginning with the hospital's cost report for its fiscal year ending in 1999, for the first time ever, the hospital has requested direct GME and IME payment for an additional 67 FTE residents because the hospital claims it has begun to incur “all or substantially all” of the costs of the dental residents training in the university's affiliated dental school, in accordance with the regulations at § 413.86(f)(4).
• A university dental school in one State has been incurring the costs of dental residency programs at its dental school for several years. Beginning in FY 1999, a teaching hospital in a neighboring State decided to begin incurring all or substantially all of the costs of the dental residents training in the dental clinics in the program (which is located in a different State from the hospital) in order to receive Medicare direct GME and IME payment for an additional 60 FTE residents.
• In another situation, a teaching hospital on the East Coast of the United States has requested direct GME and IME payment for an additional 60 FTE dental residents, some of whom are training in dental programs at nonhospital sites located in Hawaii, New Mexico, and the Netherlands, because it has begun to incur “all or substantially all” of the costs of dental residents training in those remote “nonhospital sites”. Prior to 1999, the costs for these dental programs were funded by nonhospital sources.

We note that such inappropriate cost-shifting practices are by no means limited to the dental school context. Indeed, we understand that there are some hospitals with resident counts below their direct GME and IME FTE resident caps that have recently (as of October 1, 1997, when it became possible to receive significant IME payments under the amendment made by Pub. L. 105-33) started to incur “all or substantially all” of the costs of residents who had been training at sites outside of the hospital without any financial assistance from the hospital, in order for the hospital to count those FTE residents and receive Medicare direct GME and IME payments for the additional residents. The actual costs of the programs that are being shifted from nonhospital entities to hospitals are relatively small, compared to the direct GME and IME payments that hospitals receive as a result of incurring “all or substantially all” of the training costs.

• In another example, an academic medical center hospital in one State asked Medicare to allow it to count an additional 10 FTEs for both direct GME and IME payment, beginning with its fiscal year ending 1999 cost report, because the hospital claims it is incurring all or substantially all of the costs of training osteopathic family practice residents in a walk-in clinic. The osteopathic family practice residency program had previously been sponsored by this clinic for several years and the residents do not participate in any training at the hospital.

c. Congressional Intent

Congress has delegated broad authority to the Secretary to implement a policy on the count of FTE residents for purposes of calculating direct GME and IME payments. For IME payment, section 1886(d)(5)(B) of the Act simply states that “the Secretary shall provide for an additional payment amount” which includes “the ratio of the hospital's full-time equivalent interns and residents to beds.” The methodology to compute the count of FTE residents for IME is not established in the statute. Similarly, for direct GME, section 1886(h)(4)(A) of the Act states that “the Secretary shall establish rules consistent with this paragraph for the computation of the number of full-time equivalent residents in an approved medical residency training program.”

Although not in the context of the general rules for counting FTE residents, Congress similarly acknowledged its intent to defer to the Secretary with respect to the rules for implementing “limits” or caps on the number of FTE residents hospitals may count for purposes of direct GME and IME payment. The conference agreement that accompanied Pub. L. 105-33, which established a cap on the number of allopathic and osteopathic residents a hospital may count, states—

“[T]he Conferees recognize that such limits raise complex issues, and provide for specific authority for the Secretary to promulgate regulations to address the implementation of this provision. The Conferees believe that rulemaking by the Secretary would allow careful but timely consideration of this matter, and that the record of the Secretary's rulemaking would be valuable when Congress revisits this provision.” (H.R. Conf. Rep. No. 105-217, 105th Cong., 1st Sess., 821 (1997).

The absence of statutory specificity on determining FTE counts in these situations and the declared Congressional delegations of authority to the Secretary on the subject are clear indications that Congress has given the Secretary broad discretion to promulgate reasonable regulations in order to implement the policy on the counting of residents for direct GME and IME payments.

When Congress enacted the nonhospital site provisions for both direct GME and IME, Congress intended to address application of the FTE count policy to situations where the training site had been the hospital. The intent was to create incentives for hospitals to move resident training from the hospital to nonhospital settings. We believe that Congress did not intend for hospitals to be able to add to their FTE counts residents that had historically trained outside the hospital in other settings. Training in those nonhospital settings had historically occurred without Congress offering any financial incentive to hospitals to move the training out of the hospital.

This Congressional intent is evident in the legislative history of both the direct GME and the IME provisions on nonhospital settings. First, legislative history associated with passage of the direct GME provision (as part of Pub. L. 99-509) indicates that Congress intended to broaden the scope of settings in which a hospital could train its residents and still receive separate direct GME cost reimbursement, and to provide incentives to hospitals for training residents in primary care programs. The Conference committee report indicates that “[s]ince it is difficult to find sufficient other sources of funding [than hospitals and Medicare] for the costs of such training, [that is, training in freestanding primary care settings such as family practice clinics or ambulatory surgery centers] assignments to these settings are discouraged. It is the Committee's view that training in these settings is desirable, because of the growing trend to treat more patients out of the inpatient hospital setting and because of the encouragement it gives to primary care.” (Emphasis added.) (H.R. Rep. No. 99-727, 99th Cong., 1st Sess., 70 (1986).)

Thus, from the start of the policy allowing payment for training in nonprovider sites, we believe Congress intended to create a monetary incentive for hospitals to rotate residents from the hospital to the nonhospital settings. We believe Congress did not intend for hospitals to be paid for residents who had previously been training at nonhospital sites without hospital funding.

Further, in the Conference committee report accompanying the provision of Pub. L. 105-33 on IME payment for training in nonhospital settings, Congress stated that “[t]he conference agreement includes new permission for hospitals to rotate residents through nonhospital settings, without reduction in indirect medical education funds.” (Emphasis added.) (H.R. Conf. Rep. No. 105-217, 105th Cong., 1st Sess., 817 (1997).)

We note that, prior to enactment of Pub. L. 105-33, if a hospital rotated a resident to train at a nonhospital site, the hospital could not count the time the resident spent at the nonhospital site for purposes of Medicare IME payments. As a result, the lack of IME payments acted as a disincentive and discouraged hospitals from rotating residents out of the hospital. Therefore, Congress authorized hospitals to count residents in nonhospital sites for IME purposes as a specific incentive to encourage hospitals to rotate their residents to nonhospital sites (and not to encourage hospitals to incur the costs of a program at a nonhospital site that had already been funded by other sources). This legislative intent becomes more apparent when the nature of the Medicare IME payment is considered. The Medicare IME payment is inherently a payment that reflects the increased operating costs of treating inpatients as a result of the hospital having a residency program. For example, as explained in the September 29, 1989 final rule (54 FR 40286), the indirect costs of medical education might include added costs resulting from an increased number of tests ordered by residents as compared to the number of tests normally ordered by more experienced physicians.

The IME payment is an adjustment that is made for each Medicare discharge from the areas subject to the IPPS in a teaching hospital. The authorization by Congress for IME payments relating to nonhospital services while residents are training at nonhospital sites would be absurd if not viewed as an incentive to transfer existing residency training from the hospital to the nonhospital setting. We do not believe Congress intended to permit such IME payments to be allowable to the hospital that is incurring “all or substantially all the costs” of residents training in nonhospital sites except in the situation where the hospital rotated residents from the hospital to the nonhospital settings. The illustrative situations described above in which nonhospital sites, such as dental schools, are shifting the costs of existing programs to the hospitals are not consistent with the intent of Congress to encourage hospitals to rotate residents from the hospital setting to nonhospital sites.

Thus, we believe Congress intended both cited provisions of the Act on counting residents in nonhospital sites for purposes of direct GME and IME payments to be limited to situations in which hospitals rotate residents from the hospital to the nonhospital settings, and not situations in which nonhospital sites transfer the costs of an existing program at a nonhospital site to the hospital.

d. Medicare Principles on Redistribution of Costs and Community Support

It is longstanding Medicare policy that if the community has undertaken to bear the costs of medical education, these costs are not to be assumed by the Medicare program. In addition, medical education costs that have been incurred by an educational institution may not be redistributed to the Medicare program. Indeed, these concepts, community support and redistribution of costs, have been a part of Medicare GME payment policy since the inception of the Medicare program. Both the House and Senate Committee reports accompanying Pub. L. 89-97 (the authorizing Medicare statute) indicate that Congress intended Medicare to share in the costs of medical education only in situations in which the community has not stepped in to incur them:

“Many hospitals engage in substantial education activities, including the training of medical students, internship and residency programs, the training of nurses and the training of various paramedical personnel. Educational activities enhance the quality of care in an institution and it is intended, until the community undertakes to bear such education costs in some other away, that a part of the net cost of such activities * * * should be considered as an element in the cost of patient care, to be borne to an appropriate extent by the hospital insurance program. (Emphasis added.) (S. Rep. No. 404, 89th Cong., 1st Sess., 36 (1965); H.R. Rep. No. 213, 89th Cong., 1st Sess., 32 (1965).)

The principle behind the congressional committee report language for Pub. L. 89-97 that Medicare would share in the costs of educational activities until communities bore them in some other way has guided Medicare policy on educational activities from the inception of the Medicare program. The principles of community support and redistribution of costs associated with payment for GME have been continually reiterated in various regulations, manual provisions, and implementing instructions to fiscal intermediaries. As recently as the final rule published in the Federal Register on January 12, 2001, we stated:

“We note that the proposed revisions in the proposed rule inadvertently did not include community support as the basis for an offset from the allowed cost of a GME or nursing and allied health program. In this final rule, we restate our longstanding policy that Medicare will share in the costs of educational activities of providers where communities have not assumed responsibility for financing these programs. Medicare's policy is to offset from otherwise allowable education costs, community funding for these activities.” (66 FR 3368)

We note the instructions that CMS (then HCFA) gave to its Regional Offices in the 1990 audit instructions for purposes of calculating the direct GME base period PRA specifically addressed redistribution of costs and community support in the GME context:

“Where costs for services related to medical education activities have historically been borne by the university, it is assumed the community has undertaken to support these activities, and subsequent allocation of these costs to a hospital constitutes a redistribution of costs from an educational institution to a patient care institution. In such a situation, these costs are not allowable under the Medicare program. (See 42 CFR 413.85(c) and HCFA Pub. 15-1, § 406). For example, if in the past the hospital did not identify and claim costs attributable to the time teaching physicians spent supervising I&Rs [interns and residents] working at the hospital, it is assumed that these costs were borne by the university. Therefore, the hospital may not claim these costs in subsequent cost reports.” (Instructions for Implementing Program Payments for Graduate Medical Education to ARAs for Medicare, Director of Office of Financial Operations of the Health Care Financing Administration, BPO-F12, February 12, 1990.)

Furthermore, the regulation at § 413.85(c) that was originally issued in the Federal Register on September 30, 1986 (51 FR 34793) (which was further refined, but conceptually left unchanged, as of March 12, 2001) addressed the Congressional intent not to increase program costs, as well. That paragraph (c) stated:

Educational Activities. Many providers engage in education activities including training programs for nurses, medical students, interns and residents, and various paramedical specialties.* * * Although the intent of the program is to share in the support of educational activities customarily or traditionally carried on by providers in conjunction with operations, it is not intended that this program should participate in increased costs resulting from redistribution of costs from educational institutions or units to patient care institutions or units.

The Secretary of Health and Human Services interpreted this provision to deny reimbursement of educational costs that were borne in prior years by a hospital's affiliated medical school. The U.S. Supreme Court affirmed the Secretary's interpretation of the redistribution of costs regulation in Thomas Jefferson University v. Shalala (“Thomas Jefferson”), 512 U.S. 504 (1994). The Court found of § 413.85(c) that:

“The regulation provides, in unambiguous terms, that the ‘costs’ of these educational activities will not be reimbursed when they are the result of a ‘redistribution,’ or shift, of costs of an ‘educational’ facility to a ‘patient care’ facility.” (Emphasis added.) (Thomas Jefferson, 512 U.S. at 514). Thus, the Supreme Court in Thomas Jefferson held that it is well within the Secretary's discretion to interpret the language at § 413.85(c), which was specifically derived from the legislative history of the original enacting Medicare legislation quoted above, to impose a substantive limitation on medical education payment.

The Supreme Court's opinion in Thomas Jefferson lends substantial support and credibility to CMS” longstanding policy on community support and redistribution of costs in the GME context.

e. Application of Redistribution of Costs and Community Support Principles.

As we have described above, we have discovered an inappropriate application of Medicare direct GME and IME payment policies relating to the counting of FTE residents in nonhospital settings. As stated previously, we believe that: (1) Congress has given the Secretary broad discretion to implement policy on FTE resident counts; (2) Congress intended that the nonhospital site policy for both direct GME and IME would encourage hospitals to move resident training from the hospital to nonhospital settings, not to enable nonhospital sites to shift the costs of already established residency programs in the nonhospital site to the hospital; and (3) since the inception of the Medicare program, CMS” policy has been consistent with the intent of Congress that Medicare would only share in the costs of medical education until the community assumes the costs. The Supreme Court has specifically found that CMS” implementation of the redistribution of costs and community support principles is “reasonable.” (Thomas Jefferson, 512 U.S. at 514.)

Accordingly, in the May 19, 2003 proposed rule, we proposed that residents training at nonhospital sites may be counted in a hospital's FTE resident count only where the principles of redistribution of costs and community support are not violated. We proposed this policy to address the inappropriate practice of nonhospital sites shifting costs to hospitals solely to allow the hospitals to count residents training in the nonhospital sites. However, we believe the concepts of redistribution of costs and community support are equally relevant to the counting of FTEs residents by a hospital in general.

We note again that the Medicare program has a long tradition of applying redistribution of costs and community support principles to medical education payments. As we have stated above, both the House and Senate Committee reports accompanying Pub. L. 89-97 (the 1965 authorizing Medicare statute) indicate that Congress intended Medicare to share in the costs of medical education only where the community has not stepped in to incur them.

We believe it is appropriate to employ the principles of redistribution of costs and community support to specifically address the inappropriate scenarios described above whereby hospitals attempt to inflate their FTE resident counts by assuming payment of training costs for residents in nonhospital sites that were previously funded by a nonhospital entity. Therefore, we proposed to specify the application of the redistribution of costs and community support principles by adopting the definitions (with some modification to reflect the methodology for counting FTE residents applicable to GME) of “community support” and “redistribution of costs” at § 413.85(c), which relate to nursing and health education program costs, for use at § 413.86(b), which relates to GME. In addition, we proposed a general rule at proposed § 413.86(i) on the application of community support and redistribution of costs principles to the counting of FTE residents for GME. We proposed to (1) make the provisions under § 413.86(f) relating to determining the number of FTE residents subject to the provisions of the proposed § 413.86(i); (2) add a proposed § 413.86(f)(4) in order to clarify that the principles of redistribution of costs and community support are applicable to the counting of FTE residents, including when the residents are training in nonhospital settings; and (3) making the provisions of the proposed § 413.86(i) specifically applicable to determining the number of FTE residents under § 413.86(g)(4) through (6) and (g)(12).

The general rule at proposed § 413.86(i) contained two provisions. Proposed § 413.86(i)(1) stated the principles of community support and redistribution of costs: In relation to community support, we proposed that if the community has undertaken to bear the costs of medical education through community support, the training costs of residents that are paid through community support are not considered GME costs to the hospital for purposes of Medicare payment. In relation to redistribution of costs, we are proposing that the costs of training residents that constitute a redistribution of costs from an educational institution to the hospital are not considered GME costs to the hospital for purposes of Medicare payment.

In applying the redistribution of costs and community support principles, we proposed under § 413.86(i)(2) to state that a hospital must continuously incur direct GME costs of residents training in a particular program at a training site since the date the residents first began training in that site in order for the hospital to count the FTE residents in accordance with the provisions of paragraphs (f) and (g)(4) through (g)(6), and (g)(12) of § 413.86.

We note that our reasons for specifically referencing the applicability of the principles of community support and redistribution of costs at § 413.86(f)(4), the paragraph concerning counting residents training in nonhospital settings for direct GME purposes, are twofold. First, although we already proposed to make the proposed § 413.86(i) applicable to § 413.86(f), which would make the principles applicable to each paragraph under § 413.86(f), in consideration of the inappropriate applications we have identified of the GME FTE-counting policy with respect to counting residents in nonhospital sites, we believe it is appropriate to also specifically address the applicability of the redistribution of costs and community support principles to § 413.86(f)(4). In addition, we note that the proposed reference at § 413.86(f)(4) has implications for IME payment as well, as explained below.

Under existing § 412.105(f)(1)(ii)(C), the rule for the counting of FTE residents training in nonhospital settings for IME payment, there is a specific reference indicating that the criteria set forth in § 413.86(f)(4) must be met in order for a hospital to count the FTE residents training in nonhospital settings for purposes of IME payments. Thus, if under proposed § 413.86(f)(4)(iv) (the paragraph making redistribution of costs and community support principles applicable) a hospital is not permitted to count the FTE residents training in a nonhospital site because of redistribution of costs or community support, the hospital would not be permitted to count the FTE residents for purposes of IME payment as well, because the IME regulation at § 412.105(f)(1)(ii)(C) requires the criteria under § 413.86(f)(4) to be met.

As we have stated above, payment for IME is based on the concept that, as a direct result of the hospital's resident training program, the costs the hospital incurs for patient care are increased. When Congress included section 1886(d)(5)(B)(iv) of the Act as part of Pub. L. 105-33, the statute expanded the circumstances under which IME payments to a hospital could be made by allowing the hospital to count the number of residents training outside the hospital setting under certain conditions. Even though it is clear that those residents training outside the hospital cannot have any impact on patient care costs to the hospital, Congress nevertheless allowed the hospital to receive IME payments when the hospital counts FTE residents training in a nonhospital setting in accordance with section 1886(d)(5)(B)(iv) of the Act, where those residents would otherwise have trained in the hospital setting. As we have stated, Congress created an incentive (or removed a disincentive) with the provisions of Pub. L. 105-33 for hospitals to rotate residents to nonhospital settings by allowing hospitals to continue to receive IME payment as if the residents continued to train in the hospital setting. If there is a redistribution of costs or community support, we believe IME payment to the hospital would be contrary to Congressional intent to encourage the hospital to rotate residents from the hospital to the nonhospital site.

In addition, when Congress included section 1886(d)(5)(B)(iv) of the Act as part of Pub. L. 105-33, the statutory authority for IME payment was premised on the hospital incurring the direct GME costs of the residents: “all the time spent by an intern or resident in patient care activities under an approved medical residency program at an entity in a nonhospital setting shall be counted towards the determination of full-time equivalency if the hospital incurs all, or substantially all, of the costs for the training program in that setting.” (Emphasis added.) (Section 4621(b)(2) of Pub. L. 105-33; section 1886(d)(5)(B)(iv) of the Act.) We believe Congress intended the hospital to incur direct GME costs of the program in the nonhospital site in order to count the FTE residents training in nonhospital settings for purposes of IME payment. Thus, in the situation where a hospital incurred direct GME costs but there was redistribution of costs or community support, a disallowance of direct GME payments as well as a disallowance of IME payments is appropriate.

Although we are stating generally that the principles of community support and redistribution of cost have applied since the inception of Medicare to graduate medical education payment, as we have stated above, we have identified relatively recent inappropriate application of the nonhospital site policy for counting FTE residents. Therefore, we believed it was appropriate to propose to identify January 1, 1999 as the date our fiscal intermediaries should use to determine whether a hospital or another entity has been incurring the costs of training in a particular program at a training setting for purposes of determining whether there has been a redistribution of costs or community support. We proposed that January 1, 1999 be used as the date the fiscal intermediaries should use for determinations, since it may be difficult for our fiscal intermediaries to obtain from hospitals contemporaneous documentation that the hospitals have appropriately been incurring the direct GME costs in earlier fiscal years. We believe the January 1, 1999 date should simplify confirmation by our fiscal intermediaries and hospitals of whether the hospital or another entity had been incurring the costs of the program in particular training settings and whether redistribution of costs or community support had occurred. We have chosen the January 1, 1999 date because of administrative convenience and feasibility, so that necessary data are both valid and available, and in recognition of the fact that our fiscal intermediaries must prioritize their limited audit resources. While we are not requiring our fiscal intermediaries to determine whether a hospital had been incurring the training costs of a program prior to the January 1, 1999 date, if the fiscal intermediaries determine that there is a redistribution of costs or community support exists with respect to certain residents prior to January 1, 1999, a disallowance of direct GME and IME payments with respect to those FTE residents would certainly be required.

Since calculation of a hospital's FTE resident count is dependent upon whether the hospital incurred the training costs, we proposed to require each teaching hospital and its fiscal intermediary to determine which entity had been incurring the training costs at least since January 1, 1999. For example, if a nonhospital entity, such as a school of medicine or dentistry, had incurred the costs of training the residents anytime on or after January 1, 1999, and a hospital subsequently begins to incur direct GME costs of training those FTE residents, the hospital would not qualify to count those FTE residents for purposes of direct GME and IME payments.

We note that the proposal stated that a hospital must have been continuously incurring the costs of the training since the date the residents first began training in that program. Accordingly, if a hospital had at one time incurred the costs of training residents in a particular program, whether at the hospital or in a nonhospital setting, but a nonhospital institution later assumed the costs of training in that setting, even if the hospital assumed payment for the training costs again, the hospital could not then count those residents for purposes of direct GME and IME payments.

We note that if a hospital incurs the direct GME costs, whether training takes place inside the hospital or in a nonhospital setting, in a new residency program, the hospital may be eligible to count the FTE residents as specified by the regulations under § 413.86(g)(6).

Consistent with the policy on redistribution of costs and community support discussed above, if a hospital incurs the direct GME costs of additional FTE residents training in an existing program in a hospital setting where the costs of the existing program had been incurred by a nonhospital entity and the hospital has continuously funded the additional residents in the existing program in the hospital setting since the date the residents first began training there, the redistribution of costs or community support principles would not prohibit the hospital from counting the additional FTE residents for purposes of direct GME and IME payments.

We note that, under existing policy, to count residents in a nonhospital setting, a hospital is required to incur for “all or substantially all of the costs of the program” in that setting. In other words, a hospital is required to assume financial responsibility for the full complement of residents training in a nonhospital site in a particular program in order to count any FTE residents training there for purposes of IME payment. A hospital cannot count any FTE residents if it incurs “all or substantially all of the costs” for only a portion of the FTE residents in that program training setting. This policy is derived from the language of the IME and direct GME provisions of the statute on counting residents in nonhospital settings; both sections 1886(d)(5)(B)(iv) and 1886(h)(4)(E) of the Act state that the hospital must incur “all, or substantially all, of the costs for the training program in that setting.” (Emphasis added.) In contrast, as explained earlier, it is permissible under the proposed policy on the application of the redistribution of costs and community support principles for the hospital to count FTE residents where the hospital incurs direct GME costs of FTE residents that are added to an existing program, even though the hospital may not count the existing FTE residents due to the application of the redistribution of costs or community support rules. In the nonhospital setting, as a result of the interaction of these two separate FTE counting requirements—(1) that the hospital must not violate the redistribution of costs and the community support principles in order to count the resident FTEs in the nonhospital settings, and (2) that the hospital must incur “all or substantially all” of the costs for the training program in that setting—a hospital would be prohibited from counting FTE residents added to an existing program at a nonhospital site unless the hospital incurs all or substantially all of the costs of training all of the residents in that program at that setting. That is, even if the hospital incurs all or substantially all of the costs for all of the training program at the nonhospital site, the hospital would only be able to count the additional FTE residents who were not excluded by application of the redistribution of costs or community support principles.

For example, training in a general dentistry program with 10 FTE residents has taken place at a school of dentistry for 20 years. The school of dentistry has been incurring the training costs of the general dentistry residents since the inception of the program. Beginning in 2003, the school of dentistry has decided to add an additional 5 FTE residents to the program, and Hospital A decides to incur “all or substantially all” the costs of those 5 additional FTE residents only. Applying the policy concerning redistribution of costs and community support in combination with the policy on incurring all or substantially all of the costs, the hospital could not count the additional 5 FTE residents in the dental school since it is not paying for all or substantially all of the costs of the program. Even if the hospital were to incur all or substantially all of the costs for the training program for all 15 FTE residents, the hospital could not count the 10 FTEs that were part of the existing general dentistry program because of the redistribution of costs and community support principles; it would be a redistribution of costs for the hospital to begin to incur direct GME costs of the 10 FTE residents when the dental school had previously been incurring those costs.

We note that such a result does not occur when a new program is established in the nonhospital site. If, from the outset of the program, the hospital incurs direct GME costs and also incurs “all or substantially all” of the costs for the training program for all the new residents training at the site, there would be no redistribution of costs or community support, and the hospital could count all of those residents in the new program in its FTE count (subject, of course, to the hospital's 1996 FTE resident cap).

We also note that the interaction of the two provisions discussed above—redistribution of costs and community support, and “all or substantially all”—does not occur when counting FTE residents training inside the hospital, since a hospital is not required to incur “all or substantially all” of the costs for the training program inside the hospital.

Furthermore, if one hospital had incurred the direct GME costs of training residents in a particular program in a nonhospital site from one point in time, for example, 1995 through 1999, and then another hospital consecutively incurs the costs from 2000 and thereafter, the second hospital may be eligible to receive direct GME and IME payments for training the FTE residents from the point in time where the second hospital incurred the direct GME costs, and the redistribution and community support exclusions would not apply. The second hospital may be eligible to receive Medicare direct GME and IME payments because the costs were incurred previously by a hospital, and not either the community or the university. Therefore, there was neither community support nor redistribution of costs.

The following are some examples to clarify how the proposed policies would be implemented:

Example 1

Since 1995, 10 FTE residents in an internal medicine program have been training in the Community Clinic. In accordance with the current provisions of § 413.86(f), Hospital A has incurred all or substantially all of the costs of training the 10 FTE residents since 1995. Assuming the current provisions of the regulations at §§ 412.105(f)(1)(ii)(C) and 413.86(f)(3) and (f)(4) are met, Hospital A may continue to receive IME and direct GME payments for 10 FTE residents because Hospital A had incurred direct GME costs continuously (as evidenced by contemporaneous documentation since January 1, 1999), as specified in our proposed regulation.

Beginning July 1, 2004, in addition to continuing to incur all or substantially all of the costs of the first 10 FTE internal medicine residents training in the nonhospital site, Hospital A also incurs all or substantially all of the costs of training an additional 3 FTE internal medicine residents at that site. Accordingly, beginning July 1, 2004, Hospital A may count all 13 FTE residents training in the Community Clinic for purposes of direct GME and IME payments, assuming Hospital A does not exceed its FTE cap for IME and direct GME.

Example 2

Since 1995, 2.25 dental FTE residents in a dental school program were training in a dental clinic at the dental school. While the 2.25 FTEs were training at the clinic, the dental school paid for all of the costs of the dental program. Prior to July 1, 2000, Hospital A signed a written agreement with the clinic to incur all or substantially all of the costs of training the 2.25 FTE residents, from July 1, 2000 and onward. Thus, beginning with July 1, 2000, the dental school no longer incurred the costs of the program at this nonhospital site. In this scenario (even if Hospital A inappropriately received direct GME and IME payments for the 2.25 FTEs since July 1, 2000), Hospital A may not receive direct GME or IME payment for the 2.25 FTE residents training in the clinic because there would have been a redistribution of costs associated with training these 2.25 FTE residents from the dental school to the hospital.

Example 3

Since 1995, 2.25 FTE residents in a family practice program were training in a physicians' group practice. While the 2.25 FTEs were training at the physicians' practice, a school of medicine paid for the costs of the family practice residency program. Prior to July 1, 2000, Hospital A signed a written agreement with the physicians' practice to send 1 additional family practice FTE resident to the physicians' practice and to incur all or substantially all of the costs of training the original 2.25 FTE residents and the 1 additional FTE, from July 1, 2000 and onward. Thus, beginning with July 1, 2000, the school of medicine no longer incurred the costs of the program at this nonhospital site. Hospital A may not count the 2.25 FTE residents that had been training since 1995 in that physicians' practice for purposes of direct GME and IME payments because the training costs were shifted from the school of medicine to the hospital. However, Hospital A may count the 1 FTE resident the hospital began to rotate for training in the physicians' practice because there was no cost-shifting for that resident and Hospital A incurred “all or substantially all” of the costs of the entire family practice program in the physicians' office setting.

Example 4

Residents in a surgery program have been rotating from a hospital to two nonhospital clinics, Clinic A and Clinic B, since 1996. The training of the surgery residents in Clinic A has been supported by a nonhospital institution since 1996, while the hospital has incurred all or substantially all of the costs of the surgery residents in Clinic B since 1996. The hospital cannot count the surgery FTE residents training in Clinic A, even if it begins to pay for all of the costs of the program at that site, since a nonhospital institution had supported the training in Clinic A since 1996 (in other words, the redistribution of costs and community support principles would prohibit the hospital from counting these FTE residents). However, if the hospital continues to incur all or substantially all of the costs of the surgery residents in Clinic B, the hospital may count the FTE residents training in Clinic B for purposes of direct GME and IME payments because there would be no cost-shifting to the hospital for these residents and the hospital would incur all or substantially all of the costs for the training program in that setting.

We received a large number of comments from the public on this proposal. Following is a summary of these comments and our responses:

Comment: Some commenters supported our proposed application of redistribution of cost and community support to direct GME. One commenter stated: “We believe that the proposed changes * * * will result in more accurate and consistent reimbursement to providers. The changes provide more definitive guidance to providers and to intermediaries in applying the regulations. In addition, the changes will more closely match Medicare reimbursement with actual IPPS-type services. This is especially true in the case of dental residents, who typically spend little or no time caring for patients receiving IPPS type services.”

Response: We agree with the commenters' assertions and appreciate the commenters' support of our proposals on redistribution of costs and community support.

Comment: Many commenters disagreed with our proposed application of redistribution of cost and community support to direct GME. In general, they believed they did not receive proper notice of the application of the principles. One commenter stated: “[t]he proposed change to the rules midstream, and only with respect to subsequent payment years, distorts the balance on which the established payment formula depends.” Other commenters believed that, in the past, CMS has never suggested that incurring the costs of offsite training in the then-current year would be a condition to hospitals' claiming those costs in future years. The commenters contended that nowhere in the regulations promulgated has CMS stated that, in order to receive GME and IME payments, a hospital must meet an additional requirement of incurring the training costs since the inception of the training program. The commenters believed it is inequitable to impose such a “retroactive requirement.”

The commenters stated that many hospitals that were contemplating whether to initiate a training program in a nonhospital setting, notified CMS in advance of establishing such a program, and requested CMS's approval. One commenter stated that, in numerous cases, “including some of the cases discussed in the regulatory preamble, CMS issued a written approval of the proposed training program. In such approval letters, CMS never mentioned the redistribution of costs and community support principles.”

Finally, another commenter stated that there is nothing in the direct GME and IME statutes that supports CMS' decision to apply redistribution of costs and community support principles.

Response: The principles of redistribution of cost and community support associated with Medicare's payments for GME have been in existence for over 35 years, that is, since the inception of the Medicare program in 1965. The principles have been continually reiterated in various regulations, manual provisions, and implementing instructions to fiscal intermediaries. We do not believe we have given the public any reason to conclude that the principles would not continue to be applicable. Several examples of our views on the principles of redistribution of cost and community support were mentioned in the proposed rule. These included:

Both the House and Senate Committee reports accompanying Pub. L. 89-97 (the authorizing Medicare statute) indicate that Congress intended Medicare to share in the costs of medical education only in situations in which the community has not stepped in to incur them:

“Many hospitals engage in substantial education activities, including the training of medical students, internship and residency programs, the training of nurses and the training of various paramedical personnel. Educational activities enhance the quality of care in an institution and it is intended, until the community undertakes to bear such education costs in some other away, that a part of the net cost of such activities * * * should be considered as an element in the cost of patient care, to be borne to an appropriate extent by the hospital insurance program.” (Emphasis added.) (S. Rept. No. 404, 89th Cong., 1st Sess., 36 (1965); H.R. Rept. No. 213, 89th Cong., 1st Sess., 32 (1965).)

The principle behind the congressional committee report language for Pub. L. 89-97 that Medicare would share in the costs of educational activities until communities bore them in some other way has guided Medicare policy on educational activities from the inception of the Medicare program.

The regulations that evolved from the authorizing legislation, first published on November 22, 1966 (31 FR 14814), as well as Chapter 4 of the Provider Reimbursement Manual in 1971, echoed the congressional committee report language from 1965 that Medicare would share in the costs of educational activities until communities bore them in some other way.

As recently as the final rule published in the Federal Register on January 12, 2001, we stated:

“We note that the proposed revisions in the proposed rule inadvertently did not include community support as the basis for an offset from the allowed cost of a GME or nursing and allied health program. In this final rule, we restate our longstanding policy that Medicare will share in the costs of educational activities of providers where communities have not assumed responsibility for financing these programs. Medicare's policy is to offset from otherwise allowable education costs, community funding for these activities.” (66 FR 3368)

Although the above language was written in the context of a regulation that clarified Medicare policy for provider (hospital) operated nursing and allied health education programs, we note that GME and nursing and allied health education programs were historically paid under the same regulations (the latest of which was codified at § 413.85) and the same cost principles. The quoted language is indicative of this relationship and the Agency's mindset that, while direct GME may have changed in the method of payment to a prospective payment, some principles, such as redistribution of cost and community support, continue to apply as they do with nursing and allied health education at § 413.85(c). Further evidence of continued application is at existing § 413.85(c) in the definition of “redistribution of cost”: “* * * costs for a school of nursing or allied health education or a medical school that were incurred by an educational institution and were not allowable to the provider [hospital] in its prospective payment or a rate-of-increase limit base year cost report, or graduate medical education per resident amount calculated under § 413.86, are not allowable costs in subsequent fiscal years.” (Emphasis added.) Therefore, even codified in regulations now is a policy that applies the principle of redistribution of cost to direct GME payments in subsequent years.

Furthermore, § 413.85(c), which was a codification of longstanding Medicare policy, was originally issued in the Federal Register on September 30, 1986 (51 FR 34793) and was further refined, but conceptually left unchanged, as of March 12, 2001 (see 66 FR 3358). Section 413.85(c) addressed the Congressional intent not to increase program costs resulting from redistribution of costs, as well. That paragraph (c) stated:

“Educational Activities. Many providers engage in education activities including training programs for nurses, medical students, interns and residents, and various paramedical specialties. * * * Although the intent of the program is to share in the support of educational activities customarily or traditionally carried on by providers in conjunction with operations, it is not intended that this program should participate in increased costs resulting from redistribution of costs from educational institutions or units to patient care institutions or units.”

We note that the guidance that CMS (then HCFA) gave to its Regional Offices in the 1990 audit instructions for purposes of calculating the direct GME base period PRA specifically addressed redistribution of costs and community support in the GME context:

“Where costs for services related to medical education activities have historically been borne by the university, it is assumed the community has undertaken to support these activities, and subsequent allocation of these costs to a hospital constitutes a redistribution of costs from an educational institution to a patient care institution. In such a situation, these costs are not allowable under the Medicare program. (See 42 CFR 413.85(c) and HCFA Pub. 15-1, section 406). For example, if in the past the hospital did not identify and claim costs attributable to the time teaching physicians spent supervising I&Rs [interns and residents] working at the hospital, it is assumed that these costs were borne by the university. Therefore, the hospital may not claim these costs in subsequent cost reports.” (Instructions for Implementing Program Payments for Graduate Medical Education to ARAs for Medicare, Director of Office of Financial Operations of the Health Care Financing Administration, BPO-F12, February 12, 1990.)

We believe we have continually put the public on notice that the Medicare program has applied and continues to apply the principles of redistribution of costs and community support to payments for education costs, including direct GME payments to hospitals. Therefore, we do not believe that we have proposed changes to the rules “in midstream” as one commenter suggested. Nor do we believe, as the commenters suggested, that we have proposed a “retroactive requirement.” We have never disavowed the principles of redistribution of cost and community support. Rather, we have continually promulgated rules and program guidance on the application of the principles since the inception of the Medicare program.

We again point to the Supreme Court case, Thomas Jefferson, to demonstrate CMS' longstanding policy on community support and redistribution of costs in the GME context. In Thomas Jefferson, the Secretary of Health and Human Services interpreted the regulation at § 413.85(c) to deny reimbursement of educational costs that were borne in prior years by a hospital's affiliated medical school for purposes of calculating the direct GME base year allowable cost for the PRA. The U.S. Supreme Court affirmed the Secretary's interpretation of the redistribution of costs regulation. The Court found that:

“The regulation [at § 413.85(c)] provides, in unambiguous terms, that the ‘costs’ of these educational activities will not be reimbursed when they are the result of a ‘redistribution,’ or shift, of costs of an ‘educational’ facility to a ‘patient care’ facility.” (Emphasis added.) (Thomas Jefferson, 512 U.S. at 514).

In addition, in response to the argument by the provider that CMS (then HCFA) had been silent in internal operating instructions in a 1978 operating memorandum on the policies of redistribution and community support, as well as in another exchange of memoranda in 1982 and other agency documentation, the Court stated that the omission in these documents of discussion of redistribution and community support is not indicative of a contrary policy on GME reimbursement: “* * * the mere failure to address [the redistribution principle in an intermediary letter] hardly establishes an inconsistent policy on the part of the Secretary.” Thomas Jefferson, 512 U.S. at 516.

Thus, the Supreme Court in Thomas Jefferson held that it is well within the Secretary's discretion to interpret the language at § 413.85(c), which was specifically derived from the legislative history of the original legislation that enacted Medicare, to impose a substantive limitation on medical education payment, even in the arguably novel context of calculating a hospital's GME costs for purposes of the base year PRA.

To address the commenters' point that CMS “never mentioned the redistribution of costs and community support principles” in CMS “approval letters” to hospitals that requested “approval” from CMS in advance of establishing a relationship with a nonhospital site in order to count the residents training in that setting, we note that when the letters were written to CMS in fiscal year end 1999-2002, it was not clear at all from the incoming correspondence that hospitals were not, in fact, rotating the hospital-based residents to the nonhospital setting in accordance with statutory intent. In other words, it was not clear from the incoming correspondence that a redistribution of costs was being contemplated by the hospitals. In addition, the letters did not explicitly mention that the costs of the program were currently being borne by the community in the contemplated arrangements. In the last 2 or 3 years, when hospitals met with or wrote to CMS for guidance on the nonhospital site policy under § 413.86(f)(4), we provided responses that were limited to the scope of the inquiries. We answered questions about the requirements of § 413.86(f)(4). It did not seem necessary to bring up the issue of “redistribution” or “community support” because it was not apparent that the community had previously incurred the direct GME costs. It was not until the relatively recent audits by our fiscal intermediaries of the fiscal year ending 1998 and 1999 cost reports of certain hospitals that CMS became aware that cost shifting was occurring. With this awareness came the necessity to explicitly reassert and explain the application of the longstanding Medicare principles of redistribution of costs and community support.

Comment: Several commenters have stated that the principles of redistribution of cost and community support do not apply in determination of a hospital's FTE resident count for direct GME. One commenter argued, in part relying on a Federal district court case, Episcopal Hospital v. Shalala, 1997 U.S. Dist. Lexis 8701 (E.Da.Pa. 1997), to state: “* * * CMS has argued, and the courts have agreed, that Medicare cost principles have no effect with respect to the direct GME payment method prescribed by section 1886(h) of the Act * * * these principles implement the statutory provision in section 1861(v) of the [Social Security] Act for payment of reasonable cost.” This commenter also quoted extensively from the September 29, 1989 final rule to argue that the GME regulation “construes the GME statute so as to preclude consideration of allowable costs incurred in connection with a resident's training.”

Similarly, another commenter believed that Congress “replaced the old reasonable cost payment system” with a prospective payment methodology, and that those principles that formed the basis for reasonable cost payments for GME were no longer relevant. The commenter believed the redistribution of costs and community support principles have no application to the current payment methodology, which relies on FTEs and PRAs.

Several commenters also disputed our citation to the Thomas Jefferson case for application of the principles to FTE counts. The commenters believed that CMS should not use this case in support of our policy because the case did not discuss applying the principles to the counting of residents. In addition, they believe the case was “very limited” and “only discussed the establishment of base year resident costs, which were used in developing base payment rates.”

Response: We disagree with the commenters that the principles of redistribution of costs and community support do not apply in determination of a hospital's FTE resident count for direct GME. When Congress enacted section 1886(h) of the Act as part of section 9202 of the Consolidated Omnibus Budget Reconciliation Act (COBRA) of 1985 (Pub. L. 99-272) on April 7, 1986, it did not altogether “preclude” consideration of allowable costs in connection with a resident's training, as the first commenter suggests. Upon enactment of the new legislation, CMS (then HCFA) considered a hospital's allowable reasonable costs, and applied reasonable cost principles (including redistribution of costs and community support, as we have explained) to calculate a hospital's direct GME costs and FTE resident count in order to determine hospital-specific PRAs in the base year. Although in cost reporting years after the PRA base year, the applicable PRAs are largely determined by the statute, we believe that costs continue to be a factor in determining the number of FTE residents that may be counted by a hospital. For example, a hospital may only count FTE residents training at the hospital for which, as repeatedly described in the September 29, 1989 final rule, the hospital almost necessarily incurs some direct GME costs. Hospitals may also count FTE residents training in nonhospital sites only if the hospital incurs all or substantially all the training costs of the program at that site (and meets other regulatory requirements.) Thus, it cannot be said that our view of the statute “precludes” consideration of allowable costs associated with training residents.

Although Congress did implement a prospective payment system for direct GME costs by enacting section 902 of COBRA 1985, we do not believe this means that all reasonable cost principles are no longer applicable under the revised system. Section 1886(h)(1) of the Act provides that: “[n]ot withstanding section 1861(v) [defining reasonable cost], instead of any amounts that are otherwise payable under this title with respect to the reasonable costs of hospitals for direct graduate medical education costs, the Secretary shall provide for payments for such costs in accordance with paragraph (3) of this subsection.” The statute literally provides that the reasonable cost payment method in section 1861(v) of the Act does not apply to section 1886(h)(3) of the Act (but those principles do apply to the remainder of section 1886(h) of the Act), which is the paragraph that specifies the general prospective payment formula for direct GME (the direct GME PRA). Thus, section 1886(h)(1) of the Act does not, as the commenter suggested, preclude any consideration of reasonable costs associated with the training of residents. Indeed, section 1886(h)(1) of the Act provides that, instead of payment under section 1861(v) of the Act, “the Secretary shall provide for payment for such costs”, which refers back to “the reasonable costs of hospitals for direct graduate medical education costs.” Thus, the statutory provisions governing direct GME payments continue to contemplate that Medicare payments to hospitals will be made for reasonable costs even under the prospective payment that is based on direct GME PRAs and FTE residents. Therefore, we do not believe the statute precludes application of reasonable cost principles, including the principles of redistribution of costs and community support.

Although we do recognize that certain reasonable cost principles are inherently contrary to a prospective payment system, others are compatible and may continue to be relevant, even upon implementation of the prospective payment. For example, in the case cited by the commenter, the Secretary and the court acknowledged that the principle of “cross-subsidization” found in section 1861(v)(1)(A) of the Act does not apply under a prospective payment context. The cross-subsidization provision requires that, in determining the reasonable costs of services, the Medicare program must ensure that it bears fully, but exclusively, “the necessary costs of efficiently delivering covered services” to Medicare beneficiaries. Simply put, the provision requires the Medicare program to pay for all the costs associated with care for its beneficiaries, and no more, so that other parties are not subsidizing care provided to Medicare beneficiaries, and Medicare is not subsidizing care provided to non-Medicare beneficiaries. However, when Medicare payments are determined prospectively, the Medicare program necessarily ceases to be concerned about whether cross-subsidization occurs—in other words, it is expected that a particular provider's costs may be higher or lower than the prospectively-determined payment (hence, the underlying premise that prospective payment systems create incentives for providers to control costs and operate efficiently).

In contrast, the principles of redistribution of costs and community support are completely congruent with the prospective payment system under section 1886(h) of the Act. Redistribution of costs and community support principles derive from legislative intent that was expressed at the enactment of the Medicare program, that the program should not assume payment for education costs that were previously funded by other sources. There is no reason to conclude that this intent changed with the enactment of the prospective payment methodology in section 1886(h) of the Act, with the addition of the FTE caps specified in section 1886(h)(4)(e) of the Act, or with the amendments that allow hospitals to count residents training in nonhospital sites for purposes of direct GME and IME payments. We do not believe that Congress intended by any of these enactments to enable an expansion in Medicare direct (or indirect) GME payments that result from cost shifting to hospitals. Rather, we believe section 1886(h) of the Act and later amendments were primarily directed toward limiting expansion of Medicare direct GME and IME payments. Therefore, we believe that the principles of redistribution of costs and community support are consistent with, and continue to be applicable under, the current direct GME payment system.

We also believe it is appropriate to cite the Supreme Court in the Thomas Jefferson case. The commenters believed that the scope of the Supreme Court's opinion that supported the agency's application of the principles of redistribution of costs and community support is limited to the calculation of hospitals' reasonable costs of GME for the purpose of determining the base period PRA. However, as we stated above, the statutory provisions governing direct GME payments continue to contemplate that Medicare payments to hospitals will be made for “such costs” even under the prospective payment methodology specified in section 1886(h) of the Act. In calculating the base year PRAs, the Agency allowed hospitals to count FTE residents where the hospitals were incurring direct GME costs associated with training those residents. This policy was clearly consistent with the principles of redistribution of costs and community support because the calculation of base year PRAs was dependent on the proper counting of FTE residents. Any opinion from the Court on the application of the principles to the base year costs would equally apply to FTE resident counts. Therefore, we believe the relevance of the Thomas Jefferson case is not limited to the establishment of base year costs, as the commenters suggested. Rather, the Court's opinion recognized that the principles of redistribution of costs and community support legitimately continue to apply under section 1886(h) of the Act. The Supreme Court's opinion is entirely relevant to the calculation of direct GME payments to hospitals in cost reporting periods on or after the PRA base year.

Finally, to address the commenters' reference to the 1989 final rule to support the argument that CMS interpreted the statute to preclude consideration of costs in connection with counting FTE residents, we note that the cited rule is replete with suggestions that CMS expected hospitals to continue to incur some level of direct GME costs for training residents, even under the direct GME PRA-based payment methodology. For example, the final rule at 54 FR 40298 states:

“Nothing in section 1886(h) of the Act indicates that the bearing of costs in connection with particular residents is a factor in determining who should be counted. The law simply requires the Secretary to determine the average amount incurred to train residents during the specified base period and to make GME payments for the residents in the hospital's programs thereafter on that basis. There was no authorization to establish a two-tiered system to account both for residents whom the hospital incurs full training costs and for residents whom hospitals incur only supervisory and overhead costs because the residents' salaries are paid by another entity.” (Ibid.)

We believe the language quoted above from the 1989 rule is exemplary of the Agency's mindset (as well as of the mindset of the commenter in that rule) that the question of whether costs were incurred by the hospital was, and would continue to be, a consideration for purposes of direct GME payment.

Comment: One commenter appeared to agree with what we stated in the proposed preamble at 68 FR 27216 that because IME regulations on counting residents at nonhospital sites cross-reference the direct GME nonhospital provisions, the provisions on redistribution of costs and community support would equally apply to IME FTE counts, as well as direct GME FTE counts, when counting residents in nonhospital settings. However, the commenter requested clarification on the issue of whether IME FTE residents counts in hospital settings would be subject to the community support and redistribution of costs provisions.

Another commenter argued that the redistribution of costs and community support principles do not apply to FTE counts for purposes of IME payment. This commenter argues that there is no evidence indicating that a teaching hospital's operating costs bear any relation to past or present sources of funding for residents' training.

Response: In response to the commenters' concerns regarding the application of the redistribution of costs principles and community support to counting residents for purposes of determining payments for IME for training in hospital settings, we agree with the commenters; the redistribution of costs and community support principles do not apply to FTE counts for residents training in hospital settings for purposes of IME payment. As we have explained in several regulations, the object of IME payments associated with resident training in hospital settings is to address the additional indirect operating costs that teaching hospitals incur in furnishing patient care (see 66 FR 39896 or 54 FR 40286). Even if the redistribution of costs and community support principles could theoretically apply to training inside the hospital, we do not know how all of these additional indirect operating costs incurred by a hospital could be “redistributed” to a nonhospital entity or could be borne by the community. As long as the hospital had consistently incurred at least some of those indirect costs, there could be no violation of redistribution of costs and community support principles, and no resulting disallowance of FTEs in calculating the hospital's IME adjustment. In any event, as stated above, we agree with the commenters because we believe the legislative history that gave rise to the principles of redistribution of costs and community support was focused on Medicare payments for direct GME.

However, we note that, for training that occurs in nonhospital settings, the application of the principles of redistribution of costs and community support to direct GME FTE counts does have implications for IME payment for residency training in nonhospital settings. Under existing § 412.105(f)(1)(ii)(C), which is the rule for the counting of FTE residents training in nonhospital settings for IME payment, there is a specific reference indicating that the criteria set forth in § 413.86(f)(4) must be met in order for a hospital to count the FTE residents training in nonhospital settings for purposes of IME payments. Thus, if under § 413.86(f)(4)(iv) (the paragraph that specifically applies redistribution of costs and community support principles to FTE counts for purposes of direct GME) a hospital is not permitted to count the FTE residents training in a nonhospital site because of redistribution of costs or community support, the hospital would not be permitted to count the FTE residents for purposes of IME payment as well, because the IME regulation at § 412.105(f)(1)(ii)(C) requires the criteria under § 413.86(f)(4) to be met.

As we have stated above, IME payments are based on the concept that, as a direct result of the hospital's resident training program, the hospital incurs increased indirect costs for patient care. When Congress added section 1886(d)(5)(B)(iv) of the Act as part of Pub. L. 105-33, the circumstances under which IME payments to a hospital could be made were broadened to allow the hospital to count the number of residents training outside the hospital setting under certain conditions, even though it is clear residents training outside the hospital cannot have any impact on the hospital's indirect patient care costs. Nevertheless, Congress authorized hospitals to receive IME payments by allowing hospitals to count FTE residents training in a nonhospital setting in accordance with section 1886(d)(5)(B)(iv) of the Act. As we have stated, we believe Congress intended the provisions of Pub. L. 105-33 to create an incentive (or remove a disincentive), for hospitals to rotate residents to nonhospital settings by allowing hospitals to continue to receive IME payment as if the residents continued to train in the hospital setting. However, we believe IME payment to the hospital would be contrary to Congressional intent if there is a redistribution of costs or community support associated with residents training in a nonhospital site. We also believe the IME payment to the hospital was only intended by Congress to encourage the hospital to rotate residents from the hospital to the nonhospital site, not to encourage (or enable) existing training programs to transfer their costs to the hospital and thereby expand the hospitals Medicare IME payments.

In addition, when Congress added section 1886(d)(5)(B)(iv) to the Act as part of Pub. L. 105-33, the statutory authority for IME payment for residents training at a nonhospital site was premised on the hospital incurring the direct GME costs of the residents: “all the time spent by an intern or resident in patient care activities under an approved medical residency program at an entity in a nonhospital setting shall be counted towards the determination of full-time equivalency if the hospital incurs all, or substantially all, of the costs for the training program in that setting.” (Emphasis added.) (Section 4621(b)(2) of Pub. L. 105-33; section 1886(d)(5)(B)(iv) of the Act.) The statute requires a hospital to incur “all or substantially all of the costs for the training program” in the nonhospital setting in order to count FTE residents training there for purposes of both direct GME and IME payment. The link between the IME regulation at existing § 412.105(f)(1)(ii)(c) and direct GME regulations at § 413.86(f)(4) implement this shared statutory requirement. As we have stated, we believe Congress intended hospitals to facilitate training in nonhospital sites that would not have occurred without the hospital's sponsorship, and for the hospital also to incur direct GME costs of the program in the nonhospital site as a precondition to counting the FTE residents training in nonhospital settings for purposes of IME payment. Thus, in the situation where a hospital currently is incurring direct GME costs at the nonhospital site but there has been a redistribution of costs or community support, a disallowance of direct GME payments, as well as a disallowance of IME payments, is appropriate.

Comment: One commenter noted that proposed § 413.86(i) (redistribution of costs and community support provision) applies not only to subparagraph (f)(4), the nonhospital site provision, but also to the remaining provisions of paragraph (f) and also to paragraphs (g)(4) through (g)(6). The commenter requested that CMS specify that the principles affect only the counting of residents in nonhospital sites and not the count of residents being trained in hospitals, both the inpatient and outpatient settings. In addition, this commenter believes such a clarification would also be consistent with other Medicare policy on counting FTE residents, such as the policy detailed in the August 1, 2002 final rule (67 FR 50077) concerning when residents rotate to other hospitals: “which entity may count the residents for IME and Direct GME payments is based on where the actual training occurs, not which hospital is incurring the costs.”

Response: While the primary reason we proposed to make the principles of redistribution of costs and community support explicit in the direct GME regulations was to specifically address the inappropriate scenarios described in the proposed rule whereby hospitals increase their FTE resident counts by assuming payment of training costs for residents in nonhospital sites that were previously funded by a nonhospital entity, we do not believe the principles are applicable in only this circumstance. In other words, the principles of community support and redistribution of costs apply generally to direct GME FTE counts, as we have explained. This holds true whether the counts relate to residents training in nonhospital sites (where we have seen the most inappropriate counting), or to residents training inside the hospital—inpatient or outpatient. Thus, it is technically possible to have a redistribution of direct GME costs for the training of residents inside the hospital setting (as well as in the nonhospital setting). Therefore, we are not adopting the commenter's suggestion to limit application of the principles to § 413.86(f)(4) (the nonhospital site provision). However, we note that we believe a redistribution of all of the direct GME costs for training that occurs in a hospital setting would be rare. All of the direct costs of the program—resident salaries, teaching physician salaries, overhead expenses, etc., would need to be redistributed to an outside entity in order for there to be a disallowance of direct GME FTE residents for training inside the hospital due to redistribution of costs or community support.

We contrast this application of the principles of redistribution of costs and community support in the current prospective payment system that depends upon PRA and FTE resident counts to application of the principles in the previous reasonable cost payment methodology that was based on cost finding and cost allocations. Under the former reasonable cost methodology, a hospital was eligible to receive direct GME payment for those direct GME costs that it incurred; however, any direct GME costs that were redistributed to the hospital were not allowable. We note that the instructions that CMS (then HCFA) gave to its Regional Offices in the 1990 audit instructions for purposes of calculating the direct GME base period PRA specifically addressed redistribution of costs and community support in the GME context:

Where costs for services related to medical education activities have historically been borne by the university, it is assumed the community has undertaken to support these activities, and subsequent allocation of these costs to a hospital constitutes a redistribution of costs from an educational institution to a patient care institution. In such a situation, these costs are not allowable under the Medicare program. (See 42 CFR 413.85(c) and HCFA Pub. 15-1, § 406). For example, if in the past the hospital did not identify and claim costs attributable to the time teaching physicians spent supervising I&Rs [interns and residents] working at the hospital, it is assumed that these costs were borne by the university. Therefore, the hospital may not claim these costs in subsequent cost reports. (Instructions for Implementing Program Payments for Graduate Medical Education to ARAs for Medicare, Director of Office of Financial Operations of the Health Care Financing Administration, BPO-F12, February 12, 1990.)

Thus, under the previous cost payment scheme, the principles of redistribution of costs and community support were applied to direct GME reasonable cost payment using a cost finding methodology. In contrast, in the current context where payment is no longer based solely on reasonable costs incurred, but on PRA and FTE resident counts, if the hospital can demonstrate that it has continuously incurred some of the direct GME costs of training the residents since the inception of the residency program at a training site, then no redistribution of costs or community support has taken place. As noted, current direct GME payments are no longer based on detailed cost finding of allowable costs of hospitals. Therefore, we believe it is appropriate to require that a hospital demonstrate that there has been no redistribution of costs or community support by proving that the hospital has incurred some of the direct GME costs of the program continuously since the inception of the program. Finally, contrary to the commenter's assertion, we believe we have been consistent with the other Medicare policies on counting residents, including the policy cited by the commenter concerning the prohibition on counting residents training at other hospitals. (See the August 1, 2002 final rule (67 FR 60077). As stated above, there would be no redistribution of costs or community support if a hospital counts a resident when another hospital incurs the resident's salary, as long as the first hospital still incurs other direct GME costs associated with the training of that resident. In any case, as we explained above and also in the proposed rule, the principles of redistribution of costs and community support are not applicable to cost shifted between the hospitals, only costs shifted between a hospital and educational institutions or other organizations that are not Medicare providers.

Comment: One commenter stated that a hospital was “required” to include in the calculation of its average per resident amount, time spent in the hospital by residents who were paid by “other entities.” This commenter quoted the September 29, 1989 final rule: “the 1989 GME rule was modified after publication of the proposed rule in order ‘to require Medicare hospitals to count residents who are working in their facility even if the residents' salaries are fully paid by other entities, either Federal or non-Federal. This revised policy will apply to both GME base period and cost reporting periods subject to the new payment methodology.’ 54 FR 40299 (emphasis added).”

Response: We believe the language quoted above by the commenter from the 1989 final rule has been taken out of context. In essence, the commenter has generalized from the language selectively quoted above to support an argument that Medicare would have required a hospital to count resident time when the residents were “paid by other entities,” thereby supporting the commenter's argument that Medicare not only condones redistribution of costs but, in fact, would seem to “require” them. However, we believe the language quoted by the commenter from a particular comment and response in the 1989 rule, if quoted in its full context, actually supports the CMS policy on the application of the principles of redistribution of costs and community support that as long as the hospital has continuously incurred at least some of the direct GME cost of the residency program since the inception of the program, there has been no redistribution of costs or community support and the hospital may count the FTE residents. Specifically, the commenters in that rule at 54 FR 40298 asked in relevant part: “A particular problem referred to was the treatment of residents who are paid by medical schools, faculty practice plans, and others rather than by hospitals that participate in Medicare. It was pointed out that teaching hospitals incur other costs such as teaching physicians' salaries and overhead costs in connection with these residents, and it would be unfair not to count these residents for payment purposes.” In our response to this comment, we stated, also in relevant part on 54 FR 40299: “we note that some of the comments have led us to believe that, in addition to Federally-employed residents (for example, residents in Veterans Administration or Department of Defense programs), a significant number of residents are paid a salary by non-Federal, nonprovider entities (for example, medical schools or philanthropic agencies). As noted by the commenters, although no hospital participating in Medicare incurs salary costs for these residents, hospitals do incur other substantial GME costs associated with these residents. Therefore, we are modifying our proposed rule to require Medicare hospitals to count residents who are working in their facility even if the residents' salaries are fully paid by other entities, either Federal or nonfederal.” (Emphasis added). It becomes apparent when the language quoted by the commenter on this final rule is read in context that, even as far as back as the 1989 final rule, we acknowledged that hospitals may count the FTE residents where other entities may have incurred the residents' salaries, but where the hospitals still “incur other substantial GME costs associated with these residents.” This view is entirely consistent with the CMS application of redistribution of costs and community support. In a scenario where a nonhospital entity, such as a medical school, incurs the residents' salaries, we continue to believe that the hospital may count the FTE residents if the hospital can demonstrate that it has incurred other direct GME costs, such as the supervisory physician salaries, since the inception of the program.

Comment: One commenter argued that when we explained our policy in the July 31, 1998 Federal Register (63 FR 40954) to require a written agreement indicating that the hospital must provide reasonable compensation for physicians' supervision of residents' training in the nonhospital setting, “nothing was said about an additional requirement that a hospital must have continuously incurred this additional cost, as well as the residents' compensation required under the prior regulations, since the inception of the training program.” This commenter further makes the point that in the final rule at 63 FR 40986, in response to a comment that hospitals did not compensate nonhospital sites for supervisory teaching physician costs and it would not be fair to shift these costs to teaching hospitals, CMS responded:

Hospitals and nonhospital sites will have 5 months following publication of this final rule to negotiate agreements that will allow hospitals to continue counting residents training in nonhospital sites for indirect and direct GME. These arrangements are related solely to financial arrangements for training in nonhospital sites. We do not believe that the agreements regarding these financial transactions will necessitate changes in the placement and training of residents.

In response to the comment that it is unfair to shift costs to the hospital, we believe that it is appropriate to include supervisory costs in the nonhospital site as part of “all or substantially all” of the costs that hospitals must incur to count the resident. Currently, the hospital is able to count the resident even though the costs for that resident may be lower during the time when the resident trains outside the hospital. At the same time, the nonhospital site may have incurred costs for which it received no compensation. We believe that requiring the hospital to incur the costs associated with training in the nonhospital site is equitable to both the hospital and the nonhospital site and is consistent with the statutory requirement that the hospital must incur “all or substantially all” of the costs.

(63 FR 40995 (emphasis added by commenter).)

The commenter believed that this explanation of the changes to the GME and IME rules, effective January 1, 1999, “belies CMS' current assertion of a longstanding policy of applying the redistribution of costs and community support principles in the determination of the resident counts used to compute payment for GME and IME.”

Response: The commenter has used the language quoted above from the 1998 final rule to argue that when CMS (then HCFA) described the policy on counting residents in nonhospital sites for IME, “nothing was said about an additional requirement that a hospital must have continuously incurred this additional cost * * * since the inception of the training program.” The commenter has inferred from the language quoted above that CMS has not had a longstanding policy of applying the redistribution of costs and community support principles. However, we believe the language actually supports the longstanding existence of our policy in two ways. First, the quoted language demonstrates the agency's view that the nonhospital site policy was written from the standpoint of addressing the counting of residents when hospitals rotate residents from the hospital to the nonhospital site. Second, the quoted language is also indicative of the Agency's policy that as long as the hospital has continuously incurred at least some of the direct GME cost of the residency program since the inception of the program, there has been no redistribution of costs or community support and the hospital may count the FTE residents (assuming that other requirements are met).

Specifically, the comment relating to the portion of the 1998 final rule quoted above stated at 63 FR 40994, in relevant part: “One commenter noted that some arrangements between hospitals and nonhospital settings for the training of residents predate the GME base year. This commenter stated that hospitals did not compensate nonhospital sites for supervisory teaching physician costs and it would not be fair to shift these costs to teaching hospitals. The commenter also stated that teaching hospitals have already entered into written agreements with nonhospital sites under the existing rules.” (Emphasis added.) In addition, as quoted above in the comment, we responded, in relevant part at 63 FR 40995 (with different emphasis):

* * * hospitals and nonhospital sites will have 5 months following publication of this final rule to negotiate agreements that will allow hospitals to continue counting residents training in nonhospital sites for indirect and direct GME. These arrangements are related solely to financial arrangements for training in nonhospital sites. We do not believe that the agreements regarding these financial transactions will necessitate changes in the placement and training of residents.

In response to the comment that it is unfair to shift costs to the hospital, we believe that it is appropriate to include supervisory costs in the nonhospital site as part of “all or substantially all” of the costs that hospitals must incur to count the resident. Currently, the hospital is able to count the resident even though the costs for that resident may be lower during the time when the resident trains outside the hospital. At the same time, the nonhospital site may have incurred costs for which it received no compensation. We believe that requiring the hospital to incur the costs associated with training in the nonhospital site is equitable to both the hospital and the nonhospital site and is consistent with the statutory requirement that the hospital must incur “all or substantially all” of the costs. Ibid.

We believe the quoted comment and response from the 1998 rule paint a picture of a hospital that has had a pre-existing relationship with a nonhospital site involving rotation of residents from the hospital to the nonhospital site for a period of time during the residency program. The language we emphasized in the response—that the hospital may “continue to count residents” when they train in the nonhospital sites, and that the hospital “may count the resident even though the costs for the resident may be lower during the time when the resident trains outside the hospital”—clearly refers to a rotational arrangement between the hospital and the nonhospital site. In addition, according to the circumstances described by the commenter in the 1998 rule, the hospitals had been incurring the residents' salaries, a direct GME cost, because they had formerly complied with the earlier regulation requiring that hospitals incur residents' salaries for purposes of meeting “all or substantially all of the costs” under § 413.86(f)(3). We had no reason to believe that the hospitals had not incurred at least the residents' salaries since the inception of the training program (the commenters state that the arrangements “predate the GME base year”). In that event, the counting of residents in the nonhospital sites would not result in a redistribution of costs if, as of January 1, 1999, the hospital was required to incur the additional direct GME cost for supervisory physician costs while the residents rotate to the nonhospital site. We believe that the commenter in the 1998 rule simply did not agree with the additional regulatory requirement finalized in the 1998 final rule that the hospital must also incur the supervisory physician costs for purposes of incurring “all or substantially all of the costs,” and hoped to label this new regulatory requirement as a “cost shift” in order to avoid it. As we have explained, it appears that there has been no redistribution in the case described by the 1998 final rule commenter because it can be inferred that the hospital had incurred at least some of the direct GME costs (the residents' salaries) since the inception of the program.

Therefore, we believe the language the commenter quotes from the 1998 rule is consistent with our clarifications in this final rule on redistribution of costs and community support. In addition, the language cited by the commenter supports our interpretation of the policy on counting residents in nonhospital sites that it was intended to address the situation when hospitals rotate residents from the hospital to the nonhospital site.

Comment: Some commenters disputed the CMS interpretation of Congressional intent as discussed in the preamble of the proposed rule (see 68 FR 27213). One commenter stated: “there is no support in the legislative history of the non-provider setting amendments [the 1986 and 1997 amendments of the Act] for the Secretary's view that these changes were not intended to shift new costs to hospitals in support of on-going training in non-provider settings * * * it can be reasonably inferred that Congress was aware, and even intended, that some costs of existing residency training programs in non-provider settings would be shifted to hospitals in order for the hospitals to qualify for direct GME and IME funding under the 1986 and 1997 amendments of the Act.” Similarly, another commenter stated that the Secretary “must look elsewhere to the statute [other than section 1886(h)(4) of the Act] for support for his proposed rule; he cannot simply create out of whole cloth an interpretation that is inconsistent with the amendment's other provisions.”

Response: The commenters would have us interpret and implement policy in a statutory vacuum. We believe we have reasonably discerned Congressional intent by interpreting the plain language of the statute at sections 1886(d)(5)(B) and 1886(h) of the Act in conjunction with the accompanying legislative history of these sections.

As we stated in the preamble to the proposed rule, Congress has delegated broad authority to the Secretary to implement a policy on the count of FTE residents for purposes of calculating direct GME and IME payments. In section 1886(d)(5)(B) of the Act (IME), the statute does not specify at all how FTE counts should be determined, and the plain language in the statute under section 1886 (h)(4) of the Act (direct GME) indicates that the Secretary “shall establish rules” for direct GME consistent with the statute. We also considered the deference expressed in the conference agreement that accompanied Pub. L. 105-33, which established a cap on the number of allopathic and osteopathic residents a hospital may count—“[T]he Conferees recognize that such limits raise complex issues, and provide for specific authority for the Secretary to promulgate regulations to address the implementation of this provision.”(H.R. Conf. Rep. No. 105-217, 105th Cong., 1st Sess., 821 (1997).

Thus, in the absence of statutory specificity on determining FTE counts and the declared Congressional delegation of authority to the Secretary on the subject are clear indications that Congress has given the Secretary broad discretion to promulgate reasonable regulations in order to implement the policy on the counting of residents for direct GME and IME payments.

In addition, we have not, as the second commenter suggests, “created out of whole cloth” an interpretation of the policy concerning counting residents in nonhospital settings that is “inconsistent with the amendment's other provisions,” nor do we at all believe that “it can be reasonably inferred that Congress was aware, and even intended, that some costs of existing residency training programs in non-provider settings would be shifted to hospitals in order for the hospitals to qualify for direct GME and IME funding under the 1986 and 1997 amendments of the Act,” as the first commenter suggests. Rather, as we have stated, we believe that when Congress created the provisions on counting resident FTEs in nonhospital settings, it was creating a monetary incentive for hospitals to rotate residents from the hospital to nonhospital settings. We have drawn this conclusion, as we explained, from the legislative history of both the direct GME and IME provisions authorizing payments to hospitals for training in nonhospital settings. First, legislative history associated with passage of the direct GME provision (as part of Pub. L. 99-509) indicates that Congress intended to broaden the scope of settings in which a hospital could train its residents and still receive separate direct GME cost reimbursement, and to provide incentives to hospitals for training residents in primary care programs. The Conference committee report indicates that “[s]ince it is difficult to find sufficient other sources of funding [than hospitals and Medicare] for the costs of such training, [that is, training in freestanding primary care settings such as family practice clinics or ambulatory surgery centers] assignments to these settings are discouraged. It is the Committee's view that training in these settings is desirable, because of the growing trend to treat more patients out of the inpatient hospital setting and because of the encouragement it gives to primary care.” (Emphasis added.) (H.R. Rep. No. 99-727, 99th Cong., 1st Sess., 70 (1986).)

Thus, from the inception of the policy allowing payment for training in nonprovider sites, we believe Congress intended to create a monetary incentive for hospitals to rotate residents from the hospital to the nonhospital settings. We do not believe Congress intended for hospitals to be paid for residents who had previously been training at nonhospital sites without hospital funding.

Further, in the Conference committee report accompanying the provision of Pub. L. 105-33 that authorizes IME payment for training in nonhospital settings, Congress stated that “[t]he conference agreement includes new permission for hospitals to rotate residents through nonhospital settings, without reduction in indirect medical education funds.” (Emphasis added.) (H.R. Conf. Rep. No. 105-217, 105th Cong., 1st Sess., 817 (1997).)

We note that, prior to enactment of Pub. L. 105-33, if a hospital rotated a resident from the hospital to train at a nonhospital site, the hospital could not count the time the resident spent at the nonhospital site for purposes of Medicare IME payments. As a result, the “loss” of IME payments acted as a disincentive and discouraged hospitals from rotating residents out of the hospital. It appears from the legislative history that Congress authorized hospitals to count residents in nonhospital sites for IME purposes as a specific incentive to encourage hospitals to rotate their residents to nonhospital sites (and not to encourage hospitals to incur the costs of a program at a nonhospital site that had already been funded by other sources). This legislative intent becomes more apparent when the nature of the Medicare IME payment is considered. The Medicare IME payment is inherently a payment that reflects the increased operating costs of treating inpatients as a result of the hospital having a residency program. For example, as explained in the September 29, 1989 final rule (54 FR 40286), the indirect costs of medical education might include added costs resulting from an increased number of tests ordered by residents as compared to the number of tests normally ordered by more experienced physicians.

The IME payment is an “add-on” adjustment that is made for each Medicare discharge from the areas subject to the IPPS in a teaching hospital. The authorization by Congress for IME payments relating to nonhospital services while residents are training at nonhospital sites would be absurd if not viewed as an incentive to transfer existing residency training from the hospital to the nonhospital setting. We do not believe Congress intended to permit IME payments to be allowable to the hospital that is incurring “all or substantially all the costs” of residents training in nonhospital sites except in the situations where either the hospital rotated residents from the hospital to the nonhospital settings or where the hospital started new programs in the nonhospital settings (and incurred the direct GME costs from the programs' inception). The illustrative situations described above and in the proposed rule in which nonhospital sites, such as dental schools, are shifting the costs of existing programs to the hospitals are not consistent with the intent of Congress to encourage hospitals to rotate residents from the hospital setting to nonhospital sites.

Thus, we believe Congress intended both cited provisions of the Act on counting residents in nonhospital sites for purposes of direct GME and IME payments to be limited to situations in which hospitals rotate residents from the hospital to the nonhospital settings, and not situations in which nonhospital sites transfer the costs of an existing program at a nonhospital site to the hospital.

Comment: One commenter cited section 1886(h)(5)(J) of the Act to support the general argument that CMS lacks the authority under the statute to “impose additional conditions” on counting FTE residents training in nonhospital sites—that is, the principles of redistribution of costs and community support. The commenter stated:

This conclusion is further supported by Congress' treatment of family practice residency programs. In 42 U.S.C. § 1395ww(h)(5)(J), Congress provided a special payment provision for family practice residency programs. Specifically, Congress authorized hospitals to claim costs related to such programs even if, during the GME prospective payment base year—a year reimbursed under the reasonable cost system and a year to which the community support principle applied—the cost of such programs had been paid by the United States, a State, a political subdivision of the State, or an instrumentality of the State or political subdivision. Congress also provided that, in the event that such program payments were part of the PRA calculation during the GME base year, the payment in future years would be reduced “in an amount equal to the proportion of such program funds received during the cost reporting period involved * * *.” Thus, Congress has spoken to the issue of whether hospitals may claim costs in the current year if those costs have been paid in the past by third parties, and it has allowed reduction in current-year payments only if: (1) During the GME PPS base year, a third party had paid for the cost of the hospital's family practice residency program; and (2) as a result, the hospital had received a PRA that included an “estimate of the amount that would have been recognized as reasonable * * * if the hospital had not received such funds.” 42 U.S.C. § 1395ww(h)(5)(J)(i). In all other situations, I submit, Congress does not permit the Secretary to reduce payments in the current year simply because, in the past, some third party may have paid the cost.

Response: We disagree with the commenter that section 1886(h)(5)(J) of the Act supports the assertion that “Congress has spoken to the issue” of whether a hospital may claim third party costs and has allowed reductions in direct GME reimbursement resulting from redistribution of costs or community support in only the very limited circumstance of that exception in the Act. Generally, section 1886(h)(5)(J) of the Act did two things: first, in subparagraph (J)(i)(1), Congress specifically allowed a hospital that only has an approved training program in family medicine and received a PRA in the base year of less than $10,000 for its family practice program, to receive a revised PRA that reflects the inclusion of “funds from the United States, a State, or a political subdivision of a State * * *” for the hospital's family practice program. Thus, the provision recognizes that ordinarily such funds would not be included in the hospital's base year per resident amount (because they were not incurred by the hospital in the base year). However, Congress explicitly created a narrow exception to the “cost finding” principles to allow such a hospital to include Federal, State, or local government grants to be included in the hospital's PRA base year calculation. Second, subparagraph (J)(i)(2) requires that direct GME payment to such a hospital that received a revised PRA amount under subparagraph (J)(i)(1) must also be reduced in subsequent cost reporting periods by the proportionate amount of funding the hospital receives from Federal, State, or local government payments. In other words, what subparagraph (J)(i)(2) does is to prohibit this hospital from receiving duplicative payments for the same GME program—both through the adjusted PRA and through continued Federal, State, and local government funding.

The commenter argues that subparagraph (J)(i)(2) is the “only” situation where Congress has “spoken” about reductions in current year payment because of third party reimbursement. However, as we stated above, we believe the effect of subparagraph (J)(i)(2) is to prevent of duplicative payments for the same program that could otherwise occur in the narrow circumstances of the exception provided by section 1886(h)(5)(J), and has nothing to do with the continued applicability of the principles of redistribution of costs and community support. To the contrary, as we have stated, we believe that subparagraph (J)(i)(1) addresses a limited theoretical “retroactive redistribution” of costs and community support to allow a very narrow exception of allowing costs to be included in direct GME payment. Thus, we believe section 1886(h)(5)(J) of the Act would support our assertion that Congress intends application of redistribution of costs and community support to direct GME payment (except in the narrow circumstance of the type of hospital described in that section), rather than support the commenter's contrary assertion that the section is inconsistent with our proposal on application of the principles.

Comment: One commenter suggested that the redistribution of costs and community support principles at nonhospital sites should apply on a “year-by-year basis,” such that if another entity funds a training program during a particular fiscal year, the hospital would not be allowed to include the residents in its count for that fiscal year.

Response: We believe the commenter's suggestion of a “year-by-year basis” policy is, in effect, already in place under existing Medicare policy without reference to the redistribution of costs or community support principles. Under the existing policy, where another entity funds a training program in a particular year while the residents are training at a nonhospital site—that is, incurs the residents' salaries and fringes, and the supervisory physician costs (“all or substantially all of the costs”), the hospital may not include the residents in its FTE count for that fiscal year. This requirement, of course, is independent of the redistribution of costs and community support policy. It is based on the statutory requirement that allows a hospital to count residents training at nonhospital sites only if the hospital has incurred for all or substantially all of the costs of the program at that site during the hospital's fiscal year.

Comment: One commenter stated that the 1989 final rule made clear that a hospital's resident count may also include residents for whom “community support was received” through a State or local grant. Similarly, another commenter stated “certain family medicine training programs that may have received outside funds, for example, State dollars, at any time in the past will be prohibited [by the hospital we proposed] from receiving GME reimbursement.”

Similarly, another commenter stated that “it is axiomatic” that State-supported and public teaching hospitals receive State appropriations to support their residency programs. The commenter urged CMS to clarify that the application of the redistribution of costs and community support principles would not apply to State or local appropriations to public hospitals, with respect to the counting of FTE residents in either the hospital or the nonhospital setting.

Response: As we explained in the 1989 final rule (54 FR 40302), grants that were restricted (those grants that were designated by the donor to pay for certain specified provider costs) or unrestricted were considered allowable costs of the hospital (including direct GME costs) when Medicare paid hospitals on a reasonable cost basis. The policy allowing payment to hospitals for costs that had been funded by grants was authorized by section 901 of the Omnibus Budget Reconciliation Act (OBRA) of 1980 (Pub. L. 96-499), which added section 1134 of the Act. Section 1134 of the Act applies to “the reasonable costs of services provided by nonprofit hospitals or critical access hospitals.” Section 1134(1) of the Act specifies that a “grant, gift or endowment or income therefrom which is to or for such a hospital * * *” may not be deducted from the operating costs of such hospitals that are paid on a reasonable cost basis. Therefore, when hospitals were paid on a reasonable cost basis for direct GME costs, the “community support” that came from “grants, gifts, or endowments” was allowable under Medicare. We are clarifying in this final rule, that under the direct GME prospective payment methodology under section 1886(h) of the Act, if a hospital had received a grant, gift or endowment to subsidize its residency programs at the hospital, and the hospital requested direct GME payment for training the residents, it would not be considered community support. Under section 1134 of the Act, it is as if the hospital had itself incurred the cost for which it had received the grant subsidy. For example, if in 2003 a hospital received a State grant to fund its family practice program at the hospital, the grant would not be considered community support under our regulation. This is because we would treat the hospital as if itself incurred the costs for the family practice program, instead of the State grant.

However, we note that this policy would not include ordinary State and local appropriations. As we mentioned in the January 12, 2001 final rule at 66 FR 3367, “In administrative, legal and policy matters, we have consistently maintained that State appropriations for the cost of medical education activities constitute community support that is to be offset from a provider's allowable costs.” Therefore, if a program were entirely funded by State or local appropriations, an inappropriate redistribution of costs would occur if the hospital subsequently begin to incur the costs of the residency program—for training inside or outside the hospital. Although, for most hospitals that receive State and local appropriations for their residency programs, the hospitals continuously incur (since the inception of the programs) some direct GME costs, there would be no disallowance of FTEs due to community support.

We contrast the situation of a grant to a hospital with the situation of a grant to a nonhospital site. If, hypothetically, nonhospital sites were reimbursed by Medicare on a reasonable cost basis, and the nonhospital site had received grants to subsidize all of the direct GME costs for the residency program there, under section 1134 of the Act, we would treat the costs the grant subsidized as if they were costs of the nonhospital site. If a hospital then tried to incur the direct GME costs, this could be a redistribution of costs or community support issue, since the hospital would be claiming FTE residents who had historically trained at the nonhospital site for whom the community had assumed the cost of that training, as described in the scenarios at 68 FR 27213.

Comment: Several commenters objected to the sentence in the preamble to the proposed rule that stated: “* * * a hospital is required to assume financial responsibility for the full complement of residents training in a nonhospital site in a particular program in order to count any FTE residents training there for purposes of IME.” One commenter explained that there are a number of situations where a hospital is truly incurring the cost of having a resident at a site, but the hospital is not incurring the cost of the entire complement of residents. “For example, if two different hospital programs each elect to send residents to the same clinic, under the interpretation in the [proposed rule], neither of the two hospitals would be able to count any of the residents because neither of the two programs would incur the cost of the full complement of residents.” Another commenter believed that “this change” runs contrary to other current Medicare policies that focus on the resident rather than the program. The commenter believed that both the direct GME and IME regulations “are replete with references to ‘resident’ rather than ‘program'.” The commenter believed that “residency program” is referenced only in the context of the requirement that, for residents to be counted for direct GME and IME payments, they must be part of an “approved program” (§ 413.86(f)(1)).

Response: We understand the concerns of the commenters about the requirement for a hospital to incur “all or substantially all of the cost” of training residents in a training program at a nonhospital site. However, we do not believe this is a change in policy. We believe that the policy that requires a hospital to incur the cost of “the program” in the nonhospital site has existed since the passage of the direct GME provisions, section 9314 of the Omnibus Budget Reconciliation Act of 1986 (Pub. L. 99-509), and the passage of the IME provision, section 4621(b)(2) of the Balanced Budget Act of 1997 (Pub. L. 105-33), that permitted hospitals to continue to count residents in nonhospital sites, for purposes of direct GME and IME payment, if the hospital incurred “all or substantially all of the cost” of residents training in the program.

As we explained in the proposed rule, this policy is derived from the language of the IME and direct GME provisions of the statute on counting residents in nonhospital settings; both sections 1886(d)(5)(B)(iv) and 1886(h)(4)(E) of the Act state that the hospital must incur “all, or substantially all, of the costs for the training program in that setting.” (Emphasis added.) Therefore, we believe a better reading of this language is that hospitals must incur all or substantially all of the cost for the full complement of residents in the training program at the nonhospital site.

We note that the policy that requires the hospital to incur the cost of the program does appear to be somewhat of a departure from other current Medicare policies on graduate medical education that focus on the resident rather than the program, as the commenter suggests. However, we believe the statutory provisions cited above require hospitals to assume the cost of the full complement of residents training in the program at the nonhospital sites in order to count any FTE residents training at that site.

In addition, as we noted at 68 FR 27217 of the proposed rule, and also above, under policy on the application of the redistribution of costs and community support principles, it is permissible for the hospital to count FTE residents where the hospital incurs direct GME costs of FTE residents that are added to an existing program, even though the hospital is not permitted to count the existing FTE residents due to the application of the redistribution of costs or community support rules. In the nonhospital setting, as a result of the interaction of these two separate FTE-counting requirements—(1) that the hospital must not violate the redistribution of costs and the community support principles in order to count the resident FTEs in the nonhospital settings; and (2) that the hospital must incur “all or substantially all” of the costs for the training program in that setting—a hospital would be prohibited from counting FTE residents added to an existing program at a nonhospital site unless the hospital incurs all or substantially all of the costs of training all of the residents in that program at that setting. That is, even if the hospital incurs all or substantially of the costs for all of the training program at the nonhospital site, the hospital would only be able to count the additional FTE residents who were not excluded by application of the redistribution of costs or community support principles.

Comment: Several comments cited a letter from CMS (then the Health Care Finance Administration, or “HCFA”) dated March 30, 1999 to C. Scott Litch of the American Association of Dental Schools (now the American Dental Education Association). Specifically, these commenters cited a sentence in the letter to Mr. Litch which stated: “If a hospital establishes a new relationship with a dental clinic and meets the conditions for counting residents training outside the hospital, the hospital may count more residents currently for indirect and direct graduate medical education than were counted in 1996 if those residents are dental residents.” One commenter stated that the “new relationship” referred to in the letter from CMS presupposes the existence of an ongoing program whose costs presumably had been met by means other than the hospital before the affiliation with a nonhospital dental clinic began. This commenter believed that this letter provided assurance to many hospitals that new affiliations with preexisting dental programs were permissible.

Response: We do not agree with the commenter that the sentence in the letter to Mr. Litch “presupposes the existence of an ongoing program” where the costs of such a program “had been met by means other than the hospital”. Rather, we believe the “new relationship” between the hospital and the dental clinic could be reconciled with application of the principles of redistribution of costs and community support and characterized by two possible interpretations, both of which would allow for the counting of residents in nonhospital sites—(1) where the hospital would rotate residents from the hospital to the nonhospital site; or (2) where the hospital would fund new training slots at the nonhospital site (the dental clinic referred to in the Mr. Litch's letter). Such assignments from the hospital to the dental clinic, or new residency training slots, would be the “new relationship,” but in either case, no redistribution would occur. Therefore, we do not believe the letter from 1999 is necessarily inconsistent with the principles of redistribution of costs and community support described in the proposed rule.

Comment: Many commenters, while remaining generally opposed to application of redistribution of costs and community support principles, requested that if CMS were to finalize the proposed rule, CMS apply the principles prospectively. One commenter, a dental school, explained that it had just admitted a new class of residents, many of whom will not complete their programs until 2006. The commenter believed that, in the application of the principles, CMS seeks to remove all Medicare funding for these residents retroactively. Along a similar vein, another commenter pointed out in support of the suggestion to apply the principles only prospectively, that the implementation of the proposed regulation would result in “substantial dislocation and hardship to hospitals, dental and other schools, and the residents themselves.” Therefore, the commenter believed CMS should indicate specifically in the final rule that such changes will only be applied to a provider's cost reporting period beginning on or after October 1, 2003, and CMS should not apply its final GME policy on redistribution of costs and community support to any prior cost reporting periods that remain open or unsettled, or are settled but potentially subject to reopening under the Medicare rules.

In addition, several commenters requested clarification regarding the effective date for the proposed application of the principles of redistribution of costs and community support to FTE counts. Specifically, the commenters point to the following language in the proposed rule:

• “A hospital must continuously incur direct GME costs of residents training in a particular program at a training site since the date the residents first began training in that site in order for the hospital to count the FTE residents.” (68 FR 27215)
• “We propose * * * to identify January 1, 1999, as the date our fiscal intermediaries should use to determine whether a hospital or another entity has been incurring the costs of training in a particular program at a training setting.” (68 FR 27216)
• “[i]f the fiscal intermediaries determine that there is a redistribution of costs or community support exists with respect to certain residents prior to January 1, 1999, a disallowance of direct GME and IME payment with respect to those FTE residents would certainly be required.” (68 FR 27216)
• “We are proposing that, effective October 1, 2003, in order for a hospital to receive IME and direct GME payment, the hospital must have been continuously incurring the direct GME cost of residents training in a particular program since the date the residents first began training in the program in order for the hospital to count the FTE residents.” (68 FR 27417)

Response: We have stated that we believe the principles of redistribution of costs and community support are longstanding Medicare policy. While we have reminded the public of the continuing application of the principles in various regulations and program guidance, we also recognize that CMS has not had occasion to invoke them in Agency policy expressions relating specifically to direct GME payments since the direct GME PRA base year.

As we have stated, we believe redistributions would occur only in rare circumstances for residency training inside the hospital. Between 1987 and 1997 when hospitals could count FTE residents training in nonhospital sites for purposes of direct GME payments, but not IME payments, we did not observe the kinds of inappropriate counting of FTE residents we described in our proposed rule. It is only since hospitals have been allowed to count FTE residents training in nonhospital sites for purposes of IME payment, that CMS has become aware that cost shifting has become prevalent in the hospital industry, which has implicated the principles of redistribution of costs and community support. Therefore, in general, we are implementing a prospective effective date of October 1, 2003, for purposes of payment. That is, for direct GME, effective for portions of cost reporting periods beginning with October 1, 2003, and for IME, effective for discharges occurring on or after October 1, 2003, a hospital must have been continuously incurring direct GME costs of residents training in a particular program since the date the residents first began training in the program in order for the hospital to count the FTE residents. We note that the effective dates apply only as they relate to disallowances of FTEs and bear no relation to determinations of redistributions or community support. Therefore, in general, a fiscal intermediary that determines that a redistribution of costs has taken place for a particular hospital prior to October 1, 2003, may disallow FTEs based on that determination beginning with October 1, 2003. For example, if a fiscal intermediary determines that a redistribution of costs has occurred that affected 10 FTEs for direct GME and IME during the hospital's cost report ending in fiscal year ending in 1999, the fiscal intermediary would take disallowances for those 10 FTEs, but not until October 1, 2003, for purposes of direct GME and IME payment.

In addition, because we have received a large number of public comments expressing surprise and confusion regarding our policy on these principles, we are grandfathering residents who began training in a program on or before October 1, 2003. That is, an FTE resident who began training in a residency program on or before October 1, 2003 (the effective date of this final rule), and with respect to whom there has been a redistribution or community support, may continue to be counted by a hospital for purposes of direct GME and IME payments after October 1, 2003, until the resident has completed training in that program, or until 3 years after the date the resident began training in that program, whichever comes first. We believe continued direct GME and IME payments to the hospital while the “redistributed” residents finish their training for up to 3 years is appropriate to address many situations in which nonhospital sites have made arrangements with hospitals to shift the costs of training those residents. We understand that, in nonhospital sites, virtually all dental residency programs are of a duration of 3 years in length or less. This policy addresses the situation pointed out by the dental school commenter and other commenters that a school may have just admitted a new class of residents, many of whom will not complete 3-year programs until 2006.

We note that this prospective “grandfather” policy does not apply to resident FTEs with respect to whom there has been a redistribution of costs or community support, and who begin training after October 1, 2003. In addition, those residents described above who began training in a program on or before October 1, 2003, may be counted until those particular residents finish their training in that program (or 3 years, whichever comes first). In order to count such residents, we are requiring that hospitals identify those residents (by social security number) to their fiscal intermediary and specify the length of time the hospital will be counting these FTE residents for direct GME and IME payment purposes.

We note that the policy described above that effectively “grandfathers” residents who began their training on or before October 1, 2003, applies only as it relates to payments to hospitals for those specified FTE residents, and bears no relation to determinations of whether a redistribution of costs or community support has taken place. Therefore, if a fiscal intermediary determines that a redistribution of costs has taken place with respect to residents counted by a particular hospital even prior to October 1, 2003, the intermediary will disallow any FTEs based on that determination, beginning October 1, 2003, except for the “grandfathered” residents. Hospitals that continue to count grandfathered FTE residents (where the costs of whom had been redistributed) may only do so until those residents finish their training in the specific program they were training in on or before or to October 1, 2003 (which would be no later than September 30, 2006, 3 years after October 1, 2003).

For example, a fiscal intermediary determines for a hospital's FYE December 31, 2003 cost report that a redistribution of costs has taken place with respect to certain FTEs the hospital counted for direct GME and IME (that is, the costs of training residents at a nonhospital site were incurred by a university from 1990 through 1999). Assume that 5 FTEs began training in a 2-year orthodontics program in a dental school on July 1, 2003, and another 5 residents begin their training in the same program on July 1, 2004. The 5 FTEs who began training on July 1, 2003, are “grandfathered,” and, therefore, the fiscal intermediary would not disallow these 5 FTEs as of October 1, 2003. The hospital may continue to count these 5 FTEs that began training on July 1, 2003 through June 30, 2005, when they finish the 2-year orthodontics program. We note that subsequent to completion of the 2-year orthodontics program on June 30, 2005, if any of these 5 FTEs participate in additional GME training programs, the fiscal intermediary would disallow these FTEs because disallowances for redistribution of costs and community support relate to FTE slots and not specific residents.

However, the 5 FTEs that began training in the 2-year orthodontics program on July 1, 2004 are not “grandfathered,” and, therefore, beginning July 1, 2004 of the hospital's December 31, 2004 cost report, the fiscal intermediary will disallow IME and direct GME payment associated with these 5 FTE slots.

Comment: Commenters disputed the situations we cited in the preamble to the proposed rule that were supposed to be illustrative of what we believe to be inappropriate application of Medicare direct GME and IME policy at 68 FR 27213. One commenter, in particular, requested information on the identity of programs cited in the examples.

Response: We do not believe it is appropriate to disclose the identities of those cited in the examples. Therefore, we are unable to respond to the commenters' points on the matter, except to state that the situations in the examples represent what we believed are the more “egregious” scenarios involving redistribution of costs and community support principles and inappropriate counting of FTE residents, we note that the same issues arise, and the same principles apply, whether the counting of residents relates to training that is taking place in another country, another State, or on the same hospital campus, as the hospital.

Comment: One commenter believed that CMS's policy on the application of the redistribution of costs and community support will lead to considerable, “but needless,” litigation over what it means to “incur” the costs of off-site training.

Response: We disagree with the commenter and see no reason to be concerned that these clarifications would result in any more litigation than other Medicare payment policies that are conditioned on whether a provider incurs costs. For example, for several decades, Medicare policy required that hospitals “incur” costs in order to receive payment from Medicare. The Medicare statute and regulations currently require that a hospital incur certain costs in order to count FTE residents training in nonhospital sites for purposes of direct GME and IME payments. We are unsure why the requirement under the policy on redistribution of costs and community support that a hospital “incur” the direct GME cost continuously for a residency program at a training site is any more complex than other cost requirements under Medicare.

Comment: One commenter suggested that we craft a narrower solution to the issue of inappropriate counting of FTE residents in nonhospital sites by focusing the language on salary and benefits for residents. The commenter believed that CMS could state that, unless the hospital in 1999 had incurred the costs of salary and benefits for FTE residents who were training in offsite locations, the hospital may not receive direct GME and IME payment for training those FTE residents at the nonhospital sites today.

Response: We do not believe a policy such as the one the commenter suggested—determining redistribution of costs based upon whether a hospital continuously incurs the residents' salaries and benefits during training in the nonhospital site— is necessary or appropriate. This is because, under the policy on redistribution of costs and community support we describe in the proposed rule and in this final rule, a hospital that continuously incurs the residents' salaries and benefits (from 1999 or before) while the residents train in the nonhospital site, or even inside the hospital, would not be redistributing costs if the nonhospital site later incurs the other direct GME costs (such as supervisory physician salaries) in the nonhospital site. There would be no redistribution of costs because the hospital would have continuously incurred at least some of the direct GME costs (the residents' salaries and benefits) since the inception of the program. However, we note that even if there has not been a redistribution of costs or community support with FTE residents training in a nonhospital site in such a scenario, the hospital would still need to meet the requirements in the existing regulations (at § 413.86(f) and § 412.105(1)(ii)(c)) in order to count those FTE residents for purposes of direct GME and IME payment.

For example, Hospital A has had a family practice program with 10 FTE residents for about 20 years, for which the hospital has incurred the residents' salaries and fringes and some other (but not all) direct GME costs for the program. For the first time, in fiscal year ending 2003, Hospital A rotates 2 FTE residents to an ambulatory clinic (a nonhospital site), and fulfills the requirements at § 413.86(f)(4), including incurring “all or substantially all of the costs” of the training program in the nonhospital site. There is no redistribution of costs with respect to these 2 FTE residents because Hospital A has continuously incurred some of the direct GME costs of the program—the residents' salaries—and therefore it may count the 2 FTE residents training at the clinic (up to the hospital's FTE cap), since it also has complied with the requirements at § 413.86(f)(4).

Comment: Some commenters suggested that the application of redistribution of costs and community support principles would impose large administrative burdens on hospitals to demonstrate which entity has been “continuously incurring” the costs of the residency training. One commenter stated: “[t]his burden would be additive to a policy that already is fraught with excessive administrative requirements.”

One commenter asked if hospitals would be required to document responsibility for the costs of training residents prior to January 1, 1999.

Response: If the hospital has continuously been incurring at least some of the direct GME costs (for example, resident salaries or supervisory physician salaries) since the inception of the residency program, we do not believe any additional documentation is necessary beyond which hospitals are already required to maintain. If resident or supervisory physician salaries, for instance, are paid through the hospital payroll, the hospital will have kept documentation of such costs for Federal tax purposes.

In response to the second comment, we stated in the proposed rule that January 1, 1999 should be used by our fiscal intermediaries as the date for determinations of whether a hospital or another entity has been incurring the costs of a training in a particular program at a training site for purposes of determining whether there has been a redistribution of costs or community support. This date was chosen as an administrative convenience because we believe it could otherwise be difficult for our fiscal intermediaries to obtain contemporaneous documentation that the hospitals have appropriately been incurring costs in earlier years. Therefore, we believe that, for purposes of determining redistribution of costs or community support, most hospitals would only be required to maintain appropriate documentation to demonstrate that they have continuously been incurring the direct GME costs from January 1, 1999 forward. However, as we mentioned in the proposed rule, if the fiscal intermediaries determine that there was a redistribution of costs or community support for a fiscal year ending for a cost report for a particular hospital prior to January 1, 1999, the hospital would be required to show contemporaneous documentation to prove otherwise.

Comment: One commenter stated that it may be difficult to track residents that have been funded by some type of community support. The commenter described a scenario where a program at a hospital has four internal medicine residents and one is covered by some type of community support for a 3-year period. The commenter stated that it may be difficult to track that slot over the next 5, 10, or 20 years to avoid submitting it for future direct GME or IME payments.

Response: As we stated above, we understand there may be administrative issues that hospitals must confront in their efforts to comply with the principles of redistribution of costs and community support. However, we do not believe it would very difficult to track the FTEs in a program that receives community support. Once the FTE residents for which community support is received have been identified, the hospital will know the number of FTE residents to remove from the count that is submitted in future cost reports (all of which will be subject to audit by our fiscal intermediaries). Using the commenter's example, if direct GME costs for one out of four FTEs in an internal medicine program is identified as being entirely subsidized by community support for three years (the duration of an internal medicine program), the hospital would know to refrain from counting one FTE in future cost reports, even after the 3 years of training for a particular resident has passed. This is because, as the commenter seemed to understand, the redistribution of costs and community support principles are applied to the FTE resident training slots of a hospital; the principles are not associated with a particular resident, to which the principles could apply differently from year to year.

Comment: One commenter disagreed with the choice of words used in the proposed definition of “redistribution of costs” at proposed § 413.86(b). As proposed, the definition states: “Redistribution of costs means an attempt by a hospital to increase the amount it is allowed to receive from Medicare under this section by counting FTE residents who were in medical residency programs where the costs of the programs had previously been incurred by the educational institution.” In particular, the commenter objected to the first part of the definition: “an attempt by a hospital to increase the amount it is allowed to receive from Medicare.” The commenter believed that the phrase was unnecessary to the definition and should be deleted.

Response: We understand the concern of the commenter. However, we have used “the attempt” language at § 413.86(b) for the proposed definition of “redistribution of costs” primarily because we have adopted the language of the existing regulation at § 413.85(c) that defines “redistribution of costs” (now applicable to costs of approved nursing and allied health education activities). The language was not intended to be offensive. Rather, we meant it to be descriptive of a possible motive for a redistribution of costs. In light of the commenter's suggestion, we are revising the language to be purely descriptive of the scenario of the redistribution and not reflect a possible motive. Accordingly, we are revising the language at § 413.86(b) to state: “Redistribution of costs” occurs when a hospital counts FTE residents in medical residency programs and the costs of the programs had previously been incurred by an educational institution. In the future, we will consider conforming changes to the definition of “redistribution of costs” at § 413.85(c) as well.

Comment: Some commenters believed that, through the enactment of the 1996 cap on the count of allopathic and osteopathic residents, Congress has already dealt with the problem that CMS is attempting to revisit with the proposed rule. The commenters believed that when Congress exempted the dental residents from the caps, it intended to create hospital incentives for dental training. The commenters believed that the CMS redistribution of costs and community support policy contradicts this Congressional intent.

Response: We do not believe that when Congress instituted the caps on the count of residents with the Balanced Budget Act of 1997, it was aware that inappropriate counting of FTE residents could occur through redistribution of costs. CMS, itself, did not become aware that many hospitals were engaging in these cost shifting arrangements, very often involving dental residents since at least October 1, 1997, when hospitals were authorized to count FTE residents for purposes of IME payments, as well as direct GME payments, for training in nonhospital sites. As we stated above, it is only since the audits by our fiscal intermediaries of the fiscal year ending 1998 and 1998 cost reports that have occurred within the last 2 years that CMS became aware that significant cost shifting was taking place. Therefore, we do not believe Congress would have been in a position to consider whether to authorize cost shifting in its 1997 legislation. Thus, we do not believe, as the commenters do, that Congress expected, or tacitly condone, cost shifting to dental residents as a result of exempting the dental residents from the 1996 caps. Rather, we believe that when Congress exempted dental residents from the 1996 caps, it intended to allow more dental training to occur in the hospital, not to authorize cost shifting from dental schools to hospitals and to the Medicare program.

Comment: One commenter asked what types of costs the hospital is required to incur for training in nonhospital sites in order for there to be no redistribution of costs or community support. Specifically, the commenter described a scenario under which a teaching hospital and a medical school are related parties and asked whether the teaching hospital is required to pay for the teaching physician services relating to offsite rotations at a medical school clinic before the FTE residents participating in the rotation can be counted for purposes of IME or direct GME payment.

Response: We understand from the scenario described by the commenter that hospital-based residents are being rotated to the medical school clinic. As such, we assume that the hospital is already incurring at least the residents' salary and fringe benefits. Therefore, when rotating the residents to the clinic, the hospital is incurring at least some of the direct GME costs of training the residents. Under these circumstances, a redistribution of costs has not taken place. However, according to the requirements for counting FTE residents in nonhospital settings under § 413.86(f)(4), among other requirements, the hospital is required to incur the portion of the teaching physicians' salaries and fringe benefits attributable to direct GME (by the term “related party,” we are assuming that the medical school clinic is not provider-based as specified under § 413.65, and therefore, is not considered part of the hospital). Thus, under the commenter's scenario, the hospital may be prohibited from counting the FTE residents, not because of redistribution of costs but because of failure to incur “all or substantially all of the cost” under § 413.86(f)(4) if the hospital is not incurring the supervisory physician's salary attributable to direct GME.

Comment: A number of commenters argued that the proposed application of the redistribution of costs and community support principles is bad public policy from the perspective of access, quality and cost-effectiveness of oral health care.

Response: We understand that dental training programs provide much needed oral health care to the American public and did not intentionally target them with our policy on redistribution of costs and community support. However, we believe much of the inappropriate cost sifting to hospitals and to the Medicare program is related to dental residency programs—which is probably due to the fact that dental residents are exempted from the statutory 1996 FTE caps. Although we regret that publication of this rule may upset some newly formed relationships between hospitals and dental schools, we continue to believe that the Medicare program should not pay for nonhospital dental residency training that had previously been funded by other sources, without any sponsorship by hospitals or the Medicare program.

Comment: One commenter stated that by establishing a PRA floor equal to 85 percent of the locality-adjusted national average PRA, Congress created an exception to the principles of community support and redistribution of costs. The commenter noted that this floor increased reimbursement to a number of teaching hospitals around the country whose own PRAs were low “precisely” because the community or another educational institution had been bearing the training costs in the GME PRA base year. Therefore, the commenter argued, the PRA floor “picked up” some of those disallowed costs, and that Medicare is, in effect, currently paying for those costs in the PRAs that were raised to the floor.

Response: The commenter is referring to section 311 of the Balanced Budget Refinement Act (BBRA) of 1999 (Pub. L. 106-113), which, for FY 2001, established a floor PRA at 70 percent of the locality-adjusted national average PRA, and to section 511 of the Benefits Improvement and Protection Act (BIPA) of 2000 (Pub. L. 106-554), which, for FY 2002, established a floor PRA at 85 percent of the locality-adjusted national average PRA. Regulations concerning these provisions are implemented at § 413.86(e)(4). These provisions were intended, in part, to narrow the disparities (both high and low) in direct GME payments to teaching hospitals across the country. One of the reasons a number of hospitals had low base year PRAs is because a significant amount of their GME costs in the PRA base year was incurred by another entity (that is, the “community”). (Variations in base year PRAs were otherwise due to differences in hospital-specific accounting practices and differences in reimbursement methods for supervising physician and resident salaries.) By providing for increased GME payments to certain hospitals with low PRAs, we do not believe Congress implicitly condoned, or made an exception to, the redistribution of costs and community support principles. We note that Congress provided for an increase to the floor PRA for all hospitals that had PRAs below the floor, not just to hospitals that, in the base year, did not incur certain GME costs. Rather, we believe Congress intended to provide increased GME payments to hospitals with low PRAs, regardless of the reasons those particular hospitals may have had low PRAs, in an attempt to even out some of the disparity in PRAs, nationally.

Comment: A commenter noted that the among the examples cited in the proposed rule at 68 FR 27213 as illustrative of inappropriate application of Medicare IME and direct GME policy, we described a situation where a hospital on the East Coast of the United States is counting dental residents training in nonhospital sites in Hawaii. The commenter believed that we have incorrect information regarding this program, and that there is, in fact, no redistribution of costs from the community to the Medicare program with respect to the program in Hawaii. Specifically, the commenter explained that in August 2002, a hospital in New York placed one dental resident in a clinic located in Honolulu. The New York hospital pays the costs of the resident's stipend and the supervising faculty's salary, and there is a written agreement between the hospital and the clinic. The commenter stated that in the future, the program anticipates placing additional residents at other nonhospital sites in Hawaii.

Response: As we stated in the preambles to the proposed rule and this final rule, there would be no redistribution of costs or community support if, from the outset of the program, a hospital incurs direct GME costs. Therefore, if, in fact, a hospital in New York has been incurring direct GME costs for a training program located in a clinic in Hawaii since the program's inception, then there would be no redistribution of costs or community support. The hospital in New York could count FTE residents training in the nonhospital site as long as the applicable requirements are met.

Comment: One commenter that described a scenario in which a university funded a family practice program for many years. However, in 2000, a Federally Qualified Health Center (FQHC) entered into a written agreement with the university and began reimbursing the university for “all or substantially all” of the costs of the program. The FQHC has been receiving Medicare direct GME payments since that time. The commenter stated that under the terms of the proposed rule, this FQHC would be ineligible for receipt of GME payments, since, prior to 2000, the program was funded exclusively by the university.

Response: The commenter raised the point that the redistribution of costs and community support principles are applicable to providers other than hospitals that may receive Medicare payments for residency training. Specifically, FQHCs and RHCs under § 405.2468, CAHs under § 413.70, and Medicare+Choice organizations (MCO) under § 422.270 may qualify to receive payments for direct GME costs. We note that the existing regulations at § 405.2468(f)(6) for FQHCs and RHCs, and at § 422.270(c) for MCOs, already clearly state that the allowable direct GME costs of these entities are subject to the redistribution of costs and community support principles in § 413.85(c). We agree with the commenter and are also clarifying the regulations at § 413.86(i) to clearly state that the principles of redistribution of costs and community support apply equally to hospitals, FQHCs, RHCs, CAHs, and MCOs. Therefore, we agree that, in the situation described by the commenter the FQHC would not be eligible for Medicare direct GME payments since the family practice program represents a redistribution of costs from the community (that is, the university) to the Medicare program (that is, the FQHC through direct GME payments).

3. Rural Track FTE Limitation for Purposes of Direct GME and IME for Urban Hospitals That Establish Separately Accredited Approved Medical Programs in a Rural Area (§§ 412.105(f)(1)(x) and 413.86(g)(12))

a. Change in the Amount of Rural Training Time Required for an Urban Hospital To Qualify for an Increase in the Rural Track FTE Limitation

To encourage the training of physicians in rural areas, section 407(c) of Pub. L. 106-113 amended sections 1886(d)(5)(B) and 1886(h)(4)(H) of the Act to add a provision that, in the case of an urban hospital that establishes separately accredited approved medical residency training programs (or rural tracks) in a rural area or has an accredited training program with an integrated rural track, an adjustment shall be made to the urban hospital's cap on the number of residents. For direct GME, the amendment applies to payments to hospitals for cost reporting periods beginning on or after April 1, 2000; for IME, the amendment applies to discharges occurring on or after April 1, 2000.

Section 407(c) of Pub. L. 106-113 did not define a “rural track” or an “integrated rural track,” nor are these terms defined elsewhere in the Act or in any applicable regulations.

Currently, there are a number of accredited 3-year primary care residency programs in which residents train for 1 year of the program at an urban hospital and are then rotated for training for the other 2 years of the 3-year program to a rural facility(ies). These separately accredited “rural track” programs are recognized by the Accreditation Council of Graduate Medical Education (ACGME) as “1-2” rural track programs. As far as CMS is able to determine, ACGME is the only accrediting body to “separately accredit” rural track residency programs, a requirement specified in Pub. L. 106-113.

We implemented the rural track program provisions of section 1886(d)(5)(B) and 1886(h)(4)(H) of the Act to address these “1-2” programs and to account for other programs that are not specifically “1-2” programs but that include rural training components. As stated above, since there is no existing definition of “rural track” or “integrated rural track,” we define at § 413.86(b) a “rural track” and an “integrated rural track” as an approved medical residency training program established by an urban hospital in which residents train for a portion of the program at the urban hospital and then rotate for a portion of the program to a rural hospital(s) or to a rural nonhospital site(s). We have previously noted that the terms “rural track” and “integrated rural track,” for purposes of this definition, are synonymous.

To implement these provisions, we revised § 413.86 to add paragraph (g)(11) (since redesignated as (g)(12)), and § 412.105 to add paragraph (f)(1)(x) to specify that, for direct GME, for cost reporting periods beginning on or after April 1, 2000, or, for IME, for discharges occurring on or after April 1, 2000, an urban hospital that establishes a new residency program, or has an existing residency program, with a rural track (or an integrated rural track) may, under certain circumstances, include in its FTE count residents in those rural tracks, in addition to the residents subject to the FTE cap at § 413.86(g)(4). (See the August 1, 2000 interim final rule with comment period (65 FR 47033) and the August 1, 2001 IPPS final rule (66 FR 39902)). These regulations specify that an urban hospital may count the residents in the rural track in excess of the hospital's FTE cap up to a “rural track FTE limitation” for that hospital. We defined this rural track FTE limitation at § 413.86(b) as the maximum number of residents (as specified in § 413.86(g)(12)) training in a rural track residency program that an urban hospital may include in its FTE count, in addition to the number of FTE residents already included in the hospital's FTE cap.

Generally, the rural track policy is divided into two categories: Rural track programs in which residents are rotated to a rural area for at least two-thirds of the duration of the program; and rural track programs in which residents are rotated to a rural area for less than two-thirds of the duration of the program. Currently, family practice is the only specialty that has separately accredited rural track programs. As previously noted, to account for other specialties that have program lengths greater than or less than 3 years, or that are not “1-2” programs, but may establish separately accredited rural track residency programs that are longer than 3 years, our regulations specify that residents must train in the rural area for “two-thirds of the duration of the program,” rather than “2 out of 3 program years,” in order for the urban hospital to count FTEs in the rural track (up to the rural track FTE limitation) in addition to the residents included in the hospital's FTE limitation. Thus, for example, under current policy, if a surgery program, which is a 5-year program, were to establish a separately accredited rural track, the urban hospital must rotate the surgery residents to the rural area for at least two-thirds of the duration of the 5-year program in order to qualify to count those FTEs in excess of the hospital's FTE cap, as provided in § 413.86(g)(12) and § 412.105(f)(1)(x).

Accordingly, our policy for determining whether an urban hospital qualifies for an adjustment to the FTE cap for training residents in rural areas is dependent upon the proportion of time the residents spend training in the rural areas. If the time spent training in rural areas (either at a rural hospital or a rural nonhospital site) constitutes at least two-thirds of the duration of the program, then the urban hospital may include the time the residents train at that urban hospital in determining GME payments. However, if the urban hospital rotates residents to rural areas for a period of time that is less than two-thirds of the duration of the program, although the rural hospital may count the time the residents train at the rural hospital if the program is new, the urban hospital may not include the time the residents train at the urban hospital for GME payment purposes (unless it can do so within the hospital's FTE cap).

When we first implemented this policy on rural tracks, it was consistent with our understanding of how the ACGME accredits rural track “1-2” programs, in which residents train for 1 year of the program at an urban hospital and are then rotated for training years 2 and 3 to a rural facility. We believed that the ACGME did not separately accredit an approved program as a rural track program unless it met this “1-2” condition; that is, the residents were spending one-third of program training in the urban area and two-thirds of the program training in the rural area. However, we have recently learned that there are a few rural track programs that are separately accredited by the ACGME as “1-2” rural track programs, but the residents in these programs are not training in rural areas for at least two-thirds of the duration of the program. We understand that in certain instances in which the case-mix of the rural facilities might not be sufficiently broad to provide the residents with an acceptable range of training opportunities, the ACGME allows the residents in program years 2 and 3 to return to the urban hospital for some training in both years. However, because the training in years 2 and 3 is predominantly occurring at the rural locations, the ACGME still separately accredits the urban and rural portions as a “1-2” program.

The existing regulations at §§ 412.105(f)(1)(x) and 413.86(g)(12) specify two main criteria for an urban hospital to count the time spent by residents training in a rural track while at the urban hospital in excess of the hospital's FTE limitation: (1) the program must be separately accredited by the ACGME; and (2) the time spent training in rural areas (either at a rural hospital or a rural nonhospital site) must constitute at least two-thirds of the duration of the program.

We believe that an urban hospital that operates a program that is separately accredited by the ACGME as a “1-2” program, but in which residents train in rural areas for more than half but less than two-thirds of the duration of the program, should still be allowed to count those FTE residents for GME payment purposes. Therefore, to be consistent with the ACGME accreditation practices, in the May 19, 2003 proposed rule, we proposed to revise our regulations. Proposed § 413.86(g)(12) still addressed our policy that an urban hospital qualifies for an adjustment to the FTE cap for training in rural areas based upon the proportion of time the residents spend training in the rural areas. However, instead of using “two-thirds” as the criterion to specify the amount of time residents training in the rural areas under regulations at §§ 413.86(g)(12)(i) through (iv) and 412.105(f)(1)(x), as under current policy, the proposal would use “one-half” as the criterion. This proposal addressed the limited cases where ACGME separately accredits programs as “1-2” rural tracks but residents in those programs train in the rural areas less than two-thirds of the time, although greater than one-half of the time. Specifically, we proposed at § 413.86(g)(12) to state:

• If an urban hospital rotates residents to a separately accredited rural track program at a rural hospital(s) for at least two-thirds of the duration of the program for cost reporting periods beginning on or after April 1, 2000 and before October 1, 2003, or for more than one-half of the duration of the program for cost reporting periods beginning on or after October 1, 2003, the urban hospital may include those residents in its FTE count for the time the rural track residents spend at the urban hospital.
• If an urban hospital rotates residents to a separately accredited rural track program at a rural nonhospital site(s) for at least two-thirds of the duration of the program for cost reporting periods beginning on or after April 1, 2000, and before October 1, 2003, or for more than one-half of the duration of the program for cost reporting periods beginning on or after October 1, 2003, the urban hospital may include those residents in its FTE count, subject to the requirements under § 413.86(f)(4).
• If an urban hospital rotates residents in the rural track program to a rural hospital(s) for less than two-thirds of the duration of the program for cost reporting periods beginning on or after April 1, 2002, and before October 1, 2003, or for one-half or less than one-half of the duration of the program for cost reporting periods beginning on or after October 1, 2003, the rural hospital may not include those residents in its FTE count (if the rural track is not a new program under § 413.86(g)(6)(iii), or if the rural hospital's FTE count exceeds that hospital's FTE cap), nor may the urban hospital include those residents when calculating its rural track FTE limitation.
• If an urban hospital rotates residents in the rural track program to a rural nonhospital site(s) for a period of time that is less than two-thirds of the duration of the program for cost reporting periods beginning on or after April 1, 2002, and before October 1, 2003, or for one-half or less than one-half of the duration of the program for cost reporting periods beginning on or after October 1, 2003, the urban hospital may include those residents in its FTE count, subject to the requirements under § 413.86(f)(4).

We also proposed to make a conforming change to § 412.105(f)(1)(x) to make these proposed provisions applicable to IME payments for discharges occurring on or after October 1, 2003.

We believe the proposal produces a more equitable result than the existing policy; the proposal encompasses what we believe to be all situations in which the ACGME separately accredits rural track programs and in which residents in the programs spend a majority of the time training in rural settings, fulfilling the intent of Congress for Medicare to provide GME payments for significant rural residency training.

Comment: Several commenters supported our proposal that, effective for cost reporting periods beginning on or after October 1, 2003, an urban hospital would be allowed to include residents in its FTE count above its FTE cap for the time that the residents train at the urban hospital, if the residents rotate to a separately accredited rural track program in a rural area for more than one-half of the duration of the program. The commenters believed that this proposed policy better reflects Congressional intent to encourage training in rural areas, while allowing residency programs the flexibility to rotate residents back to urban areas for needed clinical experiences that are not available in the rural setting.

One commenter recommended that the proposal should reduce the required rural training time even further, since research suggests that more than 50 percent of family practice residents who spend as little as 3 months training in rural areas end up practicing in rural settings.

Response: We agree with the commenters that an urban hospital that operates a program that is separately accredited by the ACGME as a “1-2” program, but in which residents train in rural areas for more than half but less than two-thirds of the duration of the program, should still be allowed to count those FTE residents for GME payment purposes. However, we do not agree that urban hospitals should be allowed to receive an increase in their FTE caps to include residents in its FTE count for the time that the residents train at the urban hospital, if the residents rotate to a rural area for one-half or less than one-half of the duration of the program. As we stated in the August 1, 2001 Federal Register (66 FR 39904-39905), we interpret section 1886(h)(4)(H)(iv) of the Act as only allowing for an urban hospital to receive an adjustment under the rural track provision if the rural track program is “separately accredited.” In order to be separately accredited as a rural track, the program must meet the ACGME's “1-2” criteria; that is, the residents are typically spending approximately two-thirds of the duration of the program in the rural area. We also explained that while we agree that post-residency retention in rural areas is important, we also believe it is important to prevent hospitals from receiving adjustments to their FTE caps in situations when only a nominal amount of training occurs in the rural area. Therefore, we are not adopting the commenter's request to allow an urban hospital to receive an increase in its FTE caps to include residents in its FTE count for the time that the residents train at the urban hospital, if the residents rotate to a rural area for one-half or less than one-half of the duration of the program.

Comment: One commenter that works for a community health center (CHC) that treats a high percentage of patients below the poverty line expressed concern about the detrimental effects that shrinking hospital revenues are having on the training of family practice residents at the CHC and at other rural and community-based settings. The commenter noted that doubling the number of CHCs is a goal of the President, and urged that, if there should be further “restraint” on teaching programs, programs that expand into CHCs should be exempt from such restrictions.

Response: We appreciate the comment. However, we note that since we did not specifically make any proposals related to residency training in community health centers, this comment is outside the scope of this final rule. Therefore, we are not responding to it at this time.

b. Inclusion of Rural Track FTE Residents in the Rolling Average Calculation

Section 1886(h)(4)(G) of the Act, as added by section 4623 of Pub. L. 105-33, provides that, for a hospital's first cost reporting period beginning on or after October 1, 1997, the hospital's FTE resident count for direct GME payment purposes equals the average of the actual FTE resident count for that cost reporting period and the preceding cost reporting period. Section 1886(h)(4)(G) of the Act requires that, for cost reporting periods beginning on or after October 1, 1998, a hospital's FTE resident count for direct GME payment purposes equals the average of the actual FTE resident count for the cost reporting period and the preceding two cost reporting periods (that is, a 3-year rolling average). This provision phases in over a 3-year period any reduction in direct GME payments to hospitals that results from a reduction in the number of FTE residents below the number allowed by the FTE cap. We first implemented this provision in the August 29, 1997 final rule with comment period (62 FR 46004) and revised § 413.86(g)(5) accordingly. Because hospitals may have two PRAs, one for residents in primary care and obstetrics and gynecology (the “primary care PRA”), and a lower PRA for nonprimary care residents, we revised our policy for computing the rolling average for direct GME payment purposes (not for IME) in the August 1, 2001 final rule (66 FR 39893) to create two separate rolling averages, one for primary care and obstetrics and gynecology residents (the “primary care rolling average”), and one for nonprimary care residents. Effective for cost reporting periods beginning on or after October 1, 2001, direct GME payments are calculated based on the sum of: (1) the product of the primary care PRA and the primary care rolling average; and (2) the product of the nonprimary care PRA and the nonprimary care FTE rolling average. (This sum is then multiplied by the Medicare patient load to determine Medicare direct GME payments).

Section 407(c) of Pub. L. 106-113, which amended sections 1886(d)(5)(B) and 1886(h)(4)(H) of the Act to create the rural track provision, provided that, in the case of an urban hospital that establishes a separately accredited rural track, “* * * the Secretary shall adjust the limitation under subparagraph (F) in an appropriate manner insofar as it applies to such programs in such rural areas in order to encourage the training of physicians in rural areas” (emphasis added). Subparagraph (F) of the Act is the provision that establishes a cap on the number of allopathic and osteopathic FTE residents that may be counted at each hospital for Medicare direct GME payment purposes. Thus, the provision authorizes the Secretary to allow for an increase to an urban hospital's FTE cap on allopathic and osteopathic residents in certain instances when an urban hospital establishes a rural track program. Although the rural track provision effectively allows an increase to the urban hospital's FTE cap by adjusting the FTE limitation under subparagraph (F), the statute makes no reference to subparagraph (G), the provision concerning the rolling average count of residents. That is, the statute does not provide for an exclusion from the rolling average for the urban hospital for those FTE residents training in a rural track.

Since we implemented this rural track provision in the August 1, 2000 interim final rule with comment period (65 FR 47033), we have interpreted this provision to mean that, except for new rural track programs begun by urban teaching hospitals that are establishing an FTE cap for the first time under § 413.86(g)(6)(i), when an urban hospital establishes a new rural track program or expands an existing rural track program, FTE residents in the rural track that are counted by the urban hospital are included in the hospital's rolling average calculation immediately. Although we have not specified in the regulations that rural track FTE residents counted by an urban hospital are included in the hospital's rolling average FTE resident count, this has been our policy. The Medicare cost report, Form CMS-2552-96 (line 3.05 on Worksheet E, Part A, for IME payments, and on line 3.02 on Worksheet E-3, Part IV, for direct GME payments), reflects this policy. Accordingly, FTE residents in a rural track program are to be included in the urban hospital's rolling average count for IME and direct GME for cost reporting periods beginning on or after April 1, 2000.

In the May 19, 2003 proposed rule, we proposed to revise the regulations at § 413.86(g)(5) to add a new paragraph (vii) to clarify that, subject to regulations at § 413.86(g)(12), except for new rural track programs begun by urban hospitals that are first establishing an FTE cap under § 413.86(g)(6)(i), when an urban hospital with an existing FTE cap establishes a new program with a rural track (or an integrated rural track), or expands an existing rural track (or an integrated rural track) program, the FTE residents in that program that are counted by the urban hospital are included in the urban hospital's rolling average FTE resident count immediately. We also proposed to revise §§ 413.86(g)(12)(i)(A), (g)(12)(ii)(B), and (g)(12)(iv)(A) to indicate that for the first 3 years of the rural track's existence, the rural track FTE limitation for each urban hospital will be the actual number of FTE residents, subject to the rolling average, training in the rural track at the urban hospital.

Comment: Commenters supported our proposal to revise § 413.86(g)(5) to clarify that the FTE residents in that program that are counted by the urban hospital are included in the urban hospital's rolling average FTE resident count immediately. The commenters stated that allowing immediate inclusion of rural track resident counts will serve to assist urban hospitals in their development of educational partnerships with rural hospitals.

Response: We appreciate the commenters support and, as explained below, are adopting revisions to the regulations concerning inclusion of rural track residents in the rolling average count of urban hospitals as final.

Except for new rural track programs begun by urban hospitals that are first establishing an FTE cap under § 413.86(g)(6)(i), or for rural hospitals that are establishing new rural track programs under § 413.86(g)(6)(iii), we are implementing sections 1886(d)(5)(B) and 1886(h)(4)(H) of the Act to require that FTE residents that are counted by an urban hospital based on the residents' participation in a rural track are included in the rolling average calculation. Accordingly, for IME and direct GME purposes, unless the rural track program is a new program under § 413.86(g)(13) and qualifies for a cap adjustment under § 413.86(g)(6)(i) or (g)(6)(iii), in instances where an urban hospital increases the number of residents it trains due to the establishment of a new or an expansion of an existing rural track program, the additional FTE residents in the rural track program are only gradually included (over a 3-year period) in the urban hospital's FTE count, since they are immediately included in the rolling average calculation of the urban hospital.

The following is an example of how residents in a rural track would be included in the rolling average calculation:

Assume that urban Hospital A, with a fiscal year end (FYE) date of June 30, had 10 unweighted FTE residents training in its cost reporting period ending June 30, 1996, thereby establishing an FTE cap of 10. Hospital A only trains primary care residents. In its cost reporting periods ending on June 30, 2002, and June 30, 2001, Hospital A again trained 10 FTE residents. However, in July 2002, Hospital A starts a rural training track program, adding 2 FTE residents. Since the additional rural track residents are included immediately in the rolling average, in FYE June 30, 2003, Hospital A's FTE residents for payment purposes equal 10.67 FTEs (12 + 10 + 10 / 3) and not 12 FTEs [(10 + 10 + 10 / 3) + 2], which would be the FTE count if FTEs in a rural track program were not subject to the rolling average calculation.

We are finalizing our proposed revision of § 413.86(g)(5) to add a new paragraph (vii) as explained above. In addition, we are finalizing our revision of §§ 413.86(g)(12)(i)(A), (g)(12)(ii)(B), and (g)(12)(iv)(A) to indicate that for the first 3 years of the rural track's existence, the rural track FTE limitation for the urban hospital will be the actual number of FTE residents, subject to the rolling average, training in the rural track at the urban hospital.

4. Technical Change Relating to Affiliated Groups and Affiliation Agreements

Section 1886(h)(4)(H)(ii) of the Act permits, but does not require, the Secretary to prescribe rules that allow institutions that are members of the same affiliated group (as defined by the Secretary) to elect to apply the FTE resident limit on an aggregate basis. This provision allows the Secretary to give hospitals flexibility in structuring rotations within a combined cap when they share a resident's time. Consistent with the broad authority conferred by the statute, we established criteria for defining an “affiliated group” and an “affiliation agreement” in both the August 29, 1997 final rule (62 FR 45965) and the May 12, 1998 final rule (63 FR 26317). We further clarified our policy concerning affiliation agreements in the August 1, 2002 final rule (67 FR 50069).

We are aware that there has been some confusion at times among members of the provider community when using the term “affiliation agreement,” since the term is used in contexts other than for Medicare GME payment purposes. For example, an “affiliation agreement” is a term historically used in the academic community that generally relates to agreements made between hospitals and medical schools or among sponsors of medical residency education programs. To help prevent further confusion, in the May 19, 2003 proposed rule, we proposed to change the term in the regulations to “Medicare GME affiliation agreement.” We believe this will help to distinguish these agreements used for purposes of GME payments from agreements used for other purposes in the provider community. We proposed to revise the regulations at § 413.86(b) to state “Medicare GME affiliated group,” and “Medicare GME affiliation agreement”. We proposed to make similar revisions to § 413.86(g)(4)(iv), (g)(7)(i) through (v), and § 412.105(f)(1)(vi) for IME payment purposes.

Comment: Commenters supported our proposal to change the terms “affiliated group” and “affiliation agreement”, as defined in § 413.86(b), to “Medicare GME affiliated group” and “Medicare GME affiliation agreement”, respectively. The commenters believed that the changes in terminology will help distinguish these terms from other affiliation agreements that are entered into by hospitals, medical schools, and other institutions that sponsor residency training.

Response: We agree with the commenters and are adopting as final the proposed changes throughout § 412.105 for IME and § 413.86 for direct GME.

Out of Scope Comments Relating to GME

Comment: Several comments addressed miscellaneous IME and direct GME issues, including the initial residency period (IRP) and volunteer physicians.

Response: Because we did not propose any changes in policy concerning these issues, we are unable to respond to these comments at this time. We will consider them for purposes of future rulemaking.

G. Updates to the Reasonable Compensation Equivalent (RCE) Limits (§ 415.70)

1. Background

Under the Medicare program, payment for services furnished by a physician is made under either the Hospital Insurance Program (Part A) or the Supplementary Medical Insurance Program (Part B), depending on the type of services furnished. In accordance with section 1848 of the Act, physicians' charges for medical or surgical services to individual Medicare patients generally are covered under Part B on a fee-for-service basis under the Medicare physician fee schedule. The compensation that physicians receive from or through a provider for services that benefit patients generally (for example, administrative services, committee work, teaching, and supervision) can be covered under Part A or Part B, depending on the provider's setting.

As required by section 1887(a)(2)(B) of the Act, allowable compensation for services furnished by physicians to providers that are paid by Medicare on a reasonable cost basis is subject to reasonable compensation equivalent (RCE) limits. Under these limits, payment is determined based on the lower of the actual cost of the services to the provider (that is, any form of compensation to the physician) or a reasonable compensation equivalent. For purposes of applying the RCE limits, physician compensation costs means monetary payments, fringe benefits, deferred compensation and any other items of value (excluding office space or billing and collection services) that a provider or other organization furnishes a physician in return for the physician's services.

The RCE limits do not apply to the costs of physician compensation that are attributable to furnishing inpatient hospital services paid under the IPPS or as GME costs. In addition, RCE limits do not apply to the costs CAHs incur in compensating physicians for services. Furthermore, compensation that a physician receives for activities that may not be paid under either Part A or Part B is not considered in applying the RCE limits.

The limits apply equally to all physician services to providers that are payable on a reasonable cost basis under Medicare. If a physician receives any compensation from a provider for his or her physician services to the provider (that is, those services that benefit patients generally), payment to those affected providers for the costs of such compensation is subject to the RCE limits. The RCE limits are not applied to payment for services that are identifiable medical or surgical services to individual patients and paid under the physician fee schedule, even if the physician agrees to accept compensation (for example, from a hospital) for those services. (However, payments to teaching hospitals that have elected to be paid for these services on a reasonable cost basis in accordance with section 1861(b)(7) of the Act are subject to the limits.)

Section 415.70(b) of the regulations specifies the methodology for determining annual RCE limits, considering average physician incomes by specialty and type of location, to the extent possible using the best available data. On October 31, 1997, the revised RCE limits update methodology was published in the Federal Register (62 FR 59075). For cost reporting periods beginning on or after January 1, 1998, updates to the RCE limits are calculated using the Medicare Economic Index (MEI). The inflation factor used to develop the initial RCE limits and, subsequently, to update those limits to reflect increases in net physician compensation was the Consumer Price Index for All Urban Consumers (CPI-U). In 1998, we revised the update methodology for the RCE limits by replacing the CPI-U with the inflation factor for the physician fee schedule (the MEI) to achieve a measure of consistency in the methodologies employed to determine reasonable payments to physicians for direct medical and surgical services furnished to individual patients and reasonable compensation levels for physicians' services that benefit provider patients generally.

2. Updated RCE Limits

In the May 19, 2003 proposed rule, we indicated our intent to publish updated payment limits on the amount of allowable compensation for services furnished by physicians to providers in this FY 2004 IPPS final rule. These revised RCE limits are based on updated economic index data and replace the limits that were published in the Federal Register on May 5, 1997 (62 FR 24483). We calculated the revised RCE limits by using the methodology published in the Federal Register on October 31, 1997 (62 FR 59075). These limits are specified in the chart below and are effective for cost reporting periods beginning on or after January 1, 2004.

The revised RCE limits are mere updates that have been calculated by applying the most recent economic index data. In the proposed rule, we did not propose to change the methodology used to determine the limits. We indicated that, in accordance with § 415.70(f), we are allowed to publish the revised RCE limits in a final rule without prior publication of a proposed rule for public comment. Furthermore, indicated our belief that publication of the revised RCE limits in a proposed rule with opportunity for public comment was unnecessary, and that we found good cause to waive the procedure.

Comment: One commenter was encouraged to learn of our proposal to publish updated RCE limits and suggested that these updates occur on an annual basis.

Response: We will continue to review the RCE limits on a regular basis by applying the most recent economic index data and publish updates as necessary.

3. Application of RCE Limits

This section, as well as the two following sections, is not describing new policy, but rather is simply a discussion of a continuation of the existing policies with respect to the application of and exceptions to the RCE limits and the geographic area classifications used for purposes of establishing the RCE limits. We will continue to use the RCE limits to compute Medicare payments when a physician is compensated by a provider that is subject to the RCE limits in some or all of its areas. We also will use these limits when the physician is compensated by any other related organization for physician administrative, supervisory, and other provider services paid under Medicare. In applying the RCE limits, the intermediary will assign each compensated physician to the most appropriate specialty category. If no specialty category is appropriate (for example, in determining the reasonable cost for an emergency room physician), the fiscal intermediary will use the RCE level for the “Total” category, which is based on income data for all physicians. The fiscal intermediary will determine the appropriate geographic area classification given in Table 9 of the addendum of this final rule.

If the physician's contractual compensation covers all duties, activities, and services furnished to the provider and to its patients and the physician is employed full-time, the appropriate specialty compensation limit will be used and adjusted by the physician's allocation agreement to arrive at the program's share of allowable costs as physician compensation costs. In the absence of an allocation agreement, we generally will assume that 100 percent of the compensation was related to services paid under the physician fee schedule and that there are no allowable costs for the physician's services to the provider.

If a physician's compensation from the provider represents payment only for services that benefit patients generally (that is, the physician bills fees for all services furnished to individual patients), the appropriate specialty compensation limit will be used. If a physician is employed by a provider to furnish services of general benefit to patients on other than a full-time basis, the RCE amount will be adjusted upward or downward to reflect the percentage of time his or her actual hours related to a full work year of 2,080 hours.

4. Exceptions to the RCE Limits

Some providers, particularly but not exclusively small or rural hospitals, may be unable to recruit or maintain an adequate number of physicians at a compensation level within the prescribed limits. In accordance with section 1887(a)(2)(C) of the Act, if a provider is able to demonstrate to the intermediary its inability to recruit or maintain physicians at a compensation level allowable under the RCE limits (as documented, for example, by unsuccessful advertising through national medical or health care publications), the intermediary may grant an exception to the RCE limits established under these rules.

5. Geographic Area Classifications for RCE Limits

We adjust the RCE limits to account for differences in salary levels by location as well as by specialty. Under our methodology for establishing limits, and in the limits set forth below, we have classified geographic areas into three types: nonmetropolitan areas, metropolitan areas less than 1 million, and metropolitan areas greater than 1 million.

As we do for purposes of the IPPS and the physician fee schedule, we use the most current MSA designations for purposes of establishing the RCE limits. In New England, we use the NECMAs for this purpose. Tables 4A and 4B of the Addendum to this final rule includes information that identifies, by type of location (urban and rural), the geographic areas affected; that is, they list all MSAs and their constituent counties and identifies whether their population are classified as large urban. Any county not listed in the tables and all other affected U.S. possessions and territories not part of a State are considered rural areas. This information will enable providers, physicians, Medicare fiscal intermediaries, and other members of the public to determine which RCE limit level will apply in specific areas.

V. PPS for Capital-Related Costs

In the May 19, 2003 proposed rule, we did not propose any changes in the policies governing the determination of the payment rates for capital-related costs for short-term acute care hospitals under the IPPS. However, for the readers' benefit, in this section of this final rule, we are providing a summary of the statutory basis for the PPS for hospital capital-related costs, the methodology used to determine capital-related payments to hospitals, and a brief description of the payment policies under the PPS for capital-related costs for new hospitals, extraordinary circumstances, and exception (regular and special) payments. (Refer to the August 1, 2001 IPPS final rule (66 FR 39910) for a more detailed discussion of the statutory basis for the system, the development and evolution of the system, the methodology used to determine capital-related payments to hospitals both during and after the transition period, and the policy for providing regular and special exceptions payments.)

Section 1886(g) of the Act requires the Secretary to pay for the capital-related costs of inpatient hospital services “in accordance with a PPS established by the Secretary.” Under the statute, the Secretary has broad authority in establishing and implementing the PPS for capital related costs. We initially implemented the capital PPS in the August 30, 1991 IPPS final rule (56 FR 43358), in which we established a 10-year transition period to change the payment methodology for Medicare hospital inpatient capital-related costs from a reasonable cost-based methodology to a prospective methodology (based fully on the Federal rate).

Federal fiscal year (FY) 2001 was the last year of the 10-year transition period established to phase in the PPS for hospital inpatient capital-related costs. Beginning in FY 2002, capital PPS payments are based solely on the Federal rate for the vast majority of hospitals. The basic methodology for determining capital prospective payments based on the Federal rate is set forth in § 412.312. For the purpose of calculating payments for each discharge, the standard Federal rate is adjusted as follows: (Standard Federal Rate) × (DRG Weight) × (Geographic Adjustment Factor (GAF)) × (Large Urban Add-on, if applicable) × (COLA Adjustment for hospitals located in Alaska and Hawaii) × (1 + DSH Adjustment Factor + IME Adjustment Factor, if applicable) Hospitals also may receive outlier payments for those cases that qualify under the thresholds established for each fiscal year that are specified in § 412.312(c) of existing regulations.

During the 10-year transition period, a new hospital (as defined at 412.300(b)) was exempt from the capital PPS for its first 2 years of operation and was paid 85 percent of its reasonable costs during that period. Originally, this provision was effective only through the transition period and, therefore, ended with cost reporting periods beginning in FY 2002. As we discussed in the August 1, 2002 final rule (67 FR 50101), this payment provision was implemented to provide special protection to new hospitals during the transition period in response to concerns that prospective payments under a DRG system may not be adequate initially to cover the capital costs of newly built hospitals. Therefore, we believe that the rationale for this policy applies to new hospitals after the transition period as well, and in that same final rule, we established regulations under § 412.304(c)(2) that provide the same special payment to new hospitals for cost reporting periods beginning on or after October 1, 2002. Therefore, a new hospital, defined under § 412.300(b), is paid 85 percent of its allowable Medicare inpatient hospital capital-related costs through its first 2 years of operation unless the new hospital elects to receive fully prospective payment based on 100 percent of the Federal rate. (For more detailed information regarding this policy, see the August 1, 2002 IPPS final rule (67 FR 50101).)

Regulations at § 412.348(f) provide that a hospital may request an additional payment if the hospital incurs unanticipated capital expenditures in excess of $5 million due to extraordinary circumstances beyond the hospital's control. This policy was established for hospitals during the 10-year transition period, but we established regulations at § 412.312(e) to specify that payments for extraordinary circumstances are also made for cost reporting periods after the transition period (that is, cost reporting periods beginning on or after October 1, 2001). (For more detailed information regarding this policy, refer to the August 1, 2002 Federal Register (67 FR 50102).)

During the transition period, under §§ 412.348(b) through (e), eligible hospitals could receive regular exception payments. These exception payments guaranteed a hospital a minimum payment of a percentage of its Medicare allowable capital-related costs depending on the class of hospital (§ 412.348(c)). However, after the end of the transition period, eligible hospitals can receive additional payments under the special exceptions provisions at § 412.348(g), which guarantees an eligible hospital a minimum payment of 70 percent of its Medicare allowable capital-related costs. Special exceptions payments may be made only for the 10 years after the cost reporting year in which the hospital completes its qualifying project, which can be no later than the hospital's cost reporting period beginning before October 1, 2001. Thus, an eligible hospital may receive special exceptions payments for up to 10 years beyond the end of the capital PPS transition period. Hospitals eligible for special exceptions payments were required to submit documentation to the intermediary indicating the completion date of their project. (For more detailed information regarding the special exceptions policy under § 412.348(g), refer to the August 1, 2001 IPPS final rule (66 FR 39911 through 39914) and the August 1, 2002 IPPS final rule (67 FR 50102).)

VI. Changes for Hospitals and Hospital Units Excluded From the IPPS

A. Payments to Excluded Hospitals and Hospital Units (§§ 413.40(c), (d), and (f))

1. Payments to Existing Excluded Hospitals and Hospital Units

Section 1886(b)(3)(H) of the Act (as amended by section 4414 of Pub. L. 105-33) established caps on the target amounts for certain existing hospitals and hospital units excluded from the IPPS for cost reporting periods beginning on or after October 1, 1997 through September 30, 2002. For this period, the caps on the target amounts apply to the following three classes of excluded hospitals or units: psychiatric hospitals and units, rehabilitation hospitals and units, and LTCHs.

In accordance with section 1886(b)(3)(H)(i) of the Act and effective for cost reporting periods beginning on or after October 1, 2002, payments to these classes of existing excluded hospitals or hospital units are no longer subject to caps on the target amounts. In accordance with existing §§ 413.40(c)(4)(ii) and (d)(1)(i) and (ii), where applicable, excluded psychiatric hospitals and units continue to be paid on a reasonable cost basis, and payments are based on their Medicare inpatient operating costs, not to exceed the ceiling. The ceiling would be computed using the hospital's or unit's target amount from the previous cost reporting period, updated by the rate-of-increase specified in § 413.40(c)(3)(viii) of the regulations, and then multiplying this figure by the number of Medicare discharges. Effective for cost reporting periods beginning on or after October 1, 2002, rehabilitation hospitals and units are paid 100 percent of the Federal rate. Effective for cost reporting periods beginning on or after October 1, 2002, LTCHs also are no longer paid on a reasonable cost basis but are paid under a DRG-based PPS. As part of the PPS for LTCHs, we established a 5-year transition period from reasonable cost-based reimbursement to a fully Federal PPS. However, a LTCH, subject to the blend methodology, may elect to be paid based on a 100 percent of the Federal prospective rate. (Sections VI.A.3. and 4. of this preamble contain a more detailed discussion of the IRF PPS and the LTCH PPS.)

2. Updated Caps for New Excluded Hospitals and Units

Section 1886(b)(7) of the Act establishes a payment limitation for new psychiatric hospitals and units, new rehabilitation hospitals and units, and new LTCHs. A discussion of how the payment limitation was calculated can be found in the August 29, 1997 final rule with comment period (62 FR 46019); the May 12, 1998 final rule (63 FR 26344); the July 31, 1998 final rule (63 FR 41000); and the July 30, 1999 final rule (64 FR 41529). Under the statute, a “new” hospital or unit is a hospital or unit that falls within one of the three classes of hospitals or units (psychiatric, rehabilitation or long-term care) that first receives payment as a hospital or unit excluded from the IPPS on or after October 1, 1997.

The amount of payment for a “new” psychiatric hospital or unit would be determined as follows:

• Under existing § 413.40(f)(2)(ii), for the first two 12-month cost reporting periods, the amount of payment is the lesser of: (1) the operating costs per case; or (2) 110 percent of the national median (as estimated by the Secretary) of the target amounts for the same class of hospital or unit for cost reporting periods ending during FY 1996, updated by the hospital market basket increase percentage to the fiscal year in which the hospital or unit first receives payments under section 1886 of the Act, as adjusted for differences in area wage levels.
• Under existing § 413.40(c)(4)(v), for cost reporting periods following the hospital's or unit's first two 12-month cost reporting periods, the target amount is equal to the amount determined under section 1886(b)(7)(A)(i) of the Act for the third period, updated by the applicable hospital market basket increase percentage.

The amounts included in the following table reflect the updated 110 percent of the national median target amounts of new excluded psychiatric hospitals and units for cost reporting periods beginning during FY 2004. These figures are updated with the most recent data available to reflect the projected market basket increase percentage of 3.4 percent. This percentage change in the market basket reflects the average change in the price of goods and services purchased by hospitals to furnish inpatient hospital services (as projected by the Office of the Actuary of CMS based on its historical experience with the IPPS). For a new provider, the labor-related share of the target amount is multiplied by the appropriate geographic area wage index, without regard to IPPS reclassifications, and added to the nonlabor-related share in order to determine the per case limit on payment under the statutory payment methodology for new providers.

Effective for cost reporting periods beginning on or after October 1, 2002, this payment limitation is no longer applicable to new LTCHs because they are paid 100 percent of the Federal rate. Under the LTCH PPS, a new LTCH is defined as a provider of inpatient hospital services that meets the qualifying criteria for LTCHs specified under § 412.23(e)(1) and (e)(2) and whose first cost reporting period as a LTCH begins on or after October 1, 2002 (§ 412.23(e)(4)). (We note that this definition of new LTCHs should not be confused with those LTCHs first paid under the TEFRA payment system for discharges occurring on or after October 1, 1997, and before October 1, 2002.) New LTCHs are paid based on 100 percent of the fully Federal prospective rate (they may not participate in the 5-year transition from cost-based reimbursement to prospective payment). In contrast, those “new” LTCHs that meet the definition of “new” under § 413.40(f)(2)(ii) and that have their first cost reporting periods beginning on or after October 1, 1997, and before October 1, 2002, may be paid under the LTCH PPS transition methodology. Since those hospitals by definition would have been considered new before October 1, 2002, they would have been subject to the updated payment limitation on new hospitals that was published in the FY 2003 IPPS final rule (67 FR 50103). Under § 413.40(f)(2)(ii), the “new” hospital would be subject to the same cap in its second cost reporting period; this cap would not be updated for the new hospital's second cost reporting year. Thus, because the same cap is to be used for the new LTCH's first two cost reporting periods, it is no longer necessary to publish an updated cap for new LTCHs.

Effective for cost reporting periods beginning on or after October 1, 2002, this payment limitation is no longer applicable to new rehabilitation hospitals and units because they are paid 100 percent of the Federal prospective rate under the IRF PPS. Therefore, it is also no longer necessary to update the payment limitation for new rehabilitation hospitals or units.

3. Implementation of a PPS for IRFs

Section 1886(j) of the Act, as added by section 4421(a) of Pub. L. 105-33, provided the phase-in of a case-mix adjusted PPS for inpatient hospital services furnished by a rehabilitation hospital or a rehabilitation hospital unit (referred to in the statute as rehabilitation facilities) for cost reporting periods beginning on or after October 1, 2000, and before October 1, 2002, with a fully implemented PPS for cost reporting periods beginning on or after October 1, 2002. Section 1886(j) of the Act was amended by section 125 of Pub. L. 106-113 to require the Secretary to use a discharge as the payment unit under the PPS for inpatient hospital services furnished by rehabilitation facilities and to establish classes of patient discharges by functional-related groups. Section 305 of Pub. L. 106-554 further amended section 1886(j) of the Act to allow rehabilitation facilities, subject to the blend methodology, to elect to be paid the full Federal prospective payment rather than the transitional period payments specified in the Act.

On August 7, 2001, we issued a final rule in the Federal Register (66 FR 41316) establishing the PPS for inpatient rehabilitation facilities, effective for cost reporting periods beginning on or after January 1, 2002. Under the IRF PPS, for cost reporting periods beginning on or after January 1, 2002, and before October 1, 2002, payment consisted of 331/3 percent of the facility-specific payment amount (based on the reasonable cost-based reimbursement methodology) and 662/3 percent of the adjusted Federal prospective payment. For cost reporting periods beginning on or after October 1, 2002, payments are based entirely on the Federal prospective payment rate determined under the IRF PPS. We plan to issue in the Federal Register by August 1, 2003 a final rule that will update the payment rates under the IRF PPS for FY 2004, to be effective for discharges occurring on or after October 1, 2003 and before October 1, 2004.

4. Development of a PPS for Inpatient Psychiatric Facilities

We are in the process of developing a proposed rule that would establish a per diem PPS for inpatient psychiatric facilities (IPFs) (previously referred to as psychiatric hospitals and units) that is required under the provisions of section 124 of Pub. L. 106.113.

5. Implementation of a PPS for LTCHs

In accordance with the requirements of section 123 of Pub. L. 106-113, as modified by section 307(b) of Pub. L. 106-554, we established a per discharge, DRG-based PPS for LTCHs as described in section 1886(d)(1)(B)(iv) of the Act for cost reporting periods beginning on or after October 1, 2002, in a final rule issued on August 30, 2002 (67 FR 55954). The LTCH PPS uses information from LTCH hospital patient records to classify patients into distinct LTC-DRGs based on clinical characteristics and expected resource needs. Separate payments are calculated for each LTC-DRG with additional adjustments applied.

As part of the implementation of the system, we established a 5-year transition period from reasonable cost-based reimbursement to the fully Federal prospective rate. A blend of the reasonable cost-based reimbursement percentage and the prospective payment Federal rate percentage would be used to determine a LTCH's total payment under the LTCH PPS during the transition period. Certain LTCHs may elect to be paid based on 100 percent of the Federal prospective rate. All LTCHs will be paid under the fully Federal prospective rate for cost reporting periods beginning on or after October 1, 2006.

We published in the Federal Register on June 6, 2003 a final rule (68 FR 34122) that updated the payment rates for the LTCH PPS and made policy changes effective for a new LTCH PPS rate year of July l, 2003 through June 30, 2004.

6. Report of Adjustment (Exception) Payments

Section 4419(b) of Pub. L. 105-33 requires the Secretary to publish annually in the Federal Register a report describing the total amount of adjustment (exception) payments made to excluded hospitals and units, by reason of section 1886(b)(4) of the Act, during the previous fiscal year. However, the data on adjustment payments made during the previous fiscal year are not available in time to publish a report describing the total amount of adjustment payments made to all excluded hospitals and units.

The process of requesting, adjudicating, and awarding an adjustment payment is likely to occur over a 2-year period or longer. First, an excluded hospital or unit must file its cost report for a fiscal year with its intermediary within 5 months after the close of its cost reporting period. The fiscal intermediary then reviews the cost report and issues a Notice of Program Reimbursement (NPR) within approximately 2 months after the filing of the cost report. If the hospital's operating costs are in excess of the ceiling, the hospital may file a request for an adjustment payment within 6 months from the date of the NPR. The intermediary, or CMS, depending on the type of adjustment requested, then reviews the request and determines if an adjustment payment is warranted. This determination is often not made until more than 6 months after the date the request is filed. Therefore, it is not possible to provide data in this final rule. However, in an attempt to provide interested parties with data on the most recent adjustments for which we do have data, we are publishing data on adjustments that were processed by the fiscal intermediary or CMS during FY 2002.

The table below includes the most recent data available from the fiscal intermediaries and CMS on adjustment payments that were adjudicated during FY 2002. As indicated above, the adjustments made during FY 2002 only pertain to cost reporting periods ending in years prior to FY 2001. Total adjustment payments awarded to excluded hospitals and units during FY 2002 are $8,541,349. The table depicts for each class of hospital, in the aggregate, the number of adjustment requests adjudicated, the excess operating cost over ceiling, and the amount of the adjustment payment.

B. Payment for Services Furnished at Hospitals-Within-Hospitals and Satellite Facilities

Existing regulations at § 412.22(e) define a hospital-within-a-hospital as a hospital that occupies space in the same building as another hospital, or in one or more entire buildings located on the same campus as buildings used by another hospital. Moreover, existing § 412.22(f) provides for the grandfathering of hospitals-within-hospitals that were in existence on or before September 30, 1995.

Sections 412.22(h) and 412.25(e), relating to satellites of hospitals and hospital units, respectively, excluded from the IPPS, define a satellite facility as a part of a hospital or unit that provides inpatient services in a building also used by another hospital, or in one or more entire buildings located on the same campus as buildings used by another hospital. Sections 412.22(h)(3) and 412.25(e)(3) provide for the grandfathering of excluded hospitals and units that were structured as satellite facilities on September 30, 1999, to the extent they operate under the same terms and conditions in effect on that date.

In providing for the grandfathering of satellite facilities of hospitals and hospital units, we believed it was appropriate to require that the satellite facilities operate under the same terms and conditions that were in effect on September 30, 1999. There are similarities between the definition of satellite facilities and the definition of hospitals-within-hospitals (that is, hospitals-within-hospitals and satellite facilities are both physically located in acute care hospitals that are paid for their inpatient services on a prospective payment basis). Also, satellite facilities of both excluded hospitals and hospital units and hospitals-within-hospitals provide inpatient hospital services that are paid at a higher rate than would apply if the facilities were treated by Medicare as part of an acute care hospital.

In the May 19, 2003 proposed rule, we proposed to revise § 412.22(f) to specify that, effective with cost reporting periods beginning on or after October 1, 2003, a hospital operating as a hospital-within-a-hospital on or before September 30, 1995, is exempt from the criteria in § 412.22(e)(1) through (e)(5) only if the hospital-within-a-hospital continues to operate under the same terms and conditions in effect as of September 30, 1995. The intent of the grandfathering provision was to ensure that hospitals that had been in existence prior to the effective date of our hospital-within-hospital requirements should not be adversely affected by those requirements. To the extent hospitals were already operating as hospitals-within-hospitals without meeting those requirements, we believe it is appropriate to limit the grandfathering provision to those hospitals that continue to operate in the same manner as they had operated prior to the effective date of those rules. However, if a hospital changes the way it operates (for example, adds more beds) subsequent to the effective date of the new rules, it should no longer receive the benefit of the grandfathering provision.

Under § 412.22(e), we specify the criteria that a hospital-within-a-hospital is required to meet in order to be excluded from the IPPS. One of these criteria, under § 412.22(e)(5)(i), requires that a hospital-within-a-hospital is able to perform basic hospital functions (for example, medical record services and nursing services) that are presently included in the Medicare hospital conditions of participation under Part 482 of the Medicare regulations. These requirements were first included in Part 412 in response to hospitals organizing themselves as what is referred to as the hospital-within-a-hospital model. Thus, to avoid recognizing nominal hospitals, while allowing hospitals adequate flexibility and opportunity for legitimate networking and sharing of services, we included, by reference, certain hospital conditions of participation as additional criteria in Part 412 for hospitals-within-hospitals that request exclusion from the IPPS. (Further discussion can be found in a final rule published in the Federal Register on September 1, 1994 (59 FR 45389).) Modifications to the conditions of participation have been made since the publication of that September 1, 1994 final rule. Thus, we need to update the references to the conditions of participation in § 412.22(e)(5)(i) to make them consistent with existing provisions under the basic hospital conditions of participation. Therefore, we also proposed to amend § 412.22(e)(5)(i) to add references to § 482.43 (discharge planning) and § 482.45 (organ, tissue, and eye procurement) as basic hospital functions that a hospital-within-a-hospital would also be required to meet.

Comment: Several commenters disagreed with our proposal to require grandfathered hospitals-within-hospitals to continue to operate under the same terms and conditions that were in place on September 30, 1995 (for example, adding beds). These commenters believed that the adoption of this proposal could result in a decertification of a number of LTCHs, thus depriving Medicare beneficiaries of specialized services and unique programs. They asserted that CMS is requiring these grandfathered hospitals-within-hospitals to either reverse their previously approved changes or lose their certification, which would retroactively reverse prior governmental approvals of LTCH changes. The commenters further asserted that there is no good reason to treat these hospitals any differently from other providers participating in the Medicare program, a practice that the commenters believed would result in inequitable treatment of patients as well as employees. Furthermore, the commenters expressed concern that the proposed effective date timeframe for implementation (that is, 60 days) is too short for purposes of implementing this proposed change because it would not allow adequate time for providers to undo previous changes.

Response: We have reviewed the commenters' concerns with regard to our proposal to require “grandfathered” hospitals-within-hospitals to continue to operate under the same terms and conditions that were in place on September 30, 1995. We understand the commenters' concern that adoption of this change as proposed could adversely impact some grandfathered hospitals-within-hospitals that, over the years, have made changes to the terms and conditions under which they operate.

After careful consideration of the comments, we have decided to revise § 412.22(f) to state that if a hospital-within-a-hospital was excluded from the IPPS under the provisions of § 412.22(f) on or before September 30, 1995, and at that time occupied space in a building also used by another hospital or in one or more buildings located on the same campus as buildings used by another hospital, the provisions of § 412.22(e) do not apply to the hospital as long as the hospital meets either of two conditions: First, under § 412.22(f)(1), the hospital continues to operate under the same terms and conditions, including the number of beds and square footage considered to be part of the hospital for purposes of Medicare participation and payment, in effect on September 30, 1995. Second, under § 412.22(f)(2) a hospital that changed the terms and conditions under which it operates after September 30, 1995 but before October 1, 2003, may continue in its grandfathered status if it continues to operate under the same terms and conditions, including the number of beds and square footage considered to be part of the hospital for purposes of Medicare participation and payment, in effect on September 30, 2003. The second condition was added in recognition of commenters who suggested that hospitals be held harmless for past changes in their terms and conditions of operation. We note that any changes occurring on or after October 1, 2003, including changes in number of beds or square footage, could lead to a loss of grandfathered status.

We want to reiterate that, in establishing grandfathering provisions, our general intent has been to protect existing hospitals from the potentially adverse impact of recent, more specific regulations that we now believe to be essential to the goals of the Medicare program. However, a hospital that continues to be excluded from the IPPS through grandfathered status may wish to alter the terms and conditions that were in effect either on September 30, 1995, or after October 1, 2003, as provided in revised § 412.22(h). In that circumstance, in order to continue being paid as a hospital excluded from the IPPS, the hospital would need to comply with the general hospital-within-a-hospital requirements set forth in § 412.22(e).

We plan to review the issue of whether further revisions to this regulation should be made to allow more changes in operation by grandfathered hospital-within-hospitals, and welcome specific suggestions on this issue.

C. Clarification of Classification Requirements for LTCHs

Under § 412.23(e)(2), to qualify to be excluded from the IPPS as a LTCH and to be paid under the LTCH PPS, a hospital must have an average Medicare length of stay of greater than 25 days (which includes all covered and noncovered days of stay for Medicare patients) as calculated under the criteria of § 412.23(e)(3). In calculating this average Medicare inpatient length of stay, data from the hospital's most recently filed cost report are used to make this determination. However, if the hospital has not yet filed a cost report or if there is an indication that the most recently filed cost report does not accurately reflect the hospital's current Medicare average length of stay, data from the most recent 6-month period are used.

Our interpretation of § 412.23(e)(3)(ii) and (e)(3)(iii) was to allow hospitals that submit data for purposes of exclusion from the IPPS to use a period of at least 5 months of the most recent data from the preceding 6-month period. This longstanding policy interpretation was necessary in order to comply with the time requirement in § 412.22(d) that specifies that, for purposes of the IPPS, status is determined at the beginning of each cost reporting period and is effective for the entire cost reporting period. Therefore, in the May 19, 2003 proposed rule, we proposed to revise §§ 412.23(e)(3)(ii) and (iii) to reflect our longstanding interpretation of the regulations.

Comment: One commenter suggested that we clarify the source of our data for computing the average length of stay for purposes of designation as a LTCH.

Response: Although we did not propose any policy change regarding the average length of stay calculation, we did describe the data source for this calculation, which is set forth at § 412.23(e)(3). Therefore, we will take this opportunity to correct an inadvertent misstatement of the data source for this calculation and clarify present data collection procedures. In the proposed rule, we stated that we relied on data from a “. . . hospital's most recently filed cost report . . .” for determining whether it qualified as a LTCH. However, the regulation does not specify or require that the hospital's cost report (Hospital and Hospital Health Care Complex Cost Report, CMS Form 2552-96) be the source of these data used in the determination for LTCH classification. Specifically, the regulation only notes that the calculation requires dividing the total Medicare inpatient days by the total number of Medicare discharges occurring for the hospital's most recent complete cost reporting period (§ 412.23(e)(3)). (A detailed description of the designation process is included in the August 30, 2002 IPPS final rule (67 FR 55970 through 55974).)

Prior to the October 1, 2002 implementation of the LTCH prospective payment system, we did rely on data from the most recently submitted cost report for this purpose. In addition, the calculation, for purposes of qualifying as a LTCH, was based on total days and discharges for all LTCH inpatients. However, with the implementation of the LTCH PPS, we revised § 412.23(e)(3)(i) to only count total days and discharges for Medicare inpatients (67 FR 55970, August 30, 2002). Presently, we are unable to capture these data on our present cost reporting forms. Therefore, until the cost reporting form is revised, for purposes of the average length of stay calculation, we will be relying upon patient census data extracted from MedPAR files that reflect each LTCH's cost reporting period. Fiscal intermediaries and LTCHs have been informed of this course of action through official agency transmittals, but we want to emphasize that this temporary shift in data sources should have no effect on the evaluation policy set forth in regulations at §§ 412.22(d) and 412.23(e)(3) and the procedures described in the August 30, 2002 final rule.

D. Criteria for Payment on a Reasonable Cost Basis for Clinical Diagnostic Laboratory Services Performed by CAHs

Section 1820 of the Act provides for the establishment of Medicare Rural Hospital Flexibility Programs, under which individual States may designate certain facilities as critical access hospitals (CAHs). Facilities that are so designated and meet the CAH conditions of participation in 42 CFR Part 485, Subpart F, will be certified as CAHs by CMS. Section 1834(g) of the Act states that the amount of payment for outpatient services furnished by a CAH will be the reasonable costs of the CAH in providing these services.

Regulations implementing section 1834(g) of the Act are set forth at § 413.70. These regulations state, in paragraph (b)(2)(iii), that payment to a CAH for outpatient clinical diagnostic laboratory tests will be made on a reasonable cost basis only if the individuals for whom the tests are performed are outpatients of the CAH, as defined in § 410.2, at the time the specimens are collected. The regulations also state that clinical diagnostic laboratory tests for persons who are not patients of the CAH at the time the specimens are collected will be paid for in accordance with the provisions of sections 1833(a)(1)(D) and 1833(a)(2)(D) of the Act. These provisions, which also are the basis for payment for clinical diagnostic laboratory tests performed by independent laboratories and by hospitals on specimens drawn at other locations, set payment at the least of: (1) charges determined under the fee schedule as set forth in section 1833(h)(1) or section 1834(d)(1) of the Act; (2) the limitation amount for that test determined under section 1833(h)(4)(B) of the Act; or (3) a negotiated rate established under section 1833(h)(6) of the Act. Payments determined under this methodology are typically referred to as “fee schedule payments,” and are so described here both for ease of reference and to differentiate them from payments determined on a reasonable cost basis.

The definition of an “outpatient” in § 410.2 states that an outpatient means a person who has not been admitted as an inpatient but who is registered on hospital or CAH records as an outpatient and receives services (rather than supplies alone) directly from the hospital or CAH.

Recently, we have received numerous questions about how Medicare pays for laboratory services that a CAH may furnish to Medicare beneficiaries in various settings other than the CAH. Specifically, the questioners have asked whether a CAH may obtain reasonable cost payment for such services to individuals in other locations by sending a CAH employee into the setting and registering the individual as a CAH patient while the blood is drawn or other specimen collection is accomplished. The settings that have been referred to most frequently are: (1) a rural health clinic (RHC), especially one that is provider-based with respect to the CAH; (2) the individual's home; and (3) an SNF.

We have considered these suggestions and understand the position taken by those who believe that nominal compliance with the requirements for outpatient status should be enough to warrant reasonable cost payment for clinical diagnostic laboratory tests for individuals at locations outside the CAH. However, we do not agree that providing reasonable cost payment under these circumstances would be appropriate. On the contrary, we believe that extending reasonable cost payment for services furnished to individuals who are not at the CAH when the specimen is drawn would duplicate existing coverage, create confusion for beneficiaries and others by blurring the distinction between CAHs and other providers, such as SNFs and HHAs, and increase the costs of care to Medicare patients without enhancing either the quality or the availability of that care.

To clarify our policies in this area and avoid possible misunderstandings about the scope of the CAH benefit, in the May 19, 2003 proposed rule, we proposed to revise § 413.70(b)(2)(iii) to state that payment to a CAH for outpatient clinical diagnostic laboratory tests will be made on a reasonable cost basis only if the individuals for whom the tests are performed are outpatients of the CAH, as defined in § 410.2, “and are physically present in the CAH” at the time the specimens are collected. (We note that, in some cases, the CAH outpatients from whom specimens are collected at the CAH may include individuals referred to the CAH from RHCs or other facilities to receive the tests.) We proposed to further revise this paragraph to state that clinical diagnostic laboratory tests for individuals who do not meet these criteria but meet other applicable requirements will be paid for only in accordance with the provisions of sections 1833(a)(1)(D) and 1833(a)(2)(D) of the Act, that is, payment will be made only on a fee schedule basis. We emphasize that the second proposal does not mean that no payment would be made for clinical diagnostic laboratory tests performed by CAHs that do not meet the revised criteria. On the contrary, such tests would be paid, but on a fee schedule basis. We believe these clarifications are appropriate, as the CAH is not providing CAH services but is acting as an independent laboratory in providing these clinical diagnostic laboratory tests.

Comment: Some commenters stated that a major goal of the Medicare Rural Hospital Flexibility Program, under which reasonable cost payment to CAHs is authorized, is to ensure that isolated rural hospitals have access to critical health care services. The commenters believed that our proposal would undermine that goal by paying less than reasonable cost amounts for certain services. These commenters stated that, in some rural communities, there may be few, or no, reasonable alternatives to having laboratory tests performed by a CAH. Because of this, the commenters believed reasonable cost payment for CAH-performed laboratory tests is warranted, even when specimens are collected in settings other than the CAH from patients who are being registered as CAH patients for the sole purpose of generating higher Medicare payment for the tests.

Response: We agree that an important goal of the CAH legislation is to pay on a reasonable cost basis for services that CAHs provide in their facilities to their inpatients and outpatients. However, we do not believe that legislation can or should be read so broadly as to authorize payment on a reasonable cost basis for laboratory services to patients who do not come to the CAH for those services, but receive them in other settings, including settings in which coverage for the services is available. We also do not agree that because the CAH may be one of only a few sources of laboratory services that the CAH should therefore be paid a higher amount for those services than would otherwise be the case. Therefore, we are not making any change to our proposal based on this comment.

Comment: Several commenters stated that even when a sample is collected outside a CAH, the cost of processing in a CAH laboratory is incurred by the CAH. Because of this circumstance, the commenters recommended that payment be based on the payment method applicable to the site where the processing is done, so that payment for laboratory tests processed at a CAH would be paid on a reasonable cost basis, not under the fee schedule.

Response: We believe the approach recommended by these commenters could create an inappropriate incentive to CAHs to expand their testing activities far beyond their normal service areas, in order to gain cost reimbursement for patients who have no other connection with the CAH other than having a specimen processed by the CAH. In some cases, this could result in payment being made on a cost basis for laboratory services to patients residing in suburban or even urban areas where there is no shortage of qualified laboratories. Such a result would only inappropriately increase payment to CAHs and create market distortions, because non-CAH laboratories performing exactly the same services may be paid substantially less for them. Therefore, we are not adopting this recommendation.

Comment: One commenter agreed with our proposal as it applies to laboratory specimens drawn in health care providers or suppliers other than CAHs, such as SNFs or RHCs, but recommended that we allow reasonable cost payment for clinical diagnostic laboratory tests on specimens drawn in physician clinics that are located in close proximity to the CAH, if the CAH owns the clinic and supplies the personnel who collect the specimens.

Response: While we considered this suggestion, we are not adopting it. A clinic of the type described by the commenter is not a part of the CAH, but is a physician office. We see no basis for treating such a non-CAH setting differently from other non-CAH facilities (such as RHCs) that are similarly owned and located. In the case of an ambulatory patient being seen in a physician office located in close proximity to the CAH, we do not believe it is unreasonable to expect the patient to go to the CAH for the laboratory service as he or she would for therapy or any other CAH outpatient service. Alternatively, the specimen may be collected during the physician visit and payment could be made to the CAH under the laboratory benefit, generally on a fee schedule basis.

Comment: Some commenters stated that the proposed revision is not a clarification but a change from past policy.

Response: We disagree with the commenter, but we do recognize from the questions raised on the issue that there has been some confusion about the policy among rural facilities. To clarify the agency policy in this area and ensure that all relevant issues are publicly noted, we set forth the clarification through notice and comment rulemaking procedures rather than through other processes, such as a program memorandum, a set of responses to “frequently asked questions,” or other document.

Comment: One commenter stated that it is inappropriate for proposed changes to CAH payment to be published in the proposed IPPS regulation. The commenter recommended that if changes are to be made to the payment methodology for those facilities excluded from the IPPS rule, they should be published separately in the Federal Register, not in a proposed rule that would not normally be reviewed by officials associated with CAHs.

Response: The IPPS proposed and final rules are published on an established and regular annual cycle and have been read for many years by a large health care population, including national, State, and local hospital associations as well as individual hospitals, including hospitals paid under the reasonable cost payment system as well as those paid under the IPPS. Because we recognize this as an important tool for disseminating information, we have used the IPPS publication in order to implement several major payment issues relating to CAHs. For example, changes in the CAH payment rules in § 413.70 were included in the IPPS final rule published on August 1, 2002 (67 FR 49982) and the IPPS final rule published on August 1, 2001 (66 FR 39828). We believe this is an appropriate vehicle in providing the information necessary to allow the CAHs access to the information they need to continue to participate knowledgeably in the Medicare program. In fact, we received over 40 comments on the provision alone.

Comment: Some commenters recommended that we withdraw our proposal because reasonable cost payment for clinical diagnostic laboratory tests on specimens collected in non-CAH settings can be an important revenue source for CAHs and yet would generate only a small amount of additional cost to the Medicare program.

Response: For the reasons stated above and in the preamble to the proposed rule, we do not believe it is appropriate to pay on a reasonable cost basis for these laboratory tests. Moreover, doing so might create an unintended incentive for laboratories processing a substantial volume of tests to affiliate with CAHs, in order to obtain the higher level of payment for tests on individuals who are only nominally patients of the CAH. Therefore, we are not adopting this recommendation.

Comment: Some commenters stated that beneficiaries, particularly frail, elderly individuals residing in remote rural areas, could be inconvenienced by our proposed clarification because they would now be required to travel to the CAH to obtain laboratory services payable on a reasonable cost basis. These commenters expressed concern that frail, elderly patients confined to nursing homes could be required by this policy to travel to CAHs to obtain needed laboratory tests.

Response: Under our proposed clarification, Medicare would not deny payment for medically necessary clinical diagnostic laboratory tests that the CAH performs on specimens collected from patients in non-CAH locations. On the contrary, clinical diagnostic laboratory tests performed by CAHs on such specimens would be paid under the same conditions as would apply to such tests furnished by an independent laboratory. In such a case, a CAH would be providing independent laboratory services and generally would be paid under the laboratory fee schedule.

Regarding the concern about the difficulty of travel for some beneficiaries, we believe it is an incorrect assumption that beneficiaries in rural areas will not have specimens collected in their homes or other locations if the CAH is not paid on a cost basis for the collection and travel. If it is medically necessary for the specimen to be collected in the patient's home, the laboratory benefit under Medicare Part B will pay the specimen collection fee (currently $3 per specimen), plus a separate travel allowance (currently at least 75 cents per mile where the average round trip is more than 20 miles) for employees of independent, mobile or hospital-based laboratories to travel to the beneficiary's home. These payments are in addition to payment for performing the tests. (For further details on how specimen collection and travel fees are calculated, see CMS Transmittal AB-98-33, Change Request #526, dated July 1998; this transmittal is available on the CMS Web site at www.cms.hhs.gov.) In many cases, the laboratories collect blood specimens in batches or groups of beneficiaries residing in neighboring areas. This can make the technicians' trips to beneficiaries' residences more cost-effective.

In addition to laboratories, home health agencies that have laboratory provider numbers can perform blood draws at a beneficiary's residence and bill Medicare under the laboratory benefit, using the appropriate codes for specimen collection and travel. Agencies would be reimbursed the $3 specimen collection fee, plus travel costs determined by the Medicare contractor.

It is also important to note that home health agencies with laboratory provider numbers may conduct some of the less complex blood tests themselves, receive the collection and travel fee, and receive a fee through the laboratory benefit for performing the tests. These are called the Clinical Laboratory Improvement Amendments (CLIA)-waived tests, and, among others, include: glucose (blood sugar levels for diabetic patients), fructosamine (also checks blood sugar levels but over longer period of time), hemoglobin (tests hemoglobin levels for patients with anemia), urine dip stick (tests urine for a variety of diseases/infections), and cholesterol/triglyceride (checks for lipid levels for patients with cardiovascular disease) tests.

A variety of other providers can draw blood at a beneficiary's home, often in conjunction with other services necessitating the laboratory tests. For example, while a physician conducts a home visit for evaluation and management, the physician may also draw a blood specimen. If the physician meets applicable requirements under the laboratory benefit, he or she may receive an additional payment for the specimen collection.

The physician also can arrange for a nurse practitioner, physician assistant, or clinical nurse specialist to conduct a home visit and draw blood when they examine the beneficiary. These clinicians are reimbursed at a rate equal to 85 percent of the physician fee schedule for a home visit, and if all applicable billing requirements are met, they are also paid specimen collection and travel fees.

Regarding tests for nursing home patients, we note that if a CAH furnishes laboratory services to a beneficiary in an SNF stay covered by Part A, nonemergency diagnostic laboratory tests—regardless of whether furnished by the SNF directly or under an arrangement with the CAH—would be included within the SNF's bundled PPS per diem payment for the covered stay itself. If a CAH furnishes laboratory services to a beneficiary in an SNF stay not covered by Part A (for example, Part A benefits exhausted; no prior qualifying hospital stay; SNF level of care requirements not met), the SNF consolidated billing restrictions do not apply. However, if the SNF nonetheless elects to bill for such a beneficiary's laboratory services, section 1888(e)(9) of the Act provides that an SNF's Part B bills are to be paid in accordance with the fee schedule that applies to the particular item or service being billed.

In the case of beneficiaries in nursing homes, patients are already under the care of an institution staffed with registered nurses, licensed practical nurses, and nursing assistants, and other health care workers who are presumably well-trained in collecting specimens for analysis, and the nursing homes are already being paid, by Medicare, Medicaid, private insurers, or other means for caring for the patient. Under these circumstances, it would not seem unreasonable to expect the nursing home to take responsibility for collecting the specimens.

Because of the many ways in which specimen collection and travel are payable under Medicare, we do not expect beneficiaries to face reduced access to services under this proposal. We specifically reject the claims made by several commenters that beneficiaries would be able to obtain needed laboratory services only by traveling to the CAH to obtain them.

Comment: Some commenters took exception to the preamble statements that allowing cost reimbursement for laboratory tests on specimens obtained by CAH personnel in non-CAH settings would duplicate existing coverage, create confusion for beneficiaries, and add to the costs of care furnished to Medicare patients. Regarding the costs of care, the commenters stated that because clinical diagnostic laboratory tests are not subject to deductible or coinsurance liability under Medicare, there would be no increase in out-of-pocket costs for beneficiaries.

Response: Regarding duplication of coverage, we have explained in a response to an earlier comment the many ways in which Medicare now pays for specimen collection fees and travel costs. Given this payment provision, adding another, more expensive payment option for the services would duplicate existing coverage without providing any benefit to anyone other than the operators of the CAHs. Despite the commenters' claims to the contrary, we continue to believe patients under the care of one provider (such as a SNF or RHC) might have questions as to why personnel from another provider are coming in to perform functions that could be performed by staff of the facility in which they are being treated. Finally, while there is no deductible or coinsurance liability associated with laboratory services, paying for services on a reasonable cost basis rather than on a fee schedule basis will ultimately drive up the cost of laboratory care provided under Medicare, increasing costs for taxpayers and contributing to general health care cost increases. To the extent Medicare Part B premiums will increase in the future because of current spending rises, we believe adopting the policy recommended by commenters would increase out-of-pocket costs for beneficiaries as well as for all other taxpayers.

Comment: One commenter asked whether the proposed clarification of our policy on payment for clinical diagnostic laboratory tests would be applied prospectively only, or also retroactively.

Response: Although this proposal represents a clarification of policy, we recognize that this policy has not been well understood in all areas. Therefore, we do not plan to direct Medicare contractors to routinely reopen and review past claims for compliance.

After full consideration of public comments on these issues as summarized above, we are adopting our proposed changes to § 413.70 as final without change.

E. Technical Change

On July 30, 1999, we published in the Federal Register a final rule (64 FR 41532) that set forth criteria for a satellite facility of a hospital or hospital unit to be excluded from the IPPS under § 412.25. Section 412.25(e)(3) of the regulations specifies that any unit structured as a satellite facility on September 30, 1999, and excluded from the IPPS on that date, is grandfathered as an excluded hospital to the extent that the unit continues operating under the same terms and conditions, including the number of beds and square footage considered to be part of the unit, in effect on September 30, 1999, except as we specified in § 412.25(e)(4). When we specified the exception for the number of beds and square footage requirement under § 412.25(e)(4), we inadvertently referred to paragraph (e)(4) as being an exception to paragraph (h)(3). We should have specified that it was an exception to paragraph (e)(3). We proposed to correct this reference.

We did not receive any comments on this proposal and, therefore, are adopting the proposed technical change as final.

VII. MedPAC Recommendations

We are required by section 1886(e)(4)(B) of the Act to respond to MedPAC's IPPS recommendations in our annual IPPS rules. We have reviewed MedPAC's March 1, 2003 “Report to the Congress: Medicare Payment Policy” and have given it careful consideration in conjunction with the policies set forth in this document. For further information relating specifically to the MedPAC report or to obtain a copy of the report, contact MedPAC at (202) 653-7220, or visit MedPAC's Web site at: http://www.medpac.gov.

MedPAC's Recommendation 2A-6 concerning the update factor for inpatient hospital operating costs and for hospitals and distinct-part hospital units excluded from the IPPS is discussed in Appendix B to this final rule. MedPAC's other recommendations relating to payments for Medicare inpatient hospital services focused mainly on the expansion of DRGs subject to the postacute care transfer policy, a reevaluation of the labor-related share of the market basket used in determining the hospital wage index, an increase in the DSH adjustment, and payments to rural hospitals. These recommendations and our responses are set forth below:

Recommendation 2A-1: The Secretary should add 13 DRGs to the postacute transfer policy in FY 2004 and then evaluate the effects on hospitals and beneficiaries before proposing further expansions.

Response: After reevaluation of this recommendation, in this final rule we are expanding the postacute care transfer policy to include 21 additional DRGs for FY 2004, although we are removing 2 DRGs from the current list. A thorough discussion of this provision, including a summary of MedPAC's analysis, can be found at section IV.A.3. of this preamble.

Recommendation 2A-2: The Congress should enact a low-volume adjustment to the rates used in the inpatient PPS. This adjustment should apply only to hospitals that are more than 15 miles from another facility offering acute inpatient care.

Response: MedPAC's analysis “revealed that hospitals with a small volume of total discharges have higher costs per discharge than larger facilities, after controlling for the other cost-related factors recognized in the payment system.” Although there are special payment protections for some rural hospitals such as CAHs, SCHs, and MDHs, MedPAC believes these provisions do not sufficiently target hospitals with low discharge volume.

This recommendation, which MedPAC estimates would increase Medicare payments to hospitals by less than $50 million in FY 2004, and others requiring Congressional action, should be considered in the context of larger discussions within Congress and between Congress and the Administration regarding Medicare reform and payment refinements. Therefore, we are not responding specifically to MedPAC's recommendation regarding a low-volume adjustment to the IPPS payments at this time.

Recommendation 2A-3: The Secretary should reevaluate the labor share used in the wage index system that geographically adjusts rates in the inpatient PPS, with any resulting change phased in over 2 years.

Response: We define the labor-related share to include costs that are likely related to, influenced by, or vary with local labor markets, even if they could be purchased in a national market. Since the implementation of the IPPS, the labor-related share has been determined by adding together the cost weights from categories in the hospital market basket that are influenced by local labor markets. When the hospital market basket weights are updated or rebased, the labor-related share is updated. The estimate of the labor-related share using the most recently revised and rebased hospital market basket (1997-based) is 72.495 percent.

In the August 1, 2002 IPPS final rule, we elected to continue to use 71.066 percent as the labor-related share applicable to the standardized amounts (67 FR 50041). At that time, we indicated that we would conduct further analysis to determine the most appropriate methodology for the labor-related share. Again, in the May 19, 2003 proposed rule, we did not propose to use the updated labor-related share for FY 2004 because we have not yet completed our research into the appropriateness of this updated measure. Specifically, we continue to review the labor-related share in two ways. First, we are performing regression analysis with the expectation that it would help give an alternative indication of the labor-related share. Second, we continue to reevaluate the methodology we currently use for determining the labor-related share using the hospital market basket.

Our regression analysis is an attempt to explain the variation in operating cost per case for a given year using many different explanatory variables, such as case-mix, DSH status, and ownership type. We described this methodology and some of our initial results in the May 9, 2002 Federal Register (67 FR 31447-31479). However, the findings from the regressions continue to be both difficult to explain and inconsistent with the underlying cost data. Thus, we believe at this point that the regression results are not robust enough to support changing the current labor-related share measurement.

We also continue to explore all options for alternative data or methodology for determining the labor-related share using the hospital market basket. We have researched various alternative data sources for use in further breaking down the cost categories in the market basket and have evaluated alternative methodologies to determine the feasibility of separating the labor-related portion or the portion that varies with local labor markets from the portion that does not vary. While each of these alternatives has strengths and weaknesses, it is not clear at this point that any one alternative data source or methodology is superior to the current methodology. We will continue to research these alternatives.

Comment: Several commenters suggested the labor share should only be adjusted by those costs (wages and salaries and benefits) that are reflected in the wage index survey. Commenters suggested that CMS should consider reducing the labor-related share for rural hospitals or having different labor shares by geographic location.

Response: We define the labor-related share to include all costs that are likely related to, influenced by, or vary with local labor markets, even if they could be purchased in a national market. This differs from the hospital wage index survey, which only collects direct labor and patient-related contract costs. Using only those direct labor costs reflected in the wage index survey would mean redefining the term labor-related share and would likely leave out many of the other costs that do vary with the local labor market.

As indicated in prior rules, we continue to research alternative methodologies for determining the labor-related share, including reexamining the labor portion of each of the individual market basket categories. However, due to a lack of one definitive data source, our analysis is still preliminary and, therefore, we will continue to use 71.066 percent as the labor-related share applicable to the standardized amounts while we conduct further analysis to determine the most appropriate methodology for determining the labor-related share.

It is currently our policy to use a national labor-related share to apply to the national PPS standardized amounts. This policy has been in effect since the implementation of the IPPS in 1983. We will consider the commenters' recommended alternative approaches, such as different labor shares for urban and rural hospitals or labor shares that vary by more detailed geographic area, as part of our ongoing research efforts. However, until we have completed our research, we will continue to use only a national labor-related share, which is currently 71.066 percent and was calculated from the 1992-based market basket.

Comment: One commenter believed that we should examine each of the categories currently included in the labor share and determine which portion of that category was actually labor-related or varied with the local labor market.

Response: We agree with the commenter that it is important that the labor-related portion of the market basket include only those categories that are actually labor-related or vary with the local labor market. As we indicated in the May 19, 2003 rule, we are continuing to explore all options for accounting for the labor-related share, including reexamining each of the categories included in the current labor share (particularly professional fees, postage, and other labor-intensive services) to make sure the labor share represents only those costs that do vary with the local labor market. However, our preliminary research has indicated that much of the data needed to break out details from each of the current market basket categories into labor and nonlabor-related components are not readily available on a national basis. We will continue to research various data sources for this information and will update the labor share as needed once our research is complete.

Recommendation 2A-4: The Congress should raise the inpatient base rate for hospitals in rural and other urban areas to the level of the rate for those in large urban areas, phased in over 2 years.

Response: This recommendation, which MedPAC estimates would increase Medicare payments to hospitals by between $200 and $600 million in FY 2004, and others requiring Congressional action, should be considered in the context of larger discussions within Congress and between Congress and the Administration regarding Medicare reform and payment refinements. Therefore, we are not responding specifically to MedPAC's recommendation regarding raising the base rate for hospitals in rural and other urban areas at this time.

Recommendation 2A-5: The Congress should raise the cap on the disproportionate share add-on a hospital can receive in the inpatient PPS from 5.25 percent to 10 percent, phased in over 2 years.

Response: This recommendation, which MedPAC estimates would increase Medicare payments to hospitals by between $50 and $200 million in FY 2004, and others requiring Congressional action, should be considered in the context of larger discussions within Congress and between Congress and the Administration regarding Medicare reform and payment refinements. Therefore, we are not responding specifically to MedPAC's recommendation regarding raising the maximum DSH adjustments at this time.

VIII. Other Required Information

A. Requests for Data From the Public

In order to respond promptly to public requests for data related to the prospective payment system, we have established a process under which commenters can gain access to raw data on an expedited basis. Generally, the data are available in computer tape or cartridge format; however, some files are available on diskette as well as on the Internet at http://www.hcfa.gov/stats/pufiles.htm. In the May 19, 2003 proposed rule, we published a list of data files that are available for purchase from CMS or that may be downloaded from the Internet free of charge (68 FR 27226 through 27228).

B. Collection of Information Requirements

This final rule directly does not impose any collection and recordkeeping requirements. Consequently, it does not need to be reviewed by the Office of Management and Budget under the authority of the Paperwork Reduction Act of 1995.

List of Subjects

42 CFR Part 412

• Administrative practice and procedure
• Health facilities
• Medicare,

42 CFR Part 413

• Health facilities
• Kidney diseases
• Medicare
• Puerto Rico
• Reporting and recordkeeping requirements

For the reasons stated in the preamble of this final rule, the Centers for Medicare & Medicaid Services amends 42 CFR chapter IV as follows:

PART 412—PROSPECTIVE PAYMENT SYSTEMS FOR INPATIENT HOSPITAL SERVICES

1. The authority citation for part 412 continues to read as follows:

Authority: Secs. 1102 and 1871 of the Social Security Act (42 U.S.C. 1302 and 1395hh).

2. Section 412.4 is amended by—

A. Revising paragraphs (b), (c), and (d).

B. In paragraph (f)(1), revising the reference “paragraph (b)(1) or (c)” to read “paragraph (b) or (c)”.

The revisions read as follows:

3. Section 412.22 is amended by:

A. Republishing the introductory text of paragraph (e)(5) and revising the first sentence of paragraph (e)(5)(i).

B. Revising paragraph (f).

The revisions read as follows:

4. Section 412.23 is amended by revising paragraphs (e)(3)(ii) and (e)(3)(iii) to read as follows:

5. In § 412.25(e)(4), introductory text, the reference “paragraph (h)(3) of this section” is revised to read “paragraph (e)(3) of this section”.

6. Section 412.87 is amended by revising paragraph (b)(3) to read as follows:

7. Section 412.105 is amended by—

A. In paragraph (a)(1), introductory text, revising the phrase “paragraph (f) of this section” to read “paragraphs (f) and (h) of this section”.

B. In paragraph (a)(1)(i), revising the phrase “affiliated groups” to read “Medicare GME affiliated groups”.

C. Revising paragraph (b).

D. Adding a sentence at the end of paragraph (f)(1)(v).

E. In paragraph (f)(1)(vi), revising the phrase “affiliated group” to read “Medicare GME affiliated group”.

F. Revising paragraph (f)(1)(x).

The revisions and additions read as follows:

7. Section 412.106 is amended by revising paragraphs (a)(1)(ii) and (b)(4)(i) to read as follows:

8. In § 412.112, the introductory text is republished and a new paragraph (d) is added to read as follows:

9. Section 412.116 is amended by revising paragraph (e) to read as follows:

10. Section 412.230 is amended by—

A. Republishing paragraph (e)(2) introductory text.

B. Revising paragraph (e)(2)(ii)(A).

The revisions read as follows:

11. Section 412.278 is amended by revising paragraph (f)(2)(i) to read as follows:

PART 413—PRINCIPLES OF REASONABLE COST REIMBURSEMENT; PAYMENT FOR END-STAGE RENAL DISEASE SERVICES; OPTIONAL PROSPECTIVELY DETERMINED PAYMENT RATES FOR SKILLED NURSING FACILITIES

1. The authority citation for part 413 is revised to read as follows:

Authority: Secs. 1102, 1812(d), 1814(b), 1815, 1833(a), (i), and (n), 1871, 1881, 1883, and 1886 of the Social Security Act (42 U.S.C. 1302, 1395d(d), 1395f(b), 1395g, 1395l(a), (i), and (n), 1395hh, 1395rr, 1395tt, and 1395ww).

2. Section 413.70 is amended by revising paragraph (b)(2)(iii), introductory text, to read as follows:

3. Section 413.85 is amended by—

A. Republishing the introductory text of paragraph (d)(1) and adding a new paragraph (d)(1)(iii).

B. Adding a new paragraph (g)(3).

C. Republishing the introductory text of paragraph (h) and revising paragraph (h)(3).

The addition and revision read as follows.

(d) General payment rules. (1) Payment for a provider's net cost of nursing and allied health education activities is determined on a reasonable cost basis, subject to the following conditions and limitations:

(iii) The costs of certain nonprovider-operated programs at wholly owned subsidiary educational institutions are reimbursable on a reasonable cost basis if the provisions of paragraph (g)(3) of this section are met.

(g) Payments for certain nonprovider-operated programs. * * *

(3) Special rule: Payment for certain nonprovider-operated programs at wholly owned subsidiary educational institutions.

(i) Effective for portions of cost reporting periods occurring on or after October 1, 2003, a provider that incurs costs for a nursing or allied health education program(s) where those program(s) had originally been provider-operated according to the criteria at paragraph (f) of this section, and then operation of the program(s) was transferred to a wholly owned subsidiary educational institution in order to meet accreditation standards prior to October 1, 2003, and where the provider has continuously incurred the costs of both the classroom and clinical training portions of the program(s) at the educational institution, may receive reasonable cost payment for such a program(s) according to the specifications under paragraphs (g)(3)(ii) and (g)(3)(iii) of this section.

(ii) Payment for the incurred costs of educational activities identified in paragraph (g)(3)(i) of this section will be made on a reasonable cost basis if a provider, as described in paragraph (g)(3)(i) of this section, received Medicare reasonable cost payment for those nursing and allied health education program(s) both prior and subsequent to the date the provider transferred operation of the program(s) to its wholly owned subsidiary educational institution (and ceased to be a provider-operated program(s) according to the criteria under paragraph (f) of this section).

(iii) The provider that meets the requirements in paragraphs (g)(3)(i) and (g)(3)(ii) of this section will be eligible to receive payment under this paragraph for: (A) the clinical training costs incurred for the program(s) as described in paragraph (g)(3)(i) of this section; and (B) classroom costs, but only those costs incurred by the provider for the courses that were included in the programs.

(h) Activities treated as normal operating costs. The costs of the following educational activities incurred by a provider but not operated by that provider are recognized only as normal operating costs and paid in accordance with the reimbursement principles specified in part 412 of this subchapter. They include:

(3) Educational seminars, workshops, and continuing education programs in which the employees participate that enhance the quality of medical care or operating efficiency of the provider and, effective October 1, 2003, do not lead to the ability to practice and begin employment in a nursing or allied health specialty.

4. Section 413.86 is amended by—

A. Under paragraph (b)—

(1) Removing the definitions of “Affiliated group” and “Affiliation agreement''.

(2) Adding definitions of “Community support”, “Medicare GME affiliated agreement”, “Medicare GME affiliated group”, and “Redistribution of costs” in alphabetical order.

(3) Under the definition of “Rural track FTE limitation”, revising the phrase “paragraph (g)(11)” to read “paragraph (g)(12)”.

B. Revising the introductory text of paragraph (f).

C. Adding a new paragraph (f)(4)(iv).

D. In paragraph (g)(1)(i), revising the reference “paragraphs (g)(1)(ii) and (g)(1)(iii)” to read “paragraphs (g)(1)(ii) through (g)(1)(iv)”.

E. Revising the introductory text of paragraph (g)(4).

F. Revising paragraph (g)(4)(iv).

G. Revising the introductory text of paragraph (g)(5).

H. Adding a new paragraph (g)(5)(vii).

I. Revising paragraphs (g)(6)(i)(D) and (g)(6)(i)(E).

J. Revising paragraph (g)(7).

K. Revising the introductory text of paragraph (g)(12).

L. Revising paragraph (g)(12)(i).

M. Revising paragraph (g)(12)(ii), introductory text.

N. Revising paragraph (g)(12)(ii)(A).

O. Revising paragraph (g)(12)(ii)(B)(1)(i).

P. Revising paragraph (g)(12)(iii).

Q. Revising paragraph (g)(12)(iv), introductory text.

R. Revising paragraph (g)(12)(iv)(A).

S. Revising paragraph (g)(12)(iv)(B)(1).

T. Redesignating paragraphs (i) and (j) as paragraphs (j) and (k), respectively, and adding a new paragraph (i).

The additions and revisions read as follows:

[Editorial Note: The following Addendum and appendic es will not appear in the Code of Federal Regulations.]

Addendum—Schedule of Standardized Amounts Effective With Discharges Occurring on or After October 1, 2003 and Update Factors and Rate-of-Increase Percentages Effective With Cost Reporting Periods Beginning on or After October 1, 2003

I. Summary and Background

In this Addendum, we are setting forth the amounts and factors for determining prospective payment rates for Medicare hospital inpatient operating costs and Medicare hospital inpatient capital-related costs. We are also setting forth rate-of-increase percentages for updating the target amounts for hospitals and hospital units excluded from the IPPS.

For discharges occurring on or after October 1, 2003, except for SCHs, MDHs, and hospitals located in Puerto Rico, each hospital's payment per discharge under the IPPS will be based on 100 percent of the Federal national rate, which will be based on the national adjusted standardized amount. This amount reflects the national average hospital costs per case from a base year, updated for inflation.

SCHs are paid based on whichever of the following rates yields the greatest aggregate payment: the Federal national rate; the updated hospital-specific rate based on FY 1982 costs per discharge; the updated hospital-specific rate based on FY 1987 costs per discharge; or the updated hospital-specific rate based on FY 1996 costs per discharge.

Under section 1886(d)(5)(G) of the Act, MDHs are paid based on the Federal national rate or, if higher, the Federal national rate plus 50 percent of the difference between the Federal national rate and the updated hospital-specific rate based on FY 1982 or FY 1987 costs per discharge, whichever is higher. MDHs do not have the option to use their FY 1996 hospital-specific rate.

For hospitals in Puerto Rico, the payment per discharge is based on the sum of 50 percent of a Puerto Rico rate that reflects base year average costs per case of Puerto Rico hospitals and 50 percent of a blended Federal national rate (a discharge-weighted average of the national large urban and other areas standardized amounts). (See section II.D.3. of this Addendum for a complete description.)

As discussed below in section II. of this Addendum, we are making changes in the determination of the prospective payment rates for Medicare inpatient operating costs for FY 2004. The changes, to be applied prospectively effective with discharges occurring on or after October 1, 2003, affect the calculation of the Federal rates. In section III. of this Addendum, we discuss our changes for determining the prospective payment rates for Medicare inpatient capital-related costs for FY 2004. Section IV. of this Addendum sets forth our changes for determining the rate-of-increase limits for hospitals excluded from the IPPS for FY 2004. Section V. of this Addendum sets forth policies on payment for blood clotting factor administered to hemophilia patients. The tables to which we refer in the preamble of this final rule are presented in section VI. of this Addendum.

II. Changes to Prospective Payment Rates for Hospital Inpatient Operating Costs for FY 2004

The basic methodology for determining prospective payment rates for hospital inpatient operating costs is set forth at § 412.63. The basic methodology for determining the prospective payment rates for hospital inpatient operating costs for hospitals located in Puerto Rico is set forth at §§ 412.210 and 412.212. Below, we discuss the factors used for determining the prospective payment rates.

In summary, the standardized amounts set forth in Tables 1A and 1C of section VI. of this Addendum reflect—

• Updates of 3.4 percent for all areas (that is, the full market basket percentage increase of 3.4 percent);
• An adjustment to ensure the proposed DRG recalibration and wage index update and changes, as well as the add-on payments for new technology, are budget neutral, as provided for under sections 1886(d)(4)(C)(iii) and (d)(3)(E) of the Act, by applying new budget neutrality adjustment factors to the large urban and other standardized amounts;
• An adjustment to ensure the effects of geographic reclassification are budget neutral, as provided for in section 1886(d)(8)(D) of the Act, by removing the FY 2003 budget neutrality factor and applying a revised factor;
• An adjustment to apply the new outlier offset by removing the FY 2003 outlier offsets and applying a new offset.

A. Calculation of Adjusted Standardized Amounts

1. Standardization of Base-Year Costs or Target Amounts

The national standardized amounts are based on per discharge averages of adjusted hospital costs from a base period (section 1886(d)(2)(A) of the Act) or, for Puerto Rico, adjusted target amounts from a base period (section 1886(d)(9)(B)(i) of the Act), updated and otherwise adjusted in accordance with the provisions of section 1886(d) of the Act. The preamble to the September 1, 1983 interim final rule (48 FR 39763) contained a detailed explanation of how base-year cost data (from cost reporting periods ending during FY 1981) were established in the initial development of standardized amounts for the IPPS. The September 1, 1987 final rule (52 FR 33043, 33066) contains a detailed explanation of how the target amounts were determined, and how they are used in computing the Puerto Rico rates.

Sections 1886(d)(2)(B) and (d)(2)(C) of the Act require us to update base-year per discharge costs for FY 1984 and then standardize the cost data in order to remove the effects of certain sources of cost variations among hospitals. These effects include case-mix, differences in area wage levels, cost-of-living adjustments for Alaska and Hawaii, indirect medical education costs, and costs to hospitals serving a disproportionate share of low-income patients.

Under sections 1886(d)(2)(H) and (d)(3)(E) of the Act, in determining payments under the IPPS, the Secretary estimates from time to time the proportion of costs that are wages and wage-related costs. Based on the estimated labor-related share, the standardized amounts are divided into labor-related and nonlabor-related amounts. As discussed in section IV. of the preamble to the August 1, 2002 IPPS final rule, when we revised the market basket in FY 2003, we did not revise the labor share of the standardized amount (the proportion adjusted by the wage index). We consider 71.1 percent of costs to be labor-related for purposes of the IPPS. The average labor share in Puerto Rico is 71.3 percent.

2. Computing Large Urban and Other Area Average Standardized Amounts

Sections 1886(d)(2)(D) and (d)(3) of the Act require the Secretary to compute two average standardized amounts for discharges occurring in a fiscal year: one for hospitals located in large urban areas and one for hospitals located in other areas. In addition, under sections 1886(d)(9)(B)(iii) and (d)(9)(C)(i) of the Act, the average standardized amount per discharge must be determined for hospitals located in large urban and other areas in Puerto Rico. In accordance with section 1886(b)(3)(B)(i) of the Act, the large urban average standardized amount is 1.6 percent higher than the other area average standardized amount.

Section 402(b) of Pub. L. 108-7 required that, effective for discharges occurring on or after April 1, 2003, and before October 1, 2003, the Federal rate for all IPPS hospitals would be based on the large urban standardized amount. However, for discharges occurring on or after October 1, 2003, the Federal rate will again be calculated based on separate average standardized amounts for hospitals in large urban areas and for hospitals in other areas.

Section 1886(d)(2)(D) of the Act defines “urban area” as those areas within a Metropolitan Statistical Area (MSA). A “large urban area” is defined as an urban area with a population of more than 1 million. In addition, section 4009(i) of Pub. L. 100-203 provides that a New England County Metropolitan Area (NECMA) with a population of more than 970,000 is classified as a large urban area. As required by section 1886(d)(2)(D) of the Act, population size is determined by the Secretary based on the latest population data published by the Bureau of the Census. Urban areas that do not meet the definition of a “large urban area” are referred to as “other urban areas.” Areas that are not included in MSAs are considered “rural areas” under section 1886(d)(2)(D) of the Act. Payment for discharges from hospitals located in large urban areas will be based on the large urban standardized amount. Payment for discharges from hospitals located in other urban and rural areas will be based on the other standardized amount.

As discussed previously, on June 6, 2003, OMB announced revised definitions of MSAs and new definitions of Micropolitan Statistical Areas and Combined Statistical Areas. In order to implement these changes for the IPPS, it is necessary to identify the new area designation for each county and hospital in the country. Because this process will have to be extensively reviewed and verified, we were unable to undertake it before publication of this final rule. Therefore, we are continuing to use MSAs based on OMB's definitions of MSAs prior to June 6, 2003. Based on those definitions, 63 areas meet the criteria to be defined as large urban areas for FY 2004. These areas are identified in Table 4A of section VI. of this Addendum.

3. Updating the Average Standardized Amounts

In accordance with section 1886(d)(3)(A)(iv) of the Act, we are updating the arge urban areas' and the other areas' average standardized amounts for FY 2004 by the full estimated market basket percentage increase for hospitals in all areas, as specified in section 1886(b)(3)(B)(i)(XIX) of the Act. The percentage change in the market basket reflects the average change in the price of goods and services purchased by hospitals to furnish inpatient care. The most recent forecast of the hospital market basket increase for FY 2004 is 3.4 percent. Thus, for FY 2004, the update to the average standardized amounts equals 3.4 percent for hospitals in all areas.

Although the update factors for FY 2004 are set by law, we are required by section 1886(e)(3) of the Act to report to the Congress our initial recommendation of update factors for FY 2004 for both IPPS hospitals and hospitals excluded from the IPPS. Our recommendation on the update factors (which is required by sections 1886(e)(4)(A) and (e)(5)(A) of the Act) is set forth as Appendix B of this final rule.

Comment: One commenter recommended an increase to the market basket that would account for large increases in the costs of malpractice, pensions, health benefits, pharmaceuticals, and new technology that hospitals are facing.

Response: The hospital market basket is structured to measure the change in prices for an exhaustive list of inputs used by hospitals in providing services. The index measures the “pure” price change of those inputs and appropriately does not measure changes in quantity or intensity. These nonprice factors include shifts in the skill mix of employees, increased amounts of labor purchased, increased malpractice coverage, the increased use of pharmaceuticals and technology in providing care, and movements toward more or less intensive pharmaceuticals and technology. Nonprice factors such as these may be contributing to the increases in cost that hospitals are currently facing.

In addition, the most recent data available are used to forecast the market basket price changes and are intended to reflect conditions that hospitals will face in the upcoming fiscal year. As it is intended, the hospital market basket measures the national average price increase and will not reflect geographic differences from one geographic area to another. In other words, while one area may see a large surge in the prices of inputs, another area may actually be experiencing much smaller increases in the prices of these inputs. This may also be contributing to the increased costs to which the commenter referred. Therefore, we believe that the market basket is an accurate representation of the national average price increase facing hospitals in providing services, and the 3.4 percent increase for FY 2004 provides an adequate update to hospitals to account for the inflationary increase in costs.

4. Other Adjustments to the Average Standardized Amounts

As in the past, we adjust the FY 2004 standardized amounts to remove the effects of the FY 2003 geographic reclassifications and outlier payments before applying the FY 2004 updates. We then apply the new offsets to the standardized amounts for outliers and geographic reclassifications for FY 2004.

We do not remove the prior year's budget neutrality adjustments for reclassification and recalibration of the DRG weights and for updated wage data because, in accordance with section 1886(d)(4)(C)(iii) of the Act, estimated aggregate payments after the changes in the DRG relative weights and wage index should equal estimated aggregate payments prior to the changes. If we removed the prior year adjustment, we would not satisfy this condition.

Budget neutrality is determined by comparing aggregate IPPS payments before and after making the changes that are required to be budget neutral (for example, reclassifying and recalibrating the DRGs, updating the wage data, and geographic reclassifications). We include outlier payments in the payment simulations because outliers may be affected by changes in these payment parameters. Because the changes to the postacute care transfer policy discussed in section IV.A. of the preamble of this final rule are not budget neutral, we included the effects of expanding this policy to additional DRGs prior to estimating the payment effects of the DRG and wage data changes.

a. Recalibration of DRG Weights and Updated Wage Index—Budget Neutrality Adjustment.—Section 1886(d)(4)(C)(iii) of the Act specifies that, beginning in FY 1991, the annual DRG reclassification and recalibration of the relative weights must be made in a manner that ensures that aggregate payments to hospitals are not affected. As discussed in section II. of the preamble, we normalized the recalibrated DRG weights by an adjustment factor, so that the average case weight after recalibration is equal to the average case weight prior to recalibration. However, equating the average case weight after recalibration to the average case weight before recalibration does not necessarily achieve budget neutrality with respect to aggregate payments to hospitals because payments to hospitals are affected by factors other than average case weight. Therefore, as we have done in past years, we are making a budget neutrality adjustment to ensure that the requirement of section 1886(d)(4)(C)(iii) of the Act is met.

Section 1886(d)(3)(E) of the Act requires us to update the hospital wage index on an annual basis beginning October 1, 1993. This provision also requires us to make any updates or adjustments to the wage index in a manner that ensures that aggregate payments to hospitals are not affected by the change in the wage index.

Section 4410 of Pub. L. 105-33 provides that, for discharges on or after October 1, 1997, the area wage index applicable to any hospital that is not located in a rural area may not be less than the area wage index applicable to hospitals located in rural areas in that State. This provision is required by section 4410(b) of Pub. L. 105-33 to be budget neutral. Therefore, we include the effects of this provision in our calculation of the wage update budget neutrality factor.

In addition, we are required to ensure that any add-on payments for new technology under section 1886(d)(5)(K) of the Act are budget neutral. As discussed in section II.E. of this final rule, we are approving two new technologies for add-on payments in FY 2004. We estimate that the total add-on payments for these new technologies will be $14.4 million for FY 2004.

To comply with the requirement that DRG reclassification and recalibration of the relative weights be budget neutral, and the requirement that the updated wage index be budget neutral, we used FY 2002 discharge data to simulate payments and compared aggregate payments using the FY 2003 relative weights, wage index, and new technology add-on payments to aggregate payments using the FY 2004 relative weights and wage index, plus the add-on payments for new technology. The same methodology was used for the FY 2003 budget neutrality adjustment.

Based on this comparison, we computed a budget neutrality adjustment factor equal to 1.005522. We also adjust the Puerto Rico-specific standardized amounts for the effect of DRG reclassification and recalibration. We computed a budget neutrality adjustment factor for Puerto Rico-specific standardized amounts equal to 1.001661. These budget neutrality adjustment factors are applied to the standardized amounts without removing the effects of the FY 2003 budget neutrality adjustments.

In addition, we are applying these same adjustment factors to the hospital-specific rates that are effective for cost reporting periods beginning on or after October 1, 2003. (See the discussion in the September 4, 1990 final rule (55 FR 36073).)

b. Reclassified Hospitals—Budget Neutrality Adjustment.—Section 1886(d)(8)(B) of the Act provides that, effective with discharges occurring on or after October 1, 1988, certain rural hospitals are deemed urban. In addition, section 1886(d)(10) of the Act provides for the reclassification of hospitals based on determinations by the MGCRB. Under section 1886(d)(10) of the Act, a hospital may be reclassified for purposes of the standardized amount or the wage index, or both.

Under section 1886(d)(8)(D) of the Act, the Secretary is required to adjust the standardized amounts so as to ensure that aggregate payments under the IPPS after implementation of the provisions of sections 1886(d)(8)(B) and (C) and 1886(d)(10) of the Act are equal to the aggregate prospective payments that would have been made absent these provisions. To calculate this budget neutrality factor, we used FY 2002 discharge data to simulate payments, and compared total IPPS payments prior to any reclassifications to total IPPS payments after reclassifications. Based on these simulations, we are applying an adjustment factor of 0.992026 to ensure that the effects of reclassification are budget neutral.

The adjustment factor is applied to the standardized amounts after removing the effects of the FY 2003 budget neutrality adjustment factor. We note that the FY 2004 adjustment reflects FY 2004 wage index and standardized amount reclassifications approved by the MGCRB or the Administrator, and the effects of section 1886(d)(10)(D)(v) of the Act to extend wage index reclassifications for 3 years.

c. Outliers.—Section 1886(d)(5)(A) of the Act provides for payments in addition to the basic prospective payments, for “outlier” cases involving extraordinarily high costs. To qualify for outlier payments, a case must have costs above a fixed-loss cost threshold amount (a dollar amount by which the costs of a case must exceed payments in order to qualify for outlier payment). To determine whether the costs of a case exceed the fixed-loss threshold, a hospital's cost-to-charge ratio is applied to the total covered charges for the case to convert the charges to costs. Payments for eligible cases are then made based on a marginal cost factor, which is a percentage of the costs above the threshold.

Under section 1886(d)(5)(A)(iv) of the Act, outlier payments for any year must be projected to be not less than 5 percent nor more than 6 percent of total operating DRG payments plus outlier payments. Section 1886(d)(3)(B) of the Act requires the Secretary to reduce the average standardized amounts by a factor to account for the estimated proportion of total DRG payments made to outlier cases. Similarly, section 1886(d)(9)(B)(iv) of the Act requires the Secretary to reduce the average standardized amounts applicable to hospitals in Puerto Rico to account for the estimated proportion of total DRG payments made to outlier cases.

i. FY 2004 outlier fixed-loss cost threshold. In the August 1, 2002 IPPS final rule (67 FR 50124), we established a threshold for FY 2003 that was equal to the prospective payment rate for the DRG, plus any IME and DSH payments and any additional payments for new technology, plus $33,560. The marginal cost factor (the percent of costs paid after costs for the case exceed the threshold) was 80 percent.

In the May 19, 2003 proposed rule, we proposed to establish a fixed-loss cost outlier threshold equal to the prospective payment rate for the DRG plus any IME and DSH payments, and any add-on payments for new technology, plus $50,645. However, we also stated that the final FY 2004 threshold was likely to be different from that proposed threshold, as a result of any changes to outlier policy subsequent to a proposed rule published on March 5, 2003. Subsequently, we published three central changes to our outlier policy in a final rule on June 9, 2003.

The first of the changes was that fiscal intermediaries will use more up-to-date data when determining the cost-to-charge ratio for each hospital. Currently, fiscal intermediaries use the hospital's most recent settled cost report. We revised our regulations to specify that fiscal intermediaries will use either the most recent settled or the most recent tentative settled cost report, whichever is from the latest reporting period.

The second change removed the requirement in our regulations specifying that a fiscal intermediary will assign a hospital the statewide average cost-to-charge ratio when the hospital has a cost-to-charge ratio that falls below an established threshold (3 standard deviations below the national geometric mean cost-to-charge ratio). We specified that hospitals will receive their actual cost-to-charge ratios no matter how low their ratios actually fall.

The third change added a provision to our regulations to provide that the outlier payments for some hospitals will become subject to reconciliation when the hospitals' cost reports are settled. In addition, outlier payments will be subject to an adjustment to account for the time value of any outlier overpayments or underpayments that are ultimately reconciled.

To calculate the FY 2004 outlier thresholds, we simulated payments by applying FY 2004 rates and policies using cases from the FY 2002 MedPAR file. Therefore, in order to determine the appropriate FY 2004 threshold, it was necessary to inflate the charges on the MedPAR claims by 2 years, from FY 2002 to FY 2004.

As discussed in the August 1, 2002 IPPS final rule (67 FR 50124), rather than use the rate-of-cost increase from hospitals' FY 1998 and FY 1999 cost reports to project the rate-of-increase from FY 2001 to FY 2003, as had been done in prior years, we used a 2-year average annual rate of change in charges per case to calculate the FY 2003 outlier threshold.

We are continuing to use the 2-year average annual rate of change in charges per case to establish the FY 2004 threshold. The 2-year average annual rate of change in charges per case from FY 2000 to FY 2001, and from FY 2001 to FY 2002, was 12.5978 percent annually, or 26.8 percent over 2 years.

In the past, we used cost-to-charge ratios from the Provider Specific File, and multiplied these ratios by the charges for each case to estimate costs. After the changes in policy enacted by the final outlier rule this year, it is necessary to calculate more recent cost-to-charge ratios because fiscal intermediaries will now use the latest tentatively settled cost report instead of the latest settled cost report to determine a hospital's cost-to-charge ratio. Therefore, to approximate using the latest tentative settled cost reports in our estimate of the FY 2004 outlier threshold, we calculated updated cost-to-charge ratios using the following three steps: for each hospital, we matched charges-per-case to costs-per-case from the most recent cost reporting year; we then divided each hospital's costs by its charges to calculate the cost-to-charge ratio for each hospital; and we multiplied charges from each case in the FY 2002 MedPAR (inflated to FY 2004) by this cost-to-charge ratio to calculate the cost per case. The final outlier rule also established the policy that fiscal intermediaries are to reconcile outlier payments at the time of cost report final settlement if a hospital's actual operating or capital cost-to-charge ratios are found to be substantially different from the cost-to-charge ratios used during that time period to make outlier payments.

However, it is difficult to project which hospitals will be subject to reconciliation of their outlier payments using available data. For example, for most hospitals, the latest available cost data are from FY 2000. In addition, the amount of fiscal intermediary resources necessary to undertake reconciliation will ultimately influence the number of hospitals reconciled. Without actual experience with the reconciliation process, it is difficult to predict the number of hospitals that will be reconciled. However, as later data become available, particularly data reflecting hospital's latest tentative settled cost-to-charge ratios, we will be better able to assess the appropriate number of hospitals to be reconciled.

Based on our analysis of hospitals that have been consistently overpaid recently for outliers, we have identified approximately 50 hospitals we believe will be reconciled. Therefore, for these hospitals, to account for the fact that the reconciliation will result in different outlier payments than predicted using the cost-to-charge ratios calculated as described above, we attempted to project each hospital's cost-to-charge ratio based on its rate of increase in charges per case based on FY 2002 charges, compared to costs (inflated to FY 2002 using actual market basket increases).

Using this methodology, we are establishing a fixed-loss cost outlier threshold equal to the prospective payment rate for the DRG, plus any IME and DSH payments, and any add-on payments for new technology, plus $31,000.

This single threshold will be applicable to qualify for both operating and capital outlier payments. We also are maintaining the marginal cost factor for cost outliers at 80 percent.

Comment: One commenter supported our changes to the outlier payment methodology but asked that we reconsider and revise the outlier threshold to at least a level of increase consistent with prior years. Other commenters asked that we lower the threshold to reflect the financial impact of the new outlier policies, to allow deserving hospitals to qualify for outlier payments and to ensure that hospitals receive the statutory mandated level of 5 to 6 percent of total DRG payments set aside for outliers. Another commenter reasoned that hospitals that have had their outlier payments dwindle to record low amounts will have no incentive to treat high-cost cases; therefore, the outlier threshold must be lowered. Another commenter noted that the current proposed threshold makes it almost impossible for hospitals to qualify for outlier payments and will cause hospitals to lose an extraordinary amount of money before additional outlier payments become available.

Other commenters indicated that they had conducted research, using the 2001 MedPAR file, which showed that the threshold required to spend 5.1 percent of total DRG payments decreased by 45 percent when the cost-to-charge ratios used to estimate costs were updated from the latest final settled to the latest tentatively settled cost report. Based on this finding, the commenters recommended a 45-percent reduction to the proposed outlier threshold, which would yield a threshold less than $28,000.

Some commenters believed that, in light of the changes adopted this year, it is appropriate that CMS revert to using changes in hospital costs to set the charge inflation factor rather than changes in hospital charges. The commenters explained that the combination of the changes made to the outlier policy and a return to using a cost inflation factor would lead to a more accurate and lower threshold. Another commenter noted the previous problems using changes in costs and recommended that CMS use a blend of the rates-of-increases for costs and charges to establish the charge inflation factor.

One commenter recommended that CMS keep the outlier threshold at $33,560 until CMS can determine the impact of using the most current cost-to-charge ratio during a full fiscal year. Other commenters also recommended that CMS eliminate any increase in the outlier threshold because the new outlier regulations will have a significant impact on Medicare outlier payments for FY 2004.

One commenter requested that CMS factor in the calculation of the threshold the fact that certain hospitals have distorted their charges significantly.

One commenter submitted a model of the outlier threshold for FY 2004 that incorporated the changes from the June 9, 2003 final rule. The commenter estimated the fixed-loss threshold to be $25,375 under these assumptions. The commenter also noted that the reconciliation process will reduce outlier payments and, accordingly, CMS should model a reduction in the outlier threshold to account for reconciliation, which would further lower the outlier threshold.

One commenter suggested that CMS lower the outlier threshold because independent studies strongly suggest that final FY 2003 outlier payments will fall short of the legislative mandate of 5 to 6 percent. Another commenter suggested that the outlier threshold remain at its current level because outlier payments for the first 3 months of FY 2003 represent 5.5 percent of total payments and, as a result, there does not seem to be any justification for such an increase. Another commenter explained that the transfer policy already reduces the payment to hospitals for short-stay cases and any increase in the outlier threshold will further penalize hospitals for treating high cost, medically complex cases.

Response: As described above, we are reflecting the changes made to outliers from the June 9, 2003 final rule. These changes have resulted in a substantial reduction in the outlier threshold from the proposed level. We estimate the outlier threshold would be approximately $50,200 without accounting for the effects of these changes. Therefore, the final threshold is 37 percent lower due to the changes described above. This reduction in the outlier threshold will allow hospitals that have been negatively impacted by the increase in the FY 2003 threshold due to those hospitals that maximized their outlier payments by dramatically increasing charges to qualify for higher outlier payments due to the lower threshold.

We are concerned that the outlier policy maintains its original intent to ensure hospitals are not significantly disadvantaged by unpredictable extraordinarily costly cases, and, therefore, we acted to close the loopholes in our prior policy through the final outlier rule. As a result of those changes, the threshold has fallen significantly from the proposed threshold.

Comment: Another commenter asked that any final outlier threshold included in the final rule be subject to a 60-day review and comment period.

Response: In the proposed rule, we noted that we would incorporate any final outlier policy changes in this final rule. We received many comments in response to the proposed rule, and we have considered them thoroughly in undertaking our analysis. Therefore, we do not believe there is any need for an additional public comment period on the changes. Accordingly, a fixed-loss threshold of $31,000 will be applied to calculate outlier payments for discharges occurring on or after October 1, 2003.

Comment: One commenter asked that CMS implement a transition period to protect those hospitals harmed by the significant changes in the June 9, 2003 final outlier rule. The commenter explained that a transition period is justified and would be consistent with previous transition methodologies employed for CMS changes, such as those proposed.

One commenter stated that any reconciliation would be inconsistent with the prospective nature of the IPPS.

Response: We responded to similar comments in the June 9, 2003 final rule on outliers (68 FR 34494). Therefore, we refer the commenters to that final rule.

Comment: Two commenters stated that the criterion in the final rule on outliers that specifically addressed our policy on reconciliation (that if a hospital's cost-to-charge ratio changed by 10 or more percentage points, a hospital would be subject to reconciliation) is flawed. The commenters believed that the criterion would tolerate vastly different rates of charge growth among hospitals, and hospitals with the lowest charges in relation to cost would be inappropriately subject to the greatest restriction in charge growth. The commenters provided an example where a hospital with a cost-to-charge ratio of .30 could mark up its charges by 50 percent in a 2-year period without triggering reconciliation, while another hospital with a cost-to-charge ratio of .80 would trigger reconciliation if charges grew by only 14 percent. The commenters recommended that, because of this inequity in this criterion, CMS modify the trigger for outlier reconciliation by promulgating a scale of cost-to-charge ratios rather than a constant amount. The scale could be based upon a rate of tolerable charge growth, which CMS would choose.

Response: We appreciate the suggestion by the commenters and will carefully evaluate the information provided by them. We note that fiscal intermediaries have discretion under the reconciliation policy to reconcile additional hospitals' cost reports based on analysis that indicates the outlier payments made to those hospitals are significantly inaccurate.

Comment: One commenter explained that one health care system agreed to accept reduced outlier payments during FY 2003. The commenter asked that this reduction be accounted for in the calculation of the threshold.

Response: Our calculation of the outlier threshold reflects the application of the outlier policies implemented by the June 9, 2003 final rule. The agreement referred to by the commenter was based upon the application of policies prior to that final rule. Therefore, it has no bearing on the calculation of the FY 2004 threshold described in this final rule.

Comment: One commenter noted that outlier payments are increasing because DRG payments are not keeping pace with the high cost of treatment. The commenter added that adjusting the outlier threshold will only add to the problem of underfunded health care and, because health care is not a priority, there will always be a struggle to pay for it. The commenter noted that there needs to be a determination of what care will be paid for, and then hospitals need to decide if they will provide the noncovered services.

Another commenter believed that the final rule on outliers would affect hospitals that have applied outlier payments appropriately. The commenter also believed that Medicare beneficiaries would be impacted as community hospitals shift care to more costly tertiary care facilities due to concerns about underpayment for potentially complex patient cases. The commenter explained that it is concerned that claims processing errors in the application of the outlier provision may result in underreporting of services provided, which will perpetuate underpayments to hospitals and lead to long-term ramifications on the integrity of the data generated by the IPPS.

Response: As discussed above, we lowered the outlier threshold in response to the new provisions on outliers. We anticipate that, as a result of the changes implemented by our June 5, 2003 final rule, outlier payments will be better targeted to truly high-cost cases. This will help alleviate the commenters' concerns.

ii. Other changes concerning outliers. As stated in the September 1, 1993 final rule (58 FR 46348), we establish outlier thresholds that are applicable to both hospital inpatient operating costs and hospital inpatient capital-related costs. When we modeled the combined operating and capital outlier payments, we found that using a common set of thresholds resulted in a higher percentage of outlier payments for capital-related costs than for operating costs. We project that the thresholds for FY 2004 will result in outlier payments equal to 5.1 percent of operating DRG payments and 4.8 percent of capital payments based on the Federal rate.

In accordance with section 1886(d)(3)(B), we reduced the FY 2004 standardized amounts by the same percentage to account for the projected proportion of payments paid to outliers. The outlier adjustment factors to be applied to the standardized amounts for FY 2004 are as follows:

We apply the outlier adjustment factors after removing the effects of the FY 2003 outlier adjustment factors on the standardized amounts.

To determine whether a case qualifies for outlier payments, we apply hospital-specific cost-to-charge ratios to the total covered charges for the case. Operating and capital costs for the case are calculated separately by applying separate operating and capital cost-to-charge ratios. These costs are then combined and compared with the fixed-loss outlier threshold.

The June 9, 2003 final rule eliminated the application of the statewide average for hospitals whose cost-to-charge ratios fall below 3 standard deviations from the national mean cost-to-charge ratio. However, for those hospitals for which the fiscal intermediary computes operating cost-to-charge ratios greater than 1.203 or capital cost-to-charge ratios greater than 0.163, or hospitals for whom the fiscal intermediary is unable to calculate a cost-to-charge ratio (as described at § 412.84(i)(3)), we are still using statewide average ratios to calculate costs to determine whether a hospital qualifies for outlier payments. Table 8A in section VI. of this Addendum contains the statewide average operating cost-to-charge ratios for urban hospitals and for rural hospitals for which the fiscal intermediary is unable to compute a hospital-specific cost-to-charge ratio within the above range. These statewide average ratios would replace the ratios published in the August 1, 2002 IPPS final rule (67 FR 50263). Table 8B in section VI. of this Addendum contains the comparable statewide average capital cost-to-charge ratios. Again, the cost-to-charge ratios in Tables 8A and 8B will be used during FY 2004 when hospital-specific cost-to-charge ratios based on the latest settled cost report are either not available or are outside the range noted above. iii. FY 2002 and FY 2003 outlier payments.

In the August 1, 2002 IPPS final rule (67 FR 50125), we stated that, based on available data, we estimated that actual FY 2002 outlier payments would be approximately 6.9 percent of actual total DRG payments. This estimate was computed based on simulations using the FY 2001 MedPAR file (discharge data for FY 2001 bills). That is, the estimate of actual outlier payments did not reflect actual FY 2002 bills but instead reflected the application of FY 2002 rates and policies to available FY 2001 bills.

Our current estimate, using available FY 2002 bills, is that actual outlier payments for FY 2002 were approximately 7.8 percent of actual total DRG payments. Thus, the data indicate that, for FY 2002, the percentage of actual outlier payments relative to actual total payments is higher than we projected before FY 2002 (and thus exceeds the percentage by which we reduced the standardized amounts for FY 2002). Nevertheless, consistent with the policy and statutory interpretation we have maintained since the inception of the IPPS, we do not plan to make retroactive adjustments to outlier payments to ensure that total outlier payments for FY 2002 are equal to 5.1 percent of total DRG payments.

We currently estimate that actual outlier payments for FY 2003 will be approximately 6.5 percent of actual total DRG payments, 1.4 percentage points higher than the 5.1 percent we projected in setting outlier policies for FY 2003. This estimate is based on simulations using the FY 2002 MedPAR file (discharge data for FY 2002 bills). We used these data to calculate an estimate of the actual outlier percentage for FY 2003 by applying FY 2003 rates and policies including an outlier threshold of $33,560 to available FY 2002 bills. This estimate does not reflect the outlier policy changes implemented in the June 9, 2003 final rule that will become effective on August 8, 2003. Due to the limited time remaining in FY 2003 during which these changes will be effective, we do not anticipate that these changes will substantially affect our estimate.

5. FY 2004 Standardized Amounts

The adjusted standardized amounts are divided into labor and nonlabor portions. Table 1A in section VI. of this Addendum contains the two national standardized amounts that we will be applying to all hospitals, except hospitals in Puerto Rico. As described in section II.A.1. of this Addendum, we are not revising the labor share of the national standardized amount from 71.1 percent.

The following table illustrates the changes from the FY 2003 national average standardized amounts. The first row in the table shows the updated (through FY 2003) average standardized amounts after restoring the FY 2003 offsets for outlier payments and geographic reclassification budget neutrality. The DRG reclassification and recalibration and wage index budget neutrality factor is cumulative. Therefore, the FY 2003 factor is not removed from the amounts in the table.

Under section 1886(d)(9)(A)(ii) of the Act, the Federal portion of the Puerto Rico payment rate is based on the discharge-weighted average of the national large urban standardized amount and the national other standardized amount (as set forth in Table 1A). The labor and nonlabor portions of the national average standardized amounts for Puerto Rico hospitals are set forth in Table 1C of section VI. of this Addendum. This table also includes the Puerto Rico standardized amounts. The labor share applied to the Puerto Rico standardized amount is 71.3 percent.

B. Adjustments for Area Wage Levels and Cost-of-Living

Tables 1A and 1C, as set forth in section VI. of this Addendum, contain the labor-related and nonlabor-related shares that we used to calculate the prospective payment rates for hospitals located in the 50 States, the District of Columbia, and Puerto Rico. This section addresses two types of adjustments to the standardized amounts that are made in determining the prospective payment rates as described in this Addendum.

1. Adjustment for Area Wage Levels

Sections 1886(d)(3)(E) and 1886(d)(9)(C)(iv) of the Act require that we make an adjustment to the labor-related portion of the national and Puerto Rico prospective payment rates, respectively, to account for area differences in hospital wage levels. This adjustment is made by multiplying the labor-related portion of the adjusted standardized amounts by the appropriate wage index for the area in which the hospital is located. In section III. of the preamble to this final rule, we discuss the data and methodology for the FY 2004 wage index. The FY 2004 wage index is set forth in Tables 4A, 4B, 4C, and 4F of section VI. of this Addendum.

2. Adjustment for Cost-of-Living in Alaska and Hawaii

Section 1886(d)(5)(H) of the Act authorizes an adjustment to take into account the unique circumstances of hospitals in Alaska and Hawaii. Higher labor-related costs for these two States are taken into account in the adjustment for area wages described above. For FY 2004, we are adjusting the payments for hospitals in Alaska and Hawaii by multiplying the nonlabor portion of the standardized amounts by the appropriate adjustment factor contained in the table below.

(The above factors are based on data obtained from the U.S. Office of Personnel Management.)

C. DRG Relative Weights

As discussed in section II. of the preamble, we have developed a classification system for all hospital discharges, assigning them into DRGs, and have developed relative weights for each DRG that reflect the resource utilization of cases in each DRG relative to Medicare cases in other DRGs. Table 5 of section VI. of this Addendum contains the relative weights that we are using for discharges occurring in FY 2004. These factors have been recalibrated as explained in section II. of the preamble of this final rule.

D. Calculation of Prospective Payment Rates for FY 2004

General Formula for Calculation of Prospective Payment Rates for FY 2004

The operating prospective payment rate for all hospitals paid under the IPPS located outside of Puerto Rico, except SCHs and MDHs, equals the Federal rate based on the amounts in Table 1A in section VI. of this Addendum.

The prospective payment rate for SCHs equals the higher of the applicable Federal rate from Table 1A or the hospital-specific rate as described below. The prospective payment rate for MDHs equals the higher of the Federal rate, or the Federal rate plus 50 percent of the difference between the Federal rate and the hospital-specific rate as described below. The prospective payment rate for Puerto Rico equals 50 percent of the Puerto Rico rate plus 50 percent of the national rate from Table 1C in section VI. of this Addendum.

1. Federal Rate

For discharges occurring on or after October 1, 2003 and before October 1, 2004, except for SCHs, MDHs, and hospitals in Puerto Rico, payment under the IPPS is based exclusively on the Federal rate.

The Federal rate is determined as follows:

Step 1—Select the appropriate average standardized amount considering the location of the hospital (large urban or other) (see Table 1A in section VI. of this Addendum).

Step 2—Multiply the labor-related portion of the standardized amount by the applicable wage index for the geographic area in which the hospital is located or the area to which the hospital is reclassified (see Tables 4A, 4B, and 4C of section VI. of this Addendum).

Step 3—For hospitals in Alaska and Hawaii, multiply the nonlabor-related portion of the standardized amount by the appropriate cost-of-living adjustment factor.

Step 4—Add the amount from Step 2 and the nonlabor-related portion of the standardized amount (adjusted, if appropriate, under Step 3).

Step 5—Multiply the final amount from Step 4 by the relative weight corresponding to the appropriate DRG (see Table 5 of section VI. of this Addendum).

The Federal rate as determined in Step 5 may then be further adjusted if the hospital qualifies for either the IME or DSH adjustment.

2. Hospital-Specific Rate (Applicable Only to SCHs and MDHs)

a. Calculation of Hospital-Specific Rate

Section 1886(b)(3)(C) of the Act provides that SCHs are paid based on whichever of the following rates yields the greatest aggregate payment: the Federal rate; the updated hospital-specific rate based on FY 1982 costs per discharge; the updated hospital-specific rate based on FY 1987 costs per discharge; or the updated hospital-specific rate based on FY 1996 costs per discharge.

Section 1886(d)(5)(G) of the Act provides that MDHs are paid based on whichever of the following rates yields the greatest aggregate payment: the Federal rate or the Federal rate plus 50 percent of the difference between the Federal rate and the greater of the updated hospital-specific rates based on either FY 1982 or FY 1987 costs per discharge. MDHs do not have the option to use their FY 1996 hospital-specific rate.

Hospital-specific rates have been determined for each of these hospitals based on either the FY 1982 costs per discharge, the FY 1987 costs per discharge or, for SCHs, the FY 1996 costs per discharge. For a more detailed discussion of the calculation of the hospital-specific rates, we refer the reader to the September 1, 1983 interim final rule (48 FR 39772); the April 20, 1990 final rule with comment (55 FR 15150); the September 4, 1990 final rule (55 FR 35994); and the August 1, 2000 final rule (65 FR 47082). In addition, for both SCHs and MDHs, the hospital-specific rate is adjusted by the budget neutrality adjustment factor (that is, by 1.005522) as discussed in section II.A.4.a. of this Addendum. The resulting rate was used in determining the payment rate an SCH or MDH will receive for its discharges beginning on or after October 1, 2003.

b. Updating the FY 1982, FY 1987, and FY 1996 Hospital-Specific Rates for FY 2004

We are increasing the hospital-specific rates by 3.4 percent (the hospital market basket percentage) for SCHs and MDHs for FY 2004. Section 1886(b)(3)(C)(iv) of the Act provides that the update factor applicable to the hospital-specific rates for SCHs is equal to the update factor provided under section 1886(b)(3)(B)(iv) of the Act, which, for SCHs in FY 2004, is the market basket rate of increase. Section 1886(b)(3)(D) of the Act provides that the update factor applicable to the hospital-specific rates for MDHs also equals the update factor provided under section 1886(b)(3)(B)(iv) of the Act, which, for FY 2004, is the market basket rate.

3. General Formula for Calculation of Prospective Payment Rates for Hospitals Located in Puerto Rico Beginning on or After October 1, 2003 and Before October 1, 2004

a. Puerto Rico Rate

The Puerto Rico prospective payment rate is determined as follows:

Step 1—Select the appropriate adjusted average standardized amount considering the large urban or other designation of the hospital (see Table 1C of section VI. of the Addendum).

Step 2—Multiply the labor-related portion of the standardized amount by the appropriate Puerto Rico-specific wage index (see Table 4F of section VI. of the Addendum).

Step 3—Add the amount from Step 2 and the nonlabor-related portion of the standardized amount.

Step 4—Multiply the result in Step 3 by 50 percent.

Step 5—Multiply the amount from Step 4 by the appropriate DRG relative weight (see Table 5 of section VI. of the Addendum).

b. National Rate

The national prospective payment rate is determined as follows:

Step 1—Multiply the labor-related portion of the national average standardized amount (see Table 1C of section VI. of the Addendum) by the appropriate national wage index (see Tables 4A and 4B of section VI. of the Addendum).

Step 2—Add the amount from Step 1 and the nonlabor-related portion of the national average standardized amount.

Step 3—Multiply the result in Step 2 by 50 percent.

Step 4—Multiply the amount from Step 3 by the appropriate DRG relative weight (see Table 5 of section VI. of the Addendum).

The sum of the Puerto Rico rate and the national rate computed above equals the prospective payment for a given discharge for a hospital located in Puerto Rico. This rate may then be further adjusted if the hospital qualifies for either the IME or DSH adjustment.

III. Changes to Payment Rates for Acute Care Hospital Inpatient Capital-Related Costs for FY 2004

The PPS for acute care hospital inpatient capital-related costs was implemented for cost reporting periods beginning on or after October 1, 1991. Effective with that cost reporting period and during a 10-year transition period extending through FY 2001, acute care hospital inpatient capital-related costs were paid on the basis of an increasing proportion of the capital PPS Federal rate and a decreasing proportion of a hospital's historical costs for capital.

The basic methodology for determining Federal capital prospective rates is set forth in regulations at §§ 412.308 through 412.352. Below we discuss the factors that we used to determine the capital Federal rate for FY 2004, which will be effective for discharges occurring on or after October 1, 2003. The 10-year transition period ended with hospital cost reporting periods beginning on or after October 1, 2001 (FY 2002). Therefore, for cost reporting periods beginning in FY 2002, all hospitals (except “new” hospitals under §§ 412.304(c)(2) and 412.324(b)) are paid based on 100 percent of the capital Federal rate.

For FY 1992, we computed the standard Federal payment rate for capital-related costs under the IPPS by updating the FY 1989 Medicare inpatient capital cost per case by an actuarial estimate of the increase in Medicare inpatient capital costs per case. Each year after FY 1992, we update the capital standard Federal rate, as provided in § 412.308(c)(1), to account for capital input price increases and other factors. Section 412.308(c)(2) provides that the capital Federal rate is adjusted annually by a factor equal to the estimated proportion of outlier payments under the capital Federal rate to total capital payments under the capital Federal rate. In addition, § 412.308(c)(3) requires that the capital Federal rate be reduced by an adjustment factor equal to the estimated proportion of payments for (regular and special) exception under § 412.348. Section 412.308(c)(4)(ii) requires that the capital standard Federal rate be adjusted so that the annual DRG reclassification and the recalibration of DRG weights and changes in the geographic adjustment factor are budget neutral.

For FYs 1992 through 1995, § 412.352 required that the capital Federal rate also be adjusted by a budget neutrality factor so that aggregate payments for inpatient hospital capital costs were projected to equal 90 percent of the payments that would have been made for capital-related costs on a reasonable cost basis during the fiscal year. That provision expired in FY 1996. Section 412.308(b)(2) describes the 7.4 percent reduction to the capital rate that was made in FY 1994, and § 412.308(b)(3) describes the 0.28 percent reduction to the capital rate made in FY 1996 as a result of the revised policy of paying for transfers. In FY 1998, we implemented section 4402 of Pub. L. 105-33, which requires that, for discharges occurring on or after October 1, 1997, and before October 1, 2002, the unadjusted capital standard Federal rate is reduced by 17.78 percent. As we discussed in the August 1, 2002 IPPS final rule (67 FR 50102) and implemented in § 412.308(b)(6)), a small part of that reduction was restored effective October 1, 2002.

To determine the appropriate budget neutrality adjustment factor and the regular exceptions payment adjustment during the 10-year transition period, we developed a dynamic model of Medicare inpatient capital-related costs, that is, a model that projected changes in Medicare inpatient capital-related costs over time. With the expiration of the budget neutrality provision, the capital cost model was only used to estimate the regular exceptions payment adjustment and other factors during the transition period. As we explained in the August 1, 2001 IPPS final rule (66 FR 39911), beginning in FY 2003, an adjustment for regular exception payments is no longer necessary because regular exception payments were only made for cost reporting periods beginning on or after October 1, 1991, and before October 1, 2001 (see § 412.348(b)). Since payments are no longer being made under the regular exception policy in FY 2003 and after, we no longer use the capital cost model. The capital cost model and its application during the transition period are described in Appendix B of the August 1, 2001 IPPS final rule (66 FR 40099).

In accordance with section 1886(d)(9)(A) of the Act, under the IPPS for acute care hospital operating costs, hospitals located in Puerto Rico are paid for operating costs under a special payment formula. Prior to FY 1998, hospitals in Puerto Rico were paid a blended capital rate that consisted of 75 percent of the applicable standardized amount specific to Puerto Rico hospitals and 25 percent of the applicable national average standardized amount. However, effective October 1, 1997, as a result of section 4406 of Pub. L. 105-33, operating payments to hospitals in Puerto Rico are based on a blend of 50 percent of the applicable standardized amount specific to Puerto Rico hospitals and 50 percent of the applicable national average standardized amount. In conjunction with this change to the operating blend percentage, effective with discharges on or after October 1, 1997, we compute capital payments to hospitals in Puerto Rico based on a blend of 50 percent of the Puerto Rico capital rate and 50 percent of the capital Federal rate.

Section 412.374 provides for the use of this blended payment system for payments to Puerto Rico hospitals under the PPS for acute care hospital inpatient capital-related costs. Accordingly, for capital-related costs, we compute a separate payment rate specific to Puerto Rico hospitals using the same methodology used to compute the national Federal rate for capital.

A. Determination of Federal Hospital Inpatient Capital-Related Prospective Payment Rate Update

In the final IPPS rule published in the Federal Register on August 1, 2002 (67 FR 50127), we established a capital Federal rate of $407.01 for FY 2003. Section 402(b) of Pub. L. 108-7 requires that, effective for discharges occurring on or after April 1, 2003, and before October 1, 2003, the capital Federal rate for operating costs for all IPPS hospitals is based on the large urban standardized amount. However, under current law for discharges occurring on or after October 1, 2003, the capital Federal rate will again be calculated based on separate average standardized amounts for hospitals in large urban areas and for hospitals in other areas. In addition, a correction notice to the FY 2003 final IPPS rule issued in the Federal Register on April 25, 2003 (68 FR 22272) contains corrections and revisions to the wage index and geographic adjustment factor (GAF). In conjunction with the change to the operating PPS standardized amounts made by Pub. L. 108-7 and the wage index and GAF corrections, we have established a capital PPS standard Federal rate of $406.93 effective for discharges occurring on or after April 1, 2003 through September 30, 2003. As we discussed in the May 19, 2003 proposed rule (68 FR 27238), the capital rates effective for discharges occurring on or after April 1, 2003 through September 30, 2003, were used in determining the final FY 2004 capital rates. As a result of the changes to the factors used to establish the capital Federal rate that are explained in this Addendum, the FY 2004 capital standard Federal rate is $415.47.

In the discussion that follows, we explain the factors that were used to determine the FY 2004 capital Federal rate. In particular, we explain why the FY 2004 capital Federal rate has increased 2.10 percent compared to the FY 2003 capital Federal rate (effective for discharges occurring on or after April 1, 2003 through September 30, 2003). We also estimate aggregate capital payments will increase by 1.4 percent during this same period. This increase is primarily due to the increase in the number of hospital admissions and the increase in case-mix. This increase in capital payments is slightly less than last year (5.81 percent), mostly due to the restoration of the 2.1 percent reduction to the capital Federal rate in FY 2003 (§ 412.308(b)(6)) and the projected decrease in outlier payments as a result of the IPPS outlier policy established in the June 9, 2003 high-cost outlier final rule (68 FR 34494).

Total payments to hospitals under the IPPS are relatively unaffected by changes in the capital prospective payments. Since capital payments constitute about 10 percent of hospital payments, a 1-percent change in the capital Federal rate yields only about 0.1 percent change in actual payments to hospitals. Aggregate payments under the capital PPS are estimated to increase in FY 2004 compared to FY 2003.

1. Capital Standard Federal Rate Update

a. Description of the Update Framework

Under § 412.308(c)(1), the capital standard Federal rate is updated on the basis of an analytical framework that takes into account changes in a capital input price index (CIPI) and several other policy adjustment factors. Specifically, we have adjusted the projected CIPI rate of increase as appropriate each year for case-mix index-related changes, for intensity, and for errors in previous CIPI forecasts. In the May 19, 2003 proposed rule (68 FR 27239), we proposed an update factor of 0.7 for FY 2004 under that framework based on the best data available at that time. Under that same update framework based on more recent data, the final update factor for FY 2004 is 0.7 percent. This final update factor is based on a 0.7 percent increase in the CIPI, a 0.0 percent adjustment for intensity, a 0.0 percent adjustment for case-mix, a 0.0 percent adjustment for the FY 2002 DRG reclassification and recalibration, and a forecast error correction of 0.0 percent. We explain the basis for the FY 2004 CIPI projection in section III.C. of this Addendum. Below we describe the policy adjustments that have been applied.

The case-mix index is the measure of the average DRG weight for cases paid under the IPPS. Because the DRG weight determines the prospective payment for each case, any percentage increase in the case-mix index corresponds to an equal percentage increase in hospital payments.

The case-mix index can change for any of several reasons:

• The average resource use of Medicare patients changes (“real” case-mix change);
• Changes in hospital coding of patient records result in higher weight DRG assignments (“coding effects”); and
• The annual DRG reclassification and recalibration changes may not be budget neutral (“reclassification effect”).

We define real case-mix change as actual changes in the mix (and resource requirements) of Medicare patients as opposed to changes in coding behavior that result in assignment of cases to higher weighted DRGs but do not reflect higher resource requirements. In the update framework for the PPS for operating costs, we adjust the update upwards to allow for real case-mix change, but remove the effects of coding changes on the case-mix index. We also remove the effect on total payments of prior year changes to the DRG classifications and relative weights, in order to retain budget neutrality for all case-mix index-related changes other than patient severity. (For example, we adjusted for the effects of the FY 2002 DRG reclassification and recalibration as part of our update for FY 2004.) We have adopted this case-mix index adjustment in the capital update framework as well.

For FY 2004, we are projecting a 1.0 percent total increase in the case-mix index. We estimate that real case-mix increase will equal 1.0 percent in FY 2004. Therefore, the net adjustment for case-mix change in FY 2004 is 0.0 percentage points.

We estimate that FY 2002 DRG reclassification and recalibration will result in a 0.0 percent change in the case-mix when compared with the case-mix index that would have resulted if we had not made the reclassification and recalibration changes to the DRGs. Therefore, we are making a 0.0 percent adjustment for DRG reclassification and recalibration in the update for FY 2004 to maintain budget neutrality.

The capital update framework contains an adjustment for forecast error. The input price index forecast is based on historical trends and relationships ascertainable at the time the update factor is established for the upcoming year. In any given year, there may be unanticipated price fluctuations that may result in differences between the actual increase in prices and the forecast used in calculating the update factors. In setting a prospective payment rate under the framework, we make an adjustment for forecast error only if our estimate of the change in the capital input price index for any year is off by 0.25 percentage points or more. There is a 2-year lag between the forecast and the measurement of the forecast error. A forecast error of 0.2 percentage points was calculated for the FY 2002 update. That is, current historical data indicate that the forecasted FY 2002 CIPI used in calculating the FY 2002 update factor (0.7 percent) overstated the actual realized price increases (0.5 percent) by 0.2 percentage points. This slight overprediction was mostly due to an underestimation of the interest rate cuts by the Federal Reserve Board in 2002, which impacted the interest component of the CIPI. However, since this estimation of the change in the CIPI is less than 0.25 percentage points, it is not reflected in the update recommended under this framework. Therefore, we are making a 0.0 percent adjustment for forecast error in the update for FY 2004.

Under the capital PPS system framework, we also make an adjustment for changes in intensity. We calculate this adjustment using the same methodology and data that are used in the framework for the operating PPS. The intensity factor for the operating update framework reflects how hospital services are utilized to produce the final product, that is, the discharge. This component accounts for changes in the use of quality-enhancing services, for changes in within-DRG severity, and for expected modification of practice patterns to remove noncost-effective services.

We calculate case-mix constant intensity as the change in total charges per admission, adjusted for price level changes (the CPI for hospital and related services) and changes in real case-mix. The use of total charges in the calculation of the intensity factor makes it a total intensity factor, that is, charges for capital services are already built into the calculation of the factor. Therefore, we have incorporated the intensity adjustment from the operating update framework into the capital update framework. Without reliable estimates of the proportions of the overall annual intensity increases that are due, respectively, to ineffective practice patterns and to the combination of quality-enhancing new technologies and within-DRG complexity, we assume, as in the operating update framework, that one-half of the annual increase is due to each of these factors. The capital update framework thus provides an add-on to the input price index rate of increase of one-half of the estimated annual increase in intensity, to allow for within-DRG severity increases and the adoption of quality-enhancing technology.

As we discussed in the May 19, 2003 proposed rule (68 FR 27239), we have developed a Medicare-specific intensity measure based on a 5-year average. Past studies of case-mix change by the RAND Corporation (“Has DRG Creep Crept Up? Decomposing the Case Mix Index Change Between 1987 and 1988” by G. M. Carter, J. P. Newhouse, and D. A. Relles, R-4098-HCFA/ProPAC (1991)) suggest that real case-mix change was not dependent on total change, but was usually a fairly steady 1.0 to 1.4 percent per year. We use 1.4 percent as the upper bound because the RAND study did not take into account that hospitals may have induced doctors to document medical records more completely in order to improve payment.

We calculate case-mix constant intensity as the change in total charges per admission, adjusted for price level changes (the CPI for hospital and related services), and changes in real case-mix. As we noted above, in accordance with § 412.308(c)(1)(ii), we began updating the capital standard Federal rate in FY 1996 using an update framework that takes into account, among other things, allowable changes in the intensity of hospital services. For FYs 1996 through 2001, we found that case-mix constant intensity was declining and we established a 0.0 percent adjustment for intensity in each of those years. For FYs 2001 and 2002, we found that case-mix constant intensity was increasing and we established a 0.3 percent adjustment and 1.0 percent adjustment for intensity, respectively.

Using the methodology described above, as we discussed in the May 19, 2003 proposed rule (68 FR 27239), for FY 2004 we examined the change in total charges per admission, adjusted for price level changes (the CPI for hospital and related services), and changes in real case-mix for FYs 1998 though 2002. We found that, over this period and in particular the last 3 years of this period (FYs 2000 through 2002), the charge data appear to be skewed. More specifically, we found a dramatic increase in hospital charges for FYs 2000 through 2002 without a corresponding increase in hospital case-mix index. If hospitals were treating new or different types of cases, which would result in an appropriate increase in charges per discharge, then we would expect hospitals' case-mix to increase proportionally.

The timing of this increase in charge growth is consistent with the dramatic increase in charges that we discussed in the June 9, 2003 high-cost outlier final rule (68 FR 34494). As we discussed in that final rule, because hospitals have the ability to increase their outlier payments through dramatic charge increases, we have made several changes in our high-cost outlier policy at §§ 412.84(i) and (m) in order to prevent hospitals from taking advantage of our current outlier policy.

As discussed above, our intensity calculation relies heavily upon charge data and we believe that this charge data may be inappropriately skewed. Therefore, in the May 19, 2003 proposed rule (68 FR 22739), we proposed a 0.0 percent adjustment for intensity for FY 2004. As we explained in that same proposed rule, in past FYs (1996 through 2000) when we found intensity to be declining, we believed a zero (rather then negative) intensity adjustment was appropriate. Similarly, we believe that it is appropriate to apply a zero intensity adjustment for FY 2004 until we believe that any increase in charges can be tied to intensity rather then to attempts to maximize outlier payments. We received no comments on our proposed 0.0 percent adjustment for intensity. Therefore, in this final rule, we are making a 0.0 percent adjustment for intensity in the update for FY 2004.

Above we described the basis of the components used to develop the 0.7 percent final capital update factor for FY 2004 as shown in the table below.

b. Comparison of CMS and MedPAC Update Recommendation

In the past, MedPAC has included update recommendations for capital PPS in a Report to Congress. As we discussed in the May 19, 2003 proposed rule (68 FR 27240), in its March 2003 Report to Congress, MedPAC did not make an update recommendation for capital PPS payments. However, in that same report, MedPAC made an update recommendation for hospital inpatient and outpatient services (page 4). MedPAC stated that hospital inpatient and outpatient services should be considered together because they are so closely interrelated. Their recommendation is based on an assessment of whether payments are adequate to cover the costs of efficient providers, an estimate of input price inflation (measured by the market basket index), and an adjustment for technological charges, which is offset by reasonable expectations in productivity gains.

2. Outlier Payment Adjustment Factor

Section 412.312(c) establishes a unified outlier methodology for inpatient operating and inpatient capital-related costs. A single set of thresholds is used to identify outlier cases for both inpatient operating and inpatient capital-related payments. Section 412.308(c)(2) provides that the standard Federal rate for inpatient capital-related costs be reduced by an adjustment factor equal to the estimated proportion of capital related outlier payments to total inpatient capital-related PPS payments. The outlier thresholds are set so that operating outlier payments are projected to be 5.1 percent of total operating DRG payments.

In the August 1, 2002 IPPS final rule (67 FR 50129), we estimated that outlier payments for capital in FY 2003 would equal 5.31 percent of inpatient capital-related payments based on the FY 2003 capital Federal rate. Accordingly, we applied an outlier adjustment factor of 0.9469 to the FY 2003 capital Federal rate. Based on the thresholds as set forth in section II.A.4.c. of this Addendum, we estimate that outlier payments for capital will equal 4.79 percent of inpatient capital-related payments based on the capital Federal rate in FY 2004. Therefore, we are establishing an outlier adjustment factor of 0.9521 to the capital Federal rate. Thus, the percentage of capital outlier payments to total capital standard payments for FY 2004 is lower than the percentage for FY 2003. This projected decrease in capital outlier payments is mostly due to the changes in the IPPS outlier policy established in the June 9, 2003 high-cost outlier final rule (68 FR 34494).

The outlier reduction factors are not built permanently into the capital rates; that is, they are not applied cumulatively in determining the capital Federal rate. Therefore, the net change in the outlier adjustment to the capital Federal rate for FY 2004 is 1.0055 (0.9521/0.9469). The outlier adjustment increases the FY 2004 capital Federal rate by 0.55 percent compared with the FY 2003 outlier adjustment.

3. Budget Neutrality Adjustment Factor for Changes in DRG Classifications and Weights and the Geographic Adjustment Factor

Section 412.308(c)(4)(ii) requires that the capital Federal rate be adjusted so that aggregate payments for the fiscal year based on the capital Federal rate after any changes resulting from the annual DRG reclassification and recalibration and changes in the geographic adjustment factor (GAF) are projected to equal aggregate payments that would have been made on the basis of the capital Federal rate without such changes.

Since we implemented a separate geographic adjustment factor for Puerto Rico, we apply separate budget neutrality adjustments for the national geographic adjustment factor and the Puerto Rico geographic adjustment factor. We apply the same budget neutrality factor for DRG reclassifications and recalibration nationally and for Puerto Rico. Separate adjustments were unnecessary for FY 1998 and earlier since the geographic adjustment factor for Puerto Rico was implemented in FY 1998.

In the past, we used the actuarial capital cost model (described in Appendix B of the August 1, 2001 IPPS final rule (66 FR 40099)) to estimate the aggregate payments that would have been made on the basis of the capital Federal rate with and without changes in the DRG classifications and weights and in the GAF to compute the adjustment required to maintain budget neutrality for changes in DRG weights and in the GAF. During the transition period, the capital cost model was also used to estimate the regular exception payment adjustment factor. As we explain in section III.A.4. of this Addendum, beginning in FY 2003 an adjustment for regular exception payments is no longer necessary. Therefore, we are no longer using the capital cost model. Instead, we are using historical data based on hospitals' actual cost experiences to determine the exceptions payment adjustment factor for special exceptions payments.

To determine the factors for FY 2004, we compared (separately for the national capital rate and the Puerto Rico capital rate) estimated aggregate capital Federal rate payments based on the FY 2003 DRG relative weights and the FY 2003 GAF to estimated aggregate capital Federal rate payments based on the FY 2004 relative weights and the FY 2004 GAF. In the August 1, 2002 IPPS final rule (67 FR 50129) for FY 2003, the budget neutrality adjustment factors were 0.9885 for the national capital rate and 0.9963 for the Puerto Rico capital rate. As a result of the revisions to the GAF effective for discharges occurring on or after April 1, 2003 through September 30, 2003, the budget neutrality adjustment factor is 0.9983 for the national capital rate for discharges occurring on or before April 1, 2003 through September 30, 2003. The budget neutrality adjustment factor for the Puerto Rico capital rate remained unchanged (0.9963). As we noted above, the capital rates effective for discharges occurring on or after April 1, 2003 through September 30, 2003 were used in determining the FY 2004 capital rates. In making the comparison, we set the regular and special exceptions reduction factors to 1.00.

To achieve budget neutrality for the changes in the national GAF, based on calculations using updated data, we are applying an incremental budget neutrality adjustment of 1.0051 for FY 2004 to the previous cumulative FY 2003 adjustment (0.9883), yielding a cumulative adjustment of 0.9933 through FY 2004. For the Puerto Rico GAF, we are applying an incremental budget neutrality adjustment of 1.0002 for FY 2004 to the previous cumulative FY 2003 adjustment (0.9963), yielding a cumulative adjustment of 0.9965 through FY 2004.

We then compared estimated aggregate capital Federal rate payments based on the FY 2003 DRG relative weights and the FY 2003 GAF to estimated aggregate capital Federal rate payments based on the FY 2004 DRG relative weights and the FY 2004 GAF. The incremental adjustment for DRG classifications and changes in relative weights is 1.0008 both nationally and for Puerto Rico. The cumulative adjustments for DRG classifications and changes in relative weights and for changes in the GAF through FY 2004 are 0.9941 nationally and 0.9973 for Puerto Rico. The following table summarizes the adjustment factors for each fiscal year:

The methodology used to determine the recalibration and geographic (DRG/GAF) budget neutrality adjustment factor for FY 2004 is similar to that used in establishing budget neutrality adjustments under the PPS for operating costs. One difference is that, under the operating PPS, the budget neutrality adjustments for the effect of geographic reclassifications are determined separately from the effects of other changes in the hospital wage index and the DRG relative weights. Under the capital PPS, there is a single DRG/GAF budget neutrality adjustment factor (the national capital rate and the Puerto Rico capital rate are determined separately) for changes in the GAF (including geographic reclassification) and the DRG relative weights. In addition, there is no adjustment for the effects that geographic reclassification has on the other payment parameters, such as the payments for serving low-income patients, indirect medical education payments, or the large urban add-on payments.

In the August 1, 2002 IPPS final rule (67 FR 50129), we calculated a GAF/DRG budget neutrality factor of 0.9957 for FY 2003. As we noted above, as a result of the revisions to the GAF effective for discharges occurring on or after April 1, 2003 through September 30, 2003 published in the Federal Register on April 25, 2003 (68 FR 22272), we calculated a GAF/DRG budget neutrality factor of 0.9956 for discharges occurring on or after April 1, 2003 through September 30, 2003. Furthermore, as noted above, the capital rates effective for discharges occurring on or after April 1, 2003 through September 30, 2003 were used in determining the FY 2004 capital rates.

In the May 19, 2003 proposed rule (68 FR 27241), for FY 2004 we calculated a GAF/DRG budget neutrality factor of 1.0038. For this final rule, based on updated data, we are establishing a GAF/DRG budget neutrality factor of 1.0059 for FY 2004. The GAF/DRG budget neutrality factors are built permanently into the capital rates; that is, they are applied cumulatively in determining the capital Federal rate. This follows from the requirement that estimated aggregate payments each year be no more or less than they would have been in the absence of the annual DRG reclassification and recalibration and changes in the GAF. The incremental change in the adjustment from FY 2003 to FY 2004 is 1.0059. The cumulative change in the capital Federal rate due to this adjustment is 0.9941 (the product of the incremental factors for FY 1993, FY 1994, FY 1995, FY 1996, FY 1997, FY 1998, FY 1999, FY 2000, FY 2001, FY 2002, FY 2003, and the incremental factor for FY 2004: 0.9980 × 1.0053 × 0.9998 × 0.9994 × 0.9987 × 0.9989 × 1.0028 × 0.9985 × 0.9979 × 0.9934 × 0.9956 × 1.0059 = 0.9941).

This factor accounts for DRG reclassifications and recalibration and for changes in the GAF. It also incorporates the effects on the GAF of FY 2004 geographic reclassification decisions made by the MGCRB compared to FY 2003 decisions. However, it does not account for changes in payments due to changes in the DSH and IME adjustment factors or in the large urban add-on.

4. Exceptions Payment Adjustment Factor

Section 412.308(c)(3) requires that the capital standard Federal rate be reduced by an adjustment factor equal to the estimated proportion of additional payments for both regular exceptions and special exceptions under § 412.348 relative to total capital PPS payments. In estimating the proportion of regular exception payments to total capital PPS payments during the transition period, we used the actuarial capital cost model originally developed for determining budget neutrality (described in Appendix B of the August 1, 2001 IPPS final rule (66 FR 40099)) to determine the exceptions payment adjustment factor, which was applied to both the Federal and hospital-specific capital rates.

An adjustment for regular exception payments is no longer necessary in determining the FY 2004 capital Federal rate because, in accordance with § 412.348(b), regular exception payments were only made for cost reporting periods beginning on or after October 1, 1991 and before October 1, 2001. Accordingly, as we explained in the August 1, 2001 IPPS final rule (66 FR 39949), in FY 2003 and subsequent fiscal years, no payments will be made under the regular exceptions provision. However, in accordance with § 412.308(c), we still need to compute a budget neutrality adjustment for special exception payments under § 412.348(g). We describe our methodology for determining the special exceptions adjustment used in calculating the FY 2004 capital Federal rate below.

Under the special exceptions provision specified at § 412.348(g)(1), eligible hospitals include SCHs, urban hospitals with at least 100 beds that have a disproportionate share percentage of at least 20.2 percent or qualify for DSH payments under § 412.106(c)(2), and hospitals with a combined Medicare and Medicaid inpatient utilization of at least 70 percent. An eligible hospital may receive special exceptions payments if it meets (1) a project need requirement as described at § 412.348(g)(2), which, in the case of certain urban hospitals, includes an excess capacity test as described at § 412.348(g)(4); (2) an age of assets test as described at § 412.348(g)(3); and (3) a project size requirement as described at § 412.348(g)(5).

As we explained in the August 1, 2001 IPPS final rule (66 FR 39912-39914), in order to determine the estimated proportion of special exceptions payments to total capital payments, we attempted to identify the universe of eligible hospitals that may potentially qualify for special exceptions payments. First, we identified hospitals that met the eligibility requirements at § 412.348(g)(1). Then we determined each hospital's average fixed asset age in the earliest available cost report starting in FY 1992 and subsequent fiscal years. For each of those hospitals, we calculated the average fixed asset age by dividing the accumulated depreciation by the current year's depreciation. In accordance with § 412.348(g)(3), a hospital must have an average age of buildings and fixed assets above the 75th percentile of all hospitals in the first year of the capital PPS. In the September 1, 1994 final rule (59 FR 45385), we stated that, based on the June 1994 update of the cost report files in HCRIS, the 75th percentile for buildings and fixed assets for FY 1992 was 16.4 years. However, we noted that we would make a final determination of that value on the basis of more complete cost report information at a later date. In the August 29, 1997 final rule (62 FR 46012), based on the December 1996 update of HCRIS and the removal of outliers, we finalized the 75th percentile for buildings and fixed assets for FY 1992 as 15.4 years. Thus, we eliminated any hospitals from the potential universe of hospitals that may qualify for special exception payments if its average age of fixed assets did not exceed 15.4 years.

For the hospitals remaining in the potential universe, we estimated project-size by using the fixed capital acquisitions shown on Worksheet A7 from the following HCRIS cost reports updated through March 2003.

Because the project phase-in may overlap 2 cost reporting years, we added together the fixed acquisitions from sequential pairs of cost reports to determine project size. Under § 412.348(g)(5), the hospital's project cost must be at least $200 million or 100 percent of its operating cost during the first 12-month cost reporting period beginning on or after October 1, 1991. We calculated the operating costs from the earliest available cost report starting in FY 1992 and later by subtracting inpatient capital costs from inpatient costs (for all payers). We did not subtract the direct medical education costs as those costs are not available on every update of the HCRIS minimum data set. If the hospital met the project size requirement, we assumed that it also met the project need requirements at § 412.348(g)(2) and the excess capacity test for urban hospitals at § 412.348(g)(4).

Because we estimate that so few hospitals will qualify for special exceptions, projecting costs, payments, and margins would result in high statistical variance. Consequently, we decided to model the effects of special exceptions using historical data based on hospitals' actual cost experiences. If we determined that a hospital may qualify for special exceptions, we modeled special exceptions payments from the project start date through the last available cost report (FY 2001). While we have not yet received all of the FY 2001 cost reports, we do have a sufficient number of FY 2001 cost reports to model a preliminary estimate of special exception payments for FY 2004. For purposes of modeling, we used the cost and payment data on the cost reports from HCRIS assuming that special exceptions would begin at the start of the qualifying project. In other words, when modeling costs and payment data, we ignored any regular exception payments that these hospitals may otherwise have received as if there had not been regular exception provision during the transition period. In projecting an eligible hospital's special exception payment, we applied the 70-percent minimum payment level, the cumulative comparison of current year capital PPS payments and costs, and the cumulative operating margin offset (excluding 75 percent of operating DSH payments).

Our modeling of special exception payments for FY 2004 produced the following results:

We note that hospitals must complete their projects by the end of PPS XVIII in order to be eligible for special exceptions payments. With complete submission of the PPS XVIII (FY 2001) cost reports, we estimate that about 30 hospitals may qualify for special exceptions payments. Thus, we project that special exception payments as a fraction of capital payments to all hospitals to be approximately 0.0005.

Because special exceptions are budget neutral, in the May 19, 2003 proposed rule, we proposed to offset the capital Federal rate by 0.05 percent for special exceptions payments for FY 2004. For this final rule, based on updated data, we are offsetting the capital Federal rate by 0.05 percent for special exceptions payments for FY 2004. Therefore, the exceptions adjustment factor is equal to 0.9995 (1^0.0005) to account for special exceptions payments in FY 2004.

In the August 1, 2002 IPPS final rule (67 FR 50131) for FY 2003, we estimated that total (special) exceptions payments would equal 0.30 percent of aggregate payments based on the capital Federal rate. Therefore, we applied an exceptions reduction factor of 0.9970 (1^0.0030) in determining the FY 2003 capital Federal rate. As we stated above, we estimate that exceptions payments in FY 2004 will equal 0.05 percent of aggregate payments based on the FY 2004 capital Federal rate. Therefore, we are applying an exceptions payment adjustment factor of 0.9995 (1^0.0005) to the capital Federal rate for FY 2004. The exceptions adjustment factor for FY 2004 is 0.25 percent higher than the factor for FY 2003 published in the August 1, 2002 IPPS final rule (67 FR 50131). This increase is primarily due to a refined analysis of more recent data.

The exceptions reduction factors are not built permanently into the capital rates; that is, the factors are not applied cumulatively in determining the capital Federal rate. Therefore, the net change in the exceptions adjustment factor used in determining the FY 2004 capital Federal rate is 0.9995/0.9970, or 1.0025.

5. Capital Standard Federal Rate for FY 2004

In the August 1, 2002 IPPS final rule (67 FR 50131) we established a capital Federal rate of $407.01 for FY 2003. As we noted above, as a result of the revisions to the GAF effective for discharges occurring on or after April 1, 2003 through September 30, 2003 published August 25, 2003 in the Federal Register (68 FR 22272), we have established a capital Federal rate of $406.93 for discharges occurring on or after April 1, 2003 through September 30, 2003. The capital rates effective for discharges occurring on or after April 1, 2003 through September 30, 2003, were used in determining the FY 2004 capital rates. In this final rule, we are establishing a capital Federal rate of $415.47 for FY 2004. The capital Federal rate for FY 2004 was calculated as follows:

• The FY 2004 update factor is 1.0070; that is, the update is 0.70 percent.
• The FY 2004 budget neutrality adjustment factor that is applied to the capital standard Federal payment rate for changes in the DRG relative weights and in the GAF is 1.0059.
• The FY 2004 outlier adjustment factor is 0.9521.
• The FY 2004 (special) exceptions payment adjustment factor is 0.9995.

Since the capital Federal rate has already been adjusted for differences in case-mix, wages, cost-of-living, indirect medical education costs, and payments to hospitals serving a disproportionate share of low-income patients, we are making no additional adjustments in the capital standard Federal rate for these factors, other than the budget neutrality factor for changes in the DRG relative weights and the GAF.

We are providing a chart that shows how each of the factors and adjustments for FY 2004 affected the computation of the FY 2004 capital Federal rate in comparison to the FY 2003 capital Federal rate. The FY 2004 update factor has the effect of increasing the capital Federal rate by 0.70 percent compared to the FY 2003 capital Federal rate, while the GAF/DRG budget neutrality factor has the effect of increasing the capital Federal rate by 0.59 percent. The FY 2004 outlier adjustment factor has the effect of increasing the capital Federal rate by 0.55 percent compared to the FY 2003 capital Federal rate. The FY 2004 exceptions payment adjustment factor has the effect of increasing the capital Federal rate by 0.25 percent compared to the exceptions payment adjustment factor for capital FY 2003. The combined effect of all the changes is to increase the capital Federal rate by 2.10 percent compared to the FY 2003 capital Federal rate.

We are also providing a chart that shows how the final FY 2004 capital Federal rate differs from the proposed FY 2004 capital Federal rate.

6. Special Capital Rate for Puerto Rico Hospitals

As explained at the beginning of section II.D. of this Addendum, hospitals in Puerto Rico are paid based on 50 percent of the Puerto Rico capital rate and 50 percent of the capital Federal rate. The Puerto Rico capital rate is derived from the costs of Puerto Rico hospitals only, while the capital Federal rate is derived from the costs of all acute care hospitals participating in the PPS (including Puerto Rico). To adjust hospitals' capital payments for geographic variations in capital costs, we apply a GAF to both portions of the blended capital rate. The GAF is calculated using the operating PPS wage index and varies, depending on the MSA or rural area in which the hospital is located. We use the Puerto Rico wage index to determine the GAF for the Puerto Rico part of the capital-blended rate and the national wage index to determine the GAF for the national part of the blended capital rate.

Because we implemented a separate GAF for Puerto Rico in FY 1998, we also apply separate budget neutrality adjustments for the national GAF and for the Puerto Rico GAF. However, we apply the same budget neutrality factor for DRG reclassifications and recalibration nationally and for Puerto Rico. As we stated in section III.A.4. of this Addendum, for Puerto Rico the GAF budget neutrality factor is 1.0002, while the DRG adjustment is 1.0008, for a combined cumulative adjustment of 0.9973.

In computing the payment for a particular Puerto Rico hospital, the Puerto Rico portion of the capital rate (50 percent) is multiplied by the Puerto Rico-specific GAF for the MSA in which the hospital is located, and the national portion of the capital rate (50 percent) is multiplied by the national GAF for the MSA in which the hospital is located (which is computed from national data for all hospitals in the United States and Puerto Rico). In FY 1998, we implemented a 17.78 percent reduction to the Puerto Rico capital rate as a result of Pub. L. 105-33. In FY 2003, a small part of that reduction was restored.

For FY 2003, before application of the GAF, the special capital rate for Puerto Rico hospitals was $198.29. With the changes we proposed to the factors used to determine the capital rate, the proposed FY 2004 special capital rate for Puerto Rico was $201.26. For this final rule, based on the final factors, the FY 2004 capital rate for Puerto Rico is $203.15.

B. Calculation of Inpatient Capital-Related Prospective Payments for FY 2004

With the end of the capital PPS transition period in FY 2001, all hospitals (except “new” hospitals under § 412.324(b) and under § 412.304(c)(2)) are paid based on 100 percent of the capital Federal rate in FY 2004. The applicable capital Federal rate was determined by making adjustments as follows:

• For outliers, by dividing the capital standard Federal rate by the outlier reduction factor for that fiscal year; and
• For the payment adjustments applicable to the hospital, by multiplying the hospital's GAF, disproportionate share adjustment factor, and IME adjustment factor, when appropriate.

For purposes of calculating payments for each discharge during FY 2004, the capital standard Federal rate is adjusted as follows: (Standard Federal Rate) × (DRG weight) × (GAF) × (Large Urban Add-on, if applicable) × (COLA adjustment for hospitals located in Alaska and Hawaii) × (1 + Disproportionate Share Adjustment Factor + IME Adjustment Factor, if applicable). The result is the adjusted capital Federal rate.

Hospitals also may receive outlier payments for those cases that qualify under the thresholds established for each fiscal year. Section 412.312(c) provides for a single set of thresholds to identify outlier cases for both inpatient operating and inpatient capital-related payments. The outlier thresholds for FY 2004 are in section II.A.4.c. of this Addendum. For FY 2004, a case qualifies as a cost outlier if the cost for the case plus the IME and DSH payments is greater than the prospective payment rate for the DRG plus $31,000.

An eligible hospital may also qualify for a special exceptions payment under § 412.348(g) for up through the 10th year beyond the end of the capital transition period if it meets: (1) a project need requirement described at § 412.348(g)(2), which in the case of certain urban hospitals includes an excess capacity test as described at § 412.348(g)(4); and (2) a project size requirement as described at § 412.348(g)(5). Eligible hospitals include sole community hospitals, urban hospitals with at least 100 beds that have a DSH patient percentage of at least 20.2 percent or qualify for DSH payments under § 412.106(c)(2), and hospitals that have a combined Medicare and Medicaid inpatient utilization of at least 70 percent. Under § 412.348(g)(8), the amount of a special exceptions payment is determined by comparing the cumulative payments made to the hospital under the capital PPS to the cumulative minimum payment level. This amount is offset by: (1) any amount by which a hospital's cumulative capital payments exceed its cumulative minimum payment levels applicable under the regular exceptions process for cost reporting periods beginning during which the hospital has been subject to the capital PPS; and (2) any amount by which a hospital's current year operating and capital payments (excluding 75 percent of operating DSH payments) exceed its operating and capital costs. Under § 412.348(g)(6), the minimum payment level is 70 percent for all eligible hospitals.

During the transition period, new hospitals (as defined under § 412.300) were exempt from the capital PPS for their first 2 years of operation and were paid 85 percent of their reasonable costs during that period. Effective with the third year of operation through the remainder of the transition period, under § 412.324(b) we paid the hospital under the appropriate transition methodology. If the hold-harmless methodology was applicable, the hold-harmless payment for assets in use during the base period would extend for 8 years, even if the hold-harmless payments extend beyond the normal transition period. As discussed in section VI.A. of the preamble of this final rule, under § 412.304(c)(2), for cost reporting periods beginning on or after October 1, 2002, we pay a new hospital 85 percent of their reasonable costs during the first 2 years of operation unless it elects to receive payment based on 100 percent of the capital Federal rate. Effective with the third year of operation, we pay the hospital based on 100 percent of the capital Federal rate (that is, the same methodology used to pay all other hospitals subject to the capital PPS).

C. Capital Input Price Index

1. Background

Like the operating input price index, the capital input price index (CIPI) is a fixed-weight price index that measures the price changes associated with capital costs during a given year. The CIPI differs from the operating input price index in one important aspect—the CIPI reflects the vintage nature of capital, which is the acquisition and use of capital over time. Capital expenses in any given year are determined by the stock of capital in that year (that is, capital that remains on hand from all current and prior capital acquisitions). An index measuring capital price changes needs to reflect this vintage nature of capital. Therefore, the CIPI was developed to capture the vintage nature of capital by using a weighted-average of past capital purchase prices up to and including the current year.

We periodically update the base year for the operating and capital input prices to reflect the changing composition of inputs for operating and capital expenses. The CIPI was last rebased to FY 1997 in the August 1, 2002 final rule (67 FR 50044).

2. Forecast of the CIPI for Federal Fiscal Year 2004

Based on historical data available through the second quarter of 2003, we forecast the CIPI to increase 0.7 percent in FY 2004. This reflects a projected 1.2 percent increase in vintage-weighted depreciation prices (building and fixed equipment, and movable equipment) and a 3.8 percent increase in other capital expense prices in FY 2004, partially offset by a 2.6 percent decline in vintage-weighted interest expenses in FY 2004. The weighted average of these three factors produces the 0.7 percent increase for the CIPI as a whole in FY 2004.

IV. Changes to Payment Rates for Excluded Hospitals and Hospital Units: Rate-of-Increase Percentages

As discussed in section VI. of the preamble of this final rule, in accordance with section 1886(b)(3)(H)(i) of the Act and effective for cost reporting periods beginning on or after October 1, 2002, payments to existing psychiatric hospitals and units, rehabilitation hospitals and units, and long-term care hospitals excluded from the IPPS are no longer subject to limits on a hospital-specific target amount (expressed in terms of the inpatient operating cost per discharge) that are set for each hospital, based on the hospital's own historical cost experience trended forward by the applicable rate-of-increase percentages (update factors).

Effective for cost reporting periods beginning on or after October 1, 2002, rehabilitation hospitals and units are no longer paid on a reasonable cost basis but are paid under the 100 percent of IRF PPS Federal rate. Effective for cost reporting periods beginning on or after October 1, 2002, LTCHs also are no longer paid on a reasonable cost basis but are paid under a LTCH DRG-based PPS. As part of the payment process for LTCHs, we established a 5-year transition period from reasonable cost-based reimbursement to a fully Federal PPS. However, a LTCH that is subject to the blend methodology may elect to be paid based on a 100 percent of the Federal prospective rate.

In accordance with existing § 413.40(c)(4)(ii) and (d)(1)(i) and (ii), where applicable, excluded psychiatric hospitals and units continue to be paid on a reasonable cost basis, and payments are based on their Medicare inpatient operating costs, not to exceed the ceiling (as defined in § 413.40(a)(3)). In addition, LTCHs that are paid under a blend methodology will have the TEFRA portion subject to the ceiling as well.

Section 1886(b)(7) of the Act had established a payment limitation for new hospitals and units excluded from the IPPS. While both rehabilitation hospitals and units and LTCHs are now paid under a PPS, psychiatric hospitals and units continue to be subject to the payment limitation. A discussion of how the payment limitation was calculated can be found in the August 29, 1997 final rule with comment period (62 FR 46019); the May 12, 1998 final rule (63 FR 26344); the July 31, 1998 final rule (63 FR 41000); and the July 30, 1999 final rule (64 FR 41529).

The amount of payment for a “new” psychiatric hospital or unit would be determined as follows:

• Under existing § 413.40(f)(2)(ii), for cost reporting periods beginning on or after October 1, 1997, the amount of payment for a new hospital or unit that was not paid as an excluded hospital or unit before October 1, 1997, is the lower of: (1) the hospital's net inpatient operating costs per case; or (2) 110 percent of the national median of the target amounts for the same class of excluded hospitals and units, adjusted for differences in wage levels and updated to the first cost reporting period in which the hospital receives payment. The second cost reporting period is subject to the same target amount applied to the first cost reporting period.
• In the case of a hospital that received payments under § 413.40(f)(2)(ii) as a newly created hospital or unit, to determine the hospital's or unit's target amount for the hospital's or unit's third 12-month cost reporting period, the payment amount determined under § 413.40(f)(2)(ii)(A) for the preceding cost reporting period is updated to the third cost reporting period.

The amounts included in the following table reflect the updated 110 percent of the national median target amounts of new excluded psychiatric hospitals and units for cost reporting periods beginning during FY 2004. These figures are updated with the most recent data available to reflect the market basket increase percentage of 3.4 percent. This percentage change in the market basket reflects the average change in the price of goods and services purchased by hospitals to furnish inpatient hospital services (as projected by CMS' Office of the Actuary based on its historical experience with the IPPS). For a new provider, the labor-related share of the target amount is multiplied by the appropriate geographic area wage index, without regard to IPPS reclassifications, and added to the nonlabor-related share in order to determine the per case limit on payment under the statutory payment methodology for new providers.

Effective for cost reporting periods beginning on or after October 1, 2002, this payment limitation is no longer applicable to new LTCHs since they will be paid 100 percent of the Federal rate. A new LTCH is a provider of inpatient hospital services that meets the qualifying criteria for LTCHs specified under § 412.23(e)(1) and (e)(2) and whose first cost reporting period as a LTCH begins on or after October 1, 2002 (§ 412.23(e)(4)). Under the LTCH PPS, new LTCHs are paid based on 100 percent of the fully Federal prospective rate (they may not participate in the 5-year transition from cost-based reimbursement to prospective payment). In contrast, those “new” LTCHs that meet the definition of “new” under § 413.40(f)(2)(ii) and that have their first cost reporting periods beginning on or after October 1, 1997, and before October 1, 2002, may be paid under the LTCH PPS transition methodology. Since those hospitals by definition would have been considered new before October 1, 2002, they would have been subject to the updated payment limitation on new hospitals that was published in the FY 2003 IPPS final rule (67 FR 50103). Under existing regulations at § 413.40(f)(2)(ii), the “new” hospital would be subject to the same cap in its second cost reporting period; this cap would not be updated for the new hospital's second cost reporting year. Thus, since the same cap is to be used for the “new” LTCH's first two cost reporting periods, it is no longer necessary to publish an updated cap.

We are in the process of developing a proposed rule that would establish a per diem PPS for inpatient psychiatric facilities (IPFs) (previously referred to as psychiatric hospitals and units) that is required under the provisions of section 124 of Pub. L. 106-113.

V. Payment for Blood Clotting Factor Administered to Hemophilia Inpatients

In December 2002, the Department implemented a policy that established the Single Drug Pricer (SDP) to correct identified discrepancies, further the legislative goal of establishing a uniform payment allowance as a reflection of the average wholesale price (AWP), and otherwise apply the existing stature and regulation more accurately and efficiently (CMS Program Memorandum AB-02-174, December 3, 2002, which can be accessed at: http://www.cms.hhs.gov/manuals ). Under the SDP, CMS will establish prices centrally, thereby resulting in greater consistency in drug pricing nationally. The SDP instruction applies to blood clotting factors furnished to hospital inpatients. The payment allowance for the single national drug price for each Medicare covered drug is based on 95 percent of the AWP, except for drugs billed to durable medical equipment regional carriers (DMERCs) and hospital outpatient drugs billed to fiscal intermediaries. We are publishing this notice here because we previously have addressed the add-on payment for the costs of administering blood clotting factor in the IPPS annual rule (see the August 1, 2000 IPPS final rule (65 FR 47116).

On a quarterly basis, CMS will furnish three SDP files to all fiscal intermediaries. Each fiscal intermediary must accept the SDP files and process claims for any drug identified on the files on the basis of the price shown on the applicable file. Previously, the fiscal intermediary performed annual update calculations based on the most recent AWP data available to the carrier. The fiscal intermediary should use the SDP to price the blood clotting factors.

VI. Tables

This section contains the tables referred to throughout the preamble to this final rule and in this Addendum. For purposes of this final rule, and to avoid confusion, we have retained the designations of Tables 1 through 5 that were first used in the September 1, 1983 initial prospective payment final rule (48 FR 39844). Tables 1A, 1C, 1D, 2, 3A, 3B, 4A, 4B, 4C, 4F, 4G, 4H, 5, 6A, 6B, 6C, 6D, 6E, 6F, 6G, 6H, 7A, 7B, 8A, 8B, 9, 10, and 11 are presented below. The tables presented below are as follows:

Table 1A—National Adjusted Operating Standardized Amounts, Labor/Nonlabor

Table 1C.—Adjusted Operating Standardized Amounts for Puerto Rico, Labor/Nonlabor

Table 1D.—Capital Standard Federal Payment Rate

Table 2.—Hospital Average Hourly Wage for Federal Fiscal Years 2002 (1998 Wage Data), 2003 (1999 Wage Data), and 2004 (2000 Wage Data) Wage -Indexes and 3-Year Average of Hospital Average Hourly Wages

Table 3A.—3-Year Average Hourly Wage for Urban Areas

Table 3B.—3-Year Average Hourly Wage for Rural Areas

Table 4A.—Wage Index and Capital Geographic Adjustment Factor (GAF) for Urban Areas

Table 4B.—Wage Index and Capital Geographic Adjustment Factor (GAF) for Rural Areas

Table 4C.—Wage Index and Capital Geographic Adjustment Factor (GAF) for Hospitals That Are Reclassified

Table 4F.—Puerto Rico Wage Index and Capital Geographic Adjustment Factor (GAF)

Table 4G.—Pre-Reclassified Wage Index for Urban Areas

Table 4H.—Pre-Reclassified Wage Index for Rural Areas

Table 5.—List of Diagnosis Related Groups (DRGs), Relative Weighting Factors, Geometric and Arithmetic Mean Length of Stay

Table 6A.—New Diagnosis Codes

Table 6B.—New Procedure Codes

Table 6C.—Invalid Diagnosis Codes

Table 6D.—Invalid Procedure Codes

Table 6E.—vised Diagnosis Code Titles

Table 6F.—Revised Procedure Code Titles

Table 6G.—Additions to the CC Exclusions List

Table 6H.—Deletions from the CC Exclusions List

Table 7A.—Medicare Prospective Payment System Selected Percentile Lengths of Stay

FY 2002 MedPAR Update March 2003 GROUPER V20.0

Table 7B.—Medicare Prospective Payment System Selected Percentile Lengths of Stay

FY 2002 MedPAR Update March 2003 GROUPER V21.0

Table 8A.—Statewide Average Operating Cost-to-Charge Ratios—July 2003

Table 8B.—Statewide Average Capital Cost-to-Charge Ratios—July 2003

Table 9.—Hospital Reclassifications and Redesignations—FY 2004

Table 10.—Mean and .75 Standard Deviation by Diagnosis-Related Groups (DRGs)-July 2003

Table 11.—LTC-DRGs Relative Weights and Geometric and Five-Sixth of the Average Length of Stay-FY 2004