Opinion
December 21, 1984
Frederick A.O. Schwarz, Jr., Corporation Counsel ( Kenneth Burford of counsel), for petitioner.
John F. Gangemi for respondent.
On August 9, 1983, petitioner, Commissioner of Social Services, as assignee of Juliana C., commenced this proceeding under Family Court Act article 5 for an order of filiation and support naming respondent as the father of Nicholas James C., born to Juliana C. on August 9, 1981. Respondent Louis T. denied paternity. Hearings were held on June 26, 1984, September 11, 1984, September 13, 1984 and September 17, 1984. The court heard testimony of Juliana C., the mother of the child, Grace S., a cousin of the mother, Dr. S. Schechter, the mother's gynecologist and obstetrician, respondent Louis T. and Prof. Richard L. Borowsky, Associate Professor of Genetics at New York University. In addition, the results of the human leucocyte antigen (HLA) test performed on the parties and the child by Dr. Leon N. Sussman of the Lindsley F. Kimball Research Institute of the New York Blood Center (NYBC) were introduced into evidence in accordance with Family Court Act § 532.
At the close of the hearing, respondent moved to dismiss the proceeding because of petitioner's failure to establish paternity by clear, convincing and entirely satisfactory evidence, based on the testimony adduced at the hearing as well as his challenge of the accuracy of the mathematical computation of the results of the HLA test.
Based on the testimony presented during the hearing with regard to the relationship between the parties as well as the expert's unrebutted findings of miscalculations of the probability of paternity of the HLA results obtained by the NYBC, the court agrees with respondent's contentions and dismisses the petition for the reasons stated hereafter.
THE HLA TEST RESULTS
Since March 2, 1981, Family Court Act § 532 authorizes the use of the results of the HLA test in paternity proceedings as evidence of the likelihood of paternity of an alleged father. In addition to a statement that a respondent is "not excluded as a father," the results of the HLA test are introduced together with a statement of "probability of paternity," synonymous with "plausibility of paternity" and of "Hummel's Predicate," which purports to quote the odds in favor of an individual's paternity. Most courts consider these results admissible and use them as positive proof of paternity giving them such weight as the court sees fit to aid the determination of paternity.
Matter of Smith v Jones ( 120 Misc.2d 834 [Fam Ct, NY County 1983]) excluded Hummel's Predicate as not authorized by Family Court Act § 532.
In the case at bar, the results of the HLA test submitted by NYBC showed:
Alleged father, A1, A3, B7, B18 Mother, A2, B18, B49 Child, A2, A3, B7, B18 HLA: Paternity Index = 11.5 (odds) Plausibility of Paternity = 93% = "likely". In order to explain the extraordinary challenge to the accuracy of the HLA test results in the present case, it is necessary to indicate the difference between this case's specific challenge and prior challenges of the HLA evidence.The Family Court in Matter of Alicia C. v Evaristo G. ( 115 Misc.2d 564 [Fam Ct, Queens County 1983]) declined to admit the results of the HLA test on account of an arbitrary, subjective determination of race by the examiner, and dismissed the petition. The Appellate Division, Second Department, reversed the case and granted the petition holding that the issue of admissibility of the test results has been settled by the Legislature and that any criticism of the test results can only go to the weight but not the admissibility of the evidence. ( Matter of Alicia C. v Evaristo G., 93 A.D.2d 820.) Since Matter of Alicia C. v Evaristo G. ( supra), various other challenges have been presented to the courts and the courts in those cases have essentially found that the results of the test as submitted by NYBC indicating a high percentage of probability of paternity together with the testimony obtained during the trial established respondent's paternity. Of course, in those cases the degree of reliance by the lower courts on the HLA results can never be determined.
Matter of Smith v Jones, supra; Matter of Alicia C. v Evaristo G., supra; Matter of Commissioner of Social Servs v Bart D., supra; Matter of Lorraine M. v Linwood M.S., supra.
As far as this court knows, the first decision which refused to give any weight to the HLA results and challenged basic concepts relating to their reliability is the decision by Honorable Stanley Gartenstein in Matter of Angela B. v Glenn D. ( 126 Misc.2d 646 [Fam Ct, N Y County 1984]). In that case, Judge Gartenstein gave no weight at all to the HLA results which indicated a 99% probability of paternity. He found the testimony of the petitioner completely unreliable, contradictory and fabricated for the purposes of the litigation. Despite the HLA test's high probability of paternity and the almost overwhelming acceptance of HLA test results as strong evidence of paternity, he found that a judge need not accept as ultimate proof of paternity the laboratory results where they contradict completely the trial testimony. In this case the probability computation itself was accepted and not challenged by the respondent. However, the court, paraphrasing testimony by Dr. Sussman, the director of the NYBC, which performs and evaluates the tests, convinced Judge Gartenstein that the results of the HLA tests "did not, could not, positively state that respondent or anyone else was the father of petitioner's child" (p 652). In order to illustrate further his conclusion that the HLA test "in contradistinction to a procedure yielding definitive results, is rather an impersonal analysis which, by its very nature, cannot mesh with the factual underpinnings of each individual case" (p 653), Judge Gartenstein summarized testimony by Dr. Sussman which acknowledged "that Dr. Alexander Weiner, his mentor, had conducted blood tests in contested litigation where the mother of the infant in issue had been having sexual intercourse with two men. HLA testing of the first man conjectured a 99.7% plausibility of paternity, results which were reported to the court, and in turn presumably formed the basis of an adjudication of paternity against the accused putative father. Because of procedural facts peculiar to that case, the second male was tested 2 1/2 years later. His tests results yielded an even higher probability of paternity, viz., 99.8%. In litigation where so much is at stake, the possibility of a miscarriage of justice such as that just illustrated is simply intolerable." ( Supra, at p 653.)
"Dr. Sussman was emphatic in stating that HLA test results did not, could not, positively state that respondent or anyone else was the father of petitioner's child" ( supra, p 652). And quoting Dr. Sussman from an article in the New York State Journal of Medicine, March 1981: "`The temptation to use a mathematical formula for calculating the probability of paternity in cases where an exclusion is not obtained, is very great. There is, however, the risk of undue influence on the court when a report of "95 percent or higher probability paternity" is presented.'" (Supra, p 652.)
Judge Gartenstein was even more puzzled by "Dr. Sussman's admission on the witness stand that despite his enthusiastic advocacy on behalf of the HLA test whose results are always issued in conjunction with `Hummel's predicate', he had no idea of Hummel's methodology other than his belief that Hummel was `universally accepted.'" ( Supra, at p 656.)
In the case at bar, the testimony offered by the mother is not as unbelievable as in Judge Gartenstein's case, but it does not by itself reach the level of clear and convincing evidence which is required to prove paternity. Therefore, the results of the HLA test, if added as strong evidence of paternity, might turn the scale in favor of petitioner. However, the respondent here has produced unrebutted testimony by an expert geneticist challenging for the first time in a paternity case the mathematical correctness of the computation of the probability of paternity which was calculated by Dr. Sussman as 93% or "likely" by Prof. Hummel's interpretation. This challenge seems especially important in view of the fact that in many cases men across this State admit paternity without requesting a trial, relying on the results of the HLA tests and courts are often inclined to accept the computation of probability of paternity and Hummel's Predicates without effective challenge by the respondents who in many cases are without means to afford genetics experts familiar with the intricacies and possible flaws of tests which the Legislature has authorized to accept in evidence and who are represented by attorneys unfamiliar with scientific facts. Thus, as a practical matter, the results of the HLA test as submitted by NYBC together with the "probability (or plausibility) of paternity" and of "Hummel's Predicate" which purports to give the English language equivalent of the tested individual's likelihood of paternity, frequently determine the outcome of paternity proceedings. In addition, it is important to note that there is no mention of "probability of paternity" or "Hummel's Predicates" in the statute or its history as opposed to the HLA test itself.
(10 Fam LQ 247, 262.) During the trial information was given to the court that Prof. Hummel is not a geneticist but a statistician.
Compare Matter of La Croix v Deyo ( 113 Misc.2d 89, 91), in which the trial court placed such heavy reliance upon the results of the HLA test, that it mooted any possible objection to the petitioning putative father's testimony of relations with the deceased mother under CPLR 4519, based solely upon what it presumed to be "incontrovertible scientific evidence".
See, Matter of Angela B. v Glenn D., supra, at p 656; Matter of Smith v Jones, supra, at p 844.
In the case at bar, Prof. Borowsky, respondent's expert witness, testified that he had discussed the results and methods used to calculate the paternity indices with Dr. Sussman of NYBC and that — based on this discussion and his own study of the HLA results — the 93% "plausibility of paternity" calculated by the NYBC was reached by an erroneous interpretation of the HLA results.
The methods used are accepted and described in 10 Family Law Quarterly 247, 260-61.
Prof. Borowsky emphasized that the HLA system is more complicated genetically than any of the red blood cells systems and that these added complications increase the difficulty of interpretation. He explained that the difficulty arises primarily because the HLA system is comprised of more than one gene locus and since the two most important gene loci are closely linked, the net effect of the close linkage is that a paternity analysis using HLA data may require a correction for uncertainty of "haplotypes" in the putative father. In the expert's opinion, such a correction was required in this case. The witness, furthermore, explained that the NYBC does apply a correction for this uncertainty but has applied it improperly and that the effect of this error resulted in its seriously overestimating the plausibility of paternity.
Prof. Borowsky's report to respondent's attorney was made available to the court and petitioner's attorney in order to simplify and assist in Prof. Borowsky's cross-examination. This decision uses part of this report which summarizes the witness' testimony.
10 Fam LQ 247, 276-77.
In order to explain this error to the court, the witness reviewed some basic principles of genetics and paternity testing through genotype. He stated that for every gene site or "locus" considered in the test, a person has two copies, "alleles," because the gene loci are on the chromosomes, which are paired structures. The alleles can either be identical, encoding the same genetic information or different, encoding variations of this genetic information. The loci used for paternity testing (e.g., ABO or HLA-A) were chosen because they have a relatively large number of alleles. The greater this diversity, the easier it is to distinguish among and identify individuals by their genotypes. The genes themselves are not examined. Rather, the presence of a particular allele is inferred because it directs the production of a specific substance or antigen which can be detected in the tested sample. Knowing the antigen characteristics or "phenotype" of a person gives information about his genetic makeup, but it does not always tell everything of importance which is the basis for the uncertainty in respondent's case.
In general, paternity testing is based on the comparison of alleles in the mother, the child and the alleged father. Any alleles not found in the mother but present in the child must be derived from the father. Such alleles are called "obligatory alleles." If the alleged father does not have the obligatory alleles he is excluded; if he has them, he is not excluded.
The NYBC uses the paternity index or "odds" method to calculate the likelihood of paternity. The paternity index (w) is the ratio of the probability of obtaining the obligatory allele from the alleged father (x), divided by the probability of obtaining it from the set of all other available males (y). The plausibility of paternity is X/Y = W This formula
In the Matter of Angela B. v Glenn D. ( supra), the manner in which the frequency of the obligatory allele in the general population is computed is challenged. The challenge in the case at bar concerns a specific miscalculation of the plausibility of paternity.
1 + X/Y.
represents the comparison of the alleged father's odds and one other nonexcluded man in the general population. The first probability is obtained from knowledge of the alleged father's genotype while the second probability is estimated from tables of allelic frequencies in the population at large, compiled by NYBC. The odds ratio can be converted to "plausibility of paternity" or "probability of paternity." In the case at bar NYBC has calculated that the probability of obtaining the obligatory HLA alleles from the respondent is .3503, while the frequency of those alleles (actually a set of alleles, called "haplotype") is .0302 among ethnic whites in the New York City area. NYBC's ratio is .3503/.0302 = 11.6 which means that the "odds" are 11.6 to 1 that respondent is the father. This can also be expressed by stating that the plausibility of paternity is 92%, because 11.6/12.6 = 92%.
This method of calculation was not challenged by the expert witness. However, according to Prof. Borowsky, an important correction for a complication was improperly applied and led to an erroneous result. Recalculating the odds figure, Prof. Borowsky obtained a plausibility of paternity of 66%, "not useful" according to Hummel's Predicates. The difference between the NYBC's and the witness' computation is based on the difference in the way an uncertainty in the determination of respondent's genetic makeup was corrected. As emphasized by the witness, the HLA system differs from all of the others because it consists of two loci (A and B) that sit next to one another on the chromosome, rather than a single locus, such as the one for ABO red blood cell types. Both HLA loci have a number of alleles but since the loci are closely linked, allelic combinations tend to be inherited as units or "haplotypes", rather than independently.
This decision does not enter into the issue whether the statute authorizes the use of Hummel's Predicates at all.
In the case at bar, the obligatory haplotype is A3 B7. Respondent was phenotyped as A1, A3, B7, B18, the mother as A2, B18, B49 and the child as A2, A3, B7, B18. Respondent therefore has the obligatory alleles of A3 B7. The uncertainty arises, however, because the specific association of the alleles on the chromosomes is unknown. Respondent could have one chromosome with A3, B7 and the other with A1, B18 or the arrangements could be A1, B7, and A3, B18. The distinction between the two alternatives is crucial. If respondent is A3, B7 and A1, B18, he could have provided the obligatory haplotype (A3, B7), and this would be strong evidence of his being the father. On the other hand, if he is A1, B7 and A3, B18 he could not be the father and is virtually excluded as the father. There is no evidence to decide between the two alternatives and therefore the calculation of the "odds" has to account for both possibilities. Dr. Sussman's table of haplotype frequencies provides the basis for calculating the probabilities of the two possibilities. The witness testified that he is in agreement with Dr. Sussman that there is a .7006 chance (roughly 70%) that respondent is A3, B7 and A1, B18 and a .2994 (roughly 30%) chance that he is A3, B18 and A1, B7. However, NYBC and the witness dispose of this uncertainty in different ways which — in the witness' opinion — erroneously creates the high probability of paternity obtained by Dr. Sussman. NYBC reasons as follows: The chance that respondent would pass the A3 B7 haplotype is .5 (usually .5 from each parent). Since the probability is only .7006 that he is of this haplotype, they reason that the a priori probability that he, as an individual, passes on that haplotype is .7006 X .5 = .3503. Since the general frequency of that haplotype among ethnic whites is .0302, the odds ratio is computed by the NYBC as .3503 .0302 = 11.6 and 11.6/12.6 = 92%. In the witness' opinion, a flaw must exist in this approach because there is a .2994 chance that respondent is excluded as the father. Thus, whatever probability of paternity is calculated for the HLA results, it cannot exceed .7006 (roughly 70%). The expert then explained that it is incorrect to say that respondent has a .3503 probability of passing on the A3, B7 haplotype. He may have a .5 probability of passing it on if he is A3, B7 and A1, B18 or he may have a zero probability of passing it on if he is A1, B7 and A3, B18 but it is sure that he, as an individual, does not have a .3503 chance of passing on any haplotype to a child. Therefore, the proper way to treat the uncertainty is to calculate the two alternatives separately and then take an average, weighted by the factors, .7006 and .2994. Based on this method of computing the probability of paternity, the expert witness obtained a 66% probability of paternity. He reasoned as follows: If the respondent has the obligatory A3, B7 and A1, B 18, the the paternity index is .5/.0302 = 16.56 = .9430 (94%) probability of paternity. If he has A3, B18 and A1, B7 then the passing on probability of A3, B7 is zero. Therefore, the over-all plausibility of paternity for the HLA data would be: (.9430 X .7006) + (0.0 X .2994) = .66 (66%). This figure is in accordance with the fact that the "plausibility of paternity" here cannot be more than .7006 (70%).
The witness illustrated his presentation with drawings to facilitate the understanding of the rather complicated nature of his scientific testimony. On cross-examination, he also testified that no further blood tests would show whether or not respondent has the obligatory haplotypes. Only testing of respondent's ancestors might lead to such results since the specific haplotypes are usually inherited.
At the close of respondent's case on September 13, 1984, the court adjourned the matter in order to give petitioner an opportunity to rebut respondent's challenge of Dr. Sussman's results concerning his computation of the HLA plausibility of paternity. On September 17, 1984, petitioner stated that no rebuttal would be offered and after final argument the court reserved decision.
Having considered the unrebutted testimony of respondent's expert witness, the court disregards as incorrect the HLA results as submitted by the NYBC.
It appears to this court that this unrebutted testimony by respondent's expert witness has thrown completely new light on the computation of "plausibility of paternity" by the NYBC, and significantly discredited the laboratory's calculations. While not contesting the HLA blood testing itself, the mathematical interpretation of computation of the odds of paternity as calculated by NYBC on which Hummel's Predicates are based are in error in this case and will have to be reexamined in future cases taking into consideration the very interesting, and to the court, convincing recalculation of the "plausibility of paternity" in the instant case. Moreover, the challenge to the reliability of the NYBC results in this case which is not based on general statistical arguments of interpretation of the HLA results but addressed directly to the correctness of the calculations themselves as used by NYBC will hopefully contribute to the continuing dialogue with the Legislature in this important area of mutual concern to the scientific and judicial community.
The "proper judicial role is to accept the plain letter of the statute and to engage in a continuing dialogue with the Legislature, particularly during this experimental period." ( Matter of Smith v Jones, 120 Misc.2d 834, 841, supra.)
FINDINGS OF FACT AND CONCLUSIONS OF LAW
The relevant testimony of all witnesses established that at no time did there exist an intimate relationship between the child's mother and respondent. The court heard about occasional meetings between the parties and their friends, about "hanging out" together at a discotheque and an after-hours club on various Saturday evenings between May 1980 and February 1981. However, the assignor admitted that the parties had never dated or gone together to the movies, a restaurant or a park. She also admitted that she never told respondent, at any time, that he was the child's father. After about her seventh month of pregnancy, in a telephone conversation with the respondent, she did tell him that she was pregnant, but she did not indicate to him that he was the responsible party. After the baby's birth, she told a cousin that the respondent was the child's father. The cousin then approached the respondent and told him of Ms. C.'s allegation of his paternity.
Ms. C.'s last menstrual period was at the end of October 1980. Her testimony, that she did not know that she was pregnant in December 1980 after two visits to her obstetrician, was credible and supported by the credible testimony of her gynecologist after negative results of two urine pregnancy tests taken during December. However, her testimony that she saw her physician again only on May 20, 1981, when she was about seven months pregnant, in order to be fitted for a diaphragm and at that time did not know that she was pregnant, appears to this court to be incredible in view of the fact that she was over 25 years old at that time and not a young, inexperienced girl. The court also has doubts about the truthfulness of her statement that she had had intercourse for the first time in her life with respondent on November 8, 1980, at the Golden Gate Motor Inn, where she claimed to have stayed with respondent for about one hour at about 5:00 A.M., and again on February 9, 1981, after a chance meeting with respondent and some friends at a discotheque. In addition, petitioner's failure to introduce any registration records of the Golden Gate Motor Inn to substantiate Ms. C.'s claim that she saw respondent register at the hotel also undermines the reliability of petitioner's evidence.
Respondent, who testified on his own behalf, confirmed the occasional meetings at the discotheque and after-hours club with Ms. C. in the presence of friends of both parties. However, he strongly denied ever having had intercourse with Juliana C. His demeanor was forthright and candid and his testimony in all major respects was consistent.
Without the aid of the HLA results to determine paternity, the court must rely solely on the testimony and demeanor of the witnesses concerning the relationship of petitioner's assignor and the respondent. Petitioner has the burden "to establish paternity by `clear and convincing' evidence, evidence which is `entirely satisfactory' and creates a genuine belief that respondent is the father of the child" ( Matter of Commissioner of Social Servs. [ Patricia A.] v Philip De G., 59 N.Y.2d 137, 141-142). The court finds from all the testimony and evidence before it in this case and without strong evidence of HLA results in favor of respondent's paternity, that this burden has not been met. Accordingly, the petition is dismissed.