IN RE: M., Petitioner, v. MARVIN S., Respondent.
This court holds that, pursuant to the 1994 amendment to FCA § 532 that creates a rebuttable presumption of paternity if blood test results indicate at least a ninety-five percent “Probability of Paternity”, it is proper for a certified laboratory that performs blood genetic marker tests to utilize an assigned “Prior Probability of Paternity” value of 0.5 in calculating a “Probability of Paternity” number, notwithstanding evidence that respondent has a fertility problem.
In this paternity proceeding, petitioner, M., seeks to establish that respondent, Marvin S., is the father of her child, S., born on April 28, 1987. Petitioner also seeks to compel respondent to furnish support for the child. Respondent denied paternity. Pursuant to FCA § 532, court-ordered blood genetic marker tests, including the human leucocyte antigen (HLA) test, were done, as well as DNA tests, on the mother, child and respondent.
The matter proceeded to a lengthy trial at which both petitioner and respondent testified. In addition, the parties each provided expert testimony and documentary evidence concerning the blood test statistical results and the significance of the respondent's alleged fertility problems.
For the reasons summarized below, the court finds by evidence that is clear, convincing, and entirely satisfactory that petitioner has met her burden of proving that respondent is the father of the subject child. See Jane PP v. Paul QQ, 65 N.Y.2d 994, 494 N.Y.S.2d 93, 484 N.E.2d 122 (1985); Social Servs. v. Philip De G., 59 N.Y.2d 137, 463 N.Y.S.2d 761, 450 N.E.2d 681 (1983).
The following is a summary of the essential trial testimony and evidence:
Petitioner testified that she met the respondent in late 1982. She said she had sexual relations with respondent on five occasions between 1983 and 1986, the latter on July 23, 1986. Petitioner detailed the July 23 encounter: After having dinner, they went to her Manhattan apartment where respondent spent the night. They had sexual intercourse involving penetration and ejaculation; no birth control was used. She testified she was otherwise abstinent for many months before and after July 23. After missing her regular August cycle, she learned she was pregnant. She contacted respondent who denied he was responsible and also told her he had a low sperm count.
The results of court ordered blood tests, introduced in evidence, conducted in September, 1994, including HLA testing, showed a Combined Paternity Index of 545 to one and a Probability of Paternity of 99.82%, using a Prior Probability of 0.5; an additional DNA analysis indicated a Combined Paternity Index of 49,111 to one and a Probability of Paternity of 99.99%, also using a Prior Probability of 0.5.
Respondent testified, consistent with petitioner, that he met her in late 1982, and had sex with her several times thereafter, the last time on July 23, 1986. Although he acknowledged penetrating petitioner on July 23, and ejaculating, he denied ejaculating inside of petitioner claiming he had no “personal knowledge” of sperm being released inside her. Respondent agreed he did not use birth control with petitioner. Respondent also acknowledged fathering a child, who was born in 1968, with his first wife. He claimed he tried to conceive a child with his second wife, during the 1970s, but was unsuccessful. He consulted various physicians, and on the basis of fertility testing in 1979 and again in 1993 believed it was improbable that he could father a child.
Each party provided expert witness testimony by a Board Certified urologist who was qualified as an expert in urology and male infertility. Although neither expert examined the respondent at any time close to when the subject child was actually conceived, and although the experts differed in their estimate of the likelihood that the respondent could conceive a child on a given occasion, they both agreed on many important points: Both agreed the respondent had poor quality sperm, i.e., low sperm count, poor motility, and poor morphology; 1 both agreed respondent had a medical condition (a bilateral varicocele, which is an abnormal swelling of veins that drain the testicles) that is a common cause of male fertility problems and the most likely cause of his poor quality sperm; both agreed respondent's potential for fertility was impaired but that he was not totally infertile.
Dr. David Bing testified as an expert witness for the petitioner. He has a doctorate in microbiology and is employed at a laboratory in Massachusetts that does paternity testing. He was qualified as an expert in microbiology and in the analysis and interpretation of paternity testing. Dr. Bing reviewed the paternity blood tests results in evidence and also familiarized himself with some aspects of the trial testimony concerning respondent's fertility issues. He agreed that the results of the paternity testing as reported by the laboratory were correct but questioned the interpretation of those results.
Dr. Bing began his direct testimony by briefly describing the derivation of the standard terms utilized in reporting paternity results: The Combined Paternity Index; the Prior Probability of Paternity; and the Probability of Paternity. In essence, the Paternity Index is a ratio comparing the frequency of the alleged father producing a set of obligatory genes (the genetic material the subject child must have inherited from its actual biological father) to the frequency of a random man of the respondent's same ethnic background producing the same set of obligatory genes. A Paternity Index is computed for each of the genetic systems used in the blood testing. The resulting indexes are then multiplied to obtain the Combined Paternity Index, which indicates how much more likely the respondent is to produce a sperm containing all of the obligatory genes compared to a randomly selected ethnically similar male. Then, a value (called the Prior Probability of Paternity) is assigned by the laboratory (which has no knowledge of the non-genetic evidence in the case) to represent the probability of paternity based on the non-genetic evidence in the case. In this case, the laboratory assumed and assigned a Prior Probability value of 0.5 which essentially assumes that the respondent alleged father is equally likely to be the biological father as he is not to be the biological father, based on the non-genetic evidence. The Combined Paternity Index and the Prior Probability of Paternity are then combined by a mathematic formula (Bayes' Theorem) to obtain a Probability of Paternity that is expressed as a percentage.
Following this explanation, Dr. Bing noted that, mathematically, as the number assigned to the Prior Probability of Paternity changes, so does the resulting Probability of Paternity. Dr. Bing then opined that in his view a Prior Probability of 0.5 is not applicable to this case. The doctor's stated reason related to the testimony and concerns raised about the respondent's fertility difficulties. According to Dr. Bing, since the respondent in fact does not have an equal chance of fathering a child on a given occasion as does a random man, because of his fertility problem, an assigned Prior Probability value of 0.5 is not a correct number. Dr. Bing testified that, in his opinion, the Prior Probability of Paternity value should be less than 1,000 or even less than 10,000. Dr. Bing illustrated how a lower Prior Probability of Paternity value would affect the final Probability of Paternity calculation: If the Prior Probability were 1/1,000 the Probability of Paternity would be 98%; if the Prior Probability were 1/10,000 the Probability of Paternity would drop to 83%; and at 1/100,000 it would reach 32%. After some reflection, Dr. Bing finally testified that in his opinion 99.9% (the Probability of Paternity reported by the laboratory and in evidence) is not correct and that, based on his knowledge of the fertility evidence, the Probability of Paternity should definitely be below 90%, with 1/10,000 his best estimate of a Prior Probability value, and therefore a Probability of Paternity figure of 83% would be the more accurate result.
On cross examination, Dr. Bing acknowledged that his own laboratory, which does thousands of paternity tests each year, uses an assigned Prior Probability figure of 0.5, as did the laboratory in this case. He further agreed that it is not the business of a paternity testing laboratory to consider fertility issues, although he insisted that it was appropriate for him, in his direct testimony, to interpret the meaning of the 0.5 assigned value. Dr. Bing further conceded that he personally knew little about fertility rates in couples with fertility difficulties. He further conceded that the Prior Probability value is affected by all the non-genetic evidence in the case, such as the facts and circumstances concerning conception, about which he knew nothing. Dr. Bing agreed that the only non-genetic fact he considered was the fertility issue and that, for example, if the respondent had, as a hypothetical fact, exclusive sex with the petitioner during the time of conception, the 0.5 Prior Probability value would surely go up, in fact, considerably up.
Dr. Clifton Ray Harris was called as an expert rebuttal witness in response to Dr. Bing. He has a doctorate in biology and genetics, and is currently an Associate Director in the Department of Paternity Evaluation at the Laboratory Corporation of America (formerly called Roche Biomedical Laboratories), the laboratory that performed the blood tests, which results are in evidence in this case. Dr. Harris was qualified as an expert in genetic marker testing and its application to parentage evaluation.
After a brief explanation, similar to Dr. Bing's, of genetic marker testing and the derivation of the statistical terminology, Dr. Harris noted (also consistent with Dr. Bing) that the Probability of Paternity term has two variables: (1) the Paternity Index which is the product of the genetic testing information; and (2) the Prior Probability, which is the sum of the non-genetic information in the case. As did Dr. Bing, Dr. Harris testified that a 0.5 Prior Probability figure is not a factually real number, but a statistical convention to reflect a neutral weighing of the non-genetic information by the laboratory. But Dr. Harris emphasized that the utilization by a laboratory of the 0.5 figure is a nationally accepted convention and that all the major laboratories use this figure for paternity test reporting purposes.
Dr. Harris interpreted the reported Combined Paternity Index in this case of 49,111/1 to mean that this respondent is 49,111 times more likely than a man chosen at random from respondent's same ethnic group to have produced a sperm cell containing the genetic markers the subject child must have received from her biological father.
With respect to interpreting the reported 99.99% Probability of Paternity, this result used the standard Prior Probability value of 0.5. In addressing Dr. Bing's argument, Dr. Harris stated that one could hypothetically vary the Prior Probability value to see its effect on the Probability of Paternity. To do this correctly, the Prior Probability should be varied using the 49,111/1 Combined Paternity Index-the reported genetic testing result-as a constant, i.e., the hypothetical effect of variations in assigned Prior Probability values should be varied in relation to the actual reported genetic test result in this specific case. When this exercise is done, it is clear that the ultimate Probability of Paternity result is highly resistant even to vary significant changes in the Prior Probability value. This same resistance to change is evident, although to a slightly lesser degree, even using the Combined Paternity Index number of 545/1, the reported genetic test results excluding the effect of the DNA testing. With a Combined Paternity Index of 49,111/1, the Probability of Paternity is still over 99% even with Prior Probability values close to zero. In effect, the Prior Probability value would have to go to zero 2 for there to be any significant change in the Probability of Paternity.
In assessing the credibility of the parties, the court found the petitioner to have been a highly credible witness and the more credible of the two. In fact, almost all of the petitioner's testimony concerning her relationship with the respondent was corroborated by the respondent's own testimony, most critically, petitioner's testimony that she had sexual intercourse with respondent during the period of conception. The court also credits petitioner's testimony that she had exclusive access with the respondent during the period of conception. Respondent's suggestion that, although acknowledging penetration and ejaculation, he might not have ejaculated inside her, is a disingenuous attempt to create doubt; the court rejects this aspect of his testimony.
In regard to the testimony of the two medical experts in male infertility, the court found petitioner's expert the more convincing in emphasizing that the respondent always possessed viable sperm 3 and that since his potential for conceiving a child was not totally compromised, nor markedly different from other men who have fathered children, the fact that he had intercourse with the petitioner during her fertile period may well have resulted in a pregnancy. In sum, respondent's undisputed below average semen analysis results are a variable to be considered, but when properly considered it should not be given the greater weight recommended by petitioner's expert.
The clash between the other two experts, Drs. Bing and Harris, both qualified as experts in paternity genetic marker testing and evaluation, is more stark and fundamental because their testimonial debate goes to the heart of the proper forensic use and interpretation of paternity blood test results. In analyzing their differences, this court rejects the conclusions of respondent's expert, Dr. Bing, as misleading, and accepts as credible and accurate the testimony of petitioner's expert, Dr. Harris. The reasons are as follows:
The debate between the two experts centers on two related questions: (1) whether it is appropriate and scientifically valid for the laboratory in this case to have employed a Prior Probability value of 0.5; and (2) how to properly interpret the laboratory's reported 99.99% Probability of Paternity in the context of this case. The starting point of any analysis of these questions is clarity about the mathematics of the statistical calculations. Then, both the probative value of the calculations as well as its inherent limitations will become more apparent.
The Combined Paternity Index in this case was 49,111 to 1. This means that the respondent was 49,111 times more likely to have produced the set of obligatory genes-the genes the child must have inherited from its actual father-than that of a random male. This statistic addresses the laboratory blood test results only, i.e., the genetic evidence. The Combined Paternity Index figure itself, of course, is not the same as an ultimate Probability of Paternity estimate since it does not consider any of the non-genetic evidence in the case. To derive a statistical estimate of the ultimate Probability of Paternity, a value would have to also be assigned to the non-genetic evidence. This value is labeled the Prior Probability of Paternity. Because laboratories customarily know little about the non-genetic evidence, or if they do are not qualified to assign a particular value, the majority of laboratories in the United States that do paternity blood testing, by convention, assign a Prior Probability value of 50% or 0.5. The Probability of Paternity estimate is then derived by combining the Combined Paternity Index value (representing the genetic evidence) with the Prior Probability of Paternity value (representing the non-genetic evidence) via a mathematical formula known as Bayes' Theorem. See Paternity Establishment Handbook, National Institute for Child Support Enforcement, Third Ed. 70-74, reprinted in Schatkin, Disputed Paternity Proceedings, Fourth Ed. (Revised), Vol. 2 4 . Both experts, Drs. Bing and Harris agree on the above mathematical derivations and conventions.
In this particular case, using a Combined Paternity Index value of 49,111 to 1, and a Prior Probability value of 0.5, and via Bayes' Theorem, the laboratory arrived at a Probability of Paternity value of 99.99%, which figures are in evidence.
Respondent's expert, Dr. Bing, focused on the use of the 0.5 Prior Probability value. Although conceding that virtually all qualified laboratories use the 0.5 value in computing the Probability of Paternity number, Dr. Bing points out, correctly, that the 0.5 figure is an assumption that can be interpreted to mean that, based on the non-genetic evidence, the respondent is equally likely to be the biological father as he is not to be the biological father. Since this assumption does not take into account the actual non-genetic evidence in any particular case, it may be inaccurate when applied to the facts of a given case. In this case, Dr. Bing claims the assumed 0.5 value is inaccurate because respondent's fertility problem means that he is less likely than not to be the biological father. Since, by mathematical definition, and operation of the conventional formula, lowering the Prior Probability value will lower the resulting Probability of Paternity, the real Probability of Paternity value should be less than the laboratory's reported value of 99.99%, in his view, considerably less.
Dr. Bing's criticism of using an arbitrarily assigned 0.5 Prior Probability value is not new. Similar criticism has been echoed by various commentators in the legal literature. See, e.g. Ellman & Kaye, Probabilities and Proof: Can HLA and Blood Group Testing Prove Paternity? 54 N.Y.U. L.Rev. 1131, 1147-1152 (1979); Martin, Probability of Paternity, NYLJ, May 12, 1989, p. 3 col. 1. In fact, Ellman & Kaye point out that utilizing a 0.5 Prior Probability assumption means that whenever the respondent is not excluded by the blood test, the formula employing Bayes' Theorem will result in a Probability of Paternity greater than 50%, mathematically meeting the burden of proof in a civil case. The authors, acknowledging that all statistical systems that attempt to combine quantifiable genetic test results with less readily quantifiable non-genetic evidence have flaws, recommend a so-called “chart approach” in which the trier of fact is allowed to see the effect of the genetic evidence on a range of assigned Prior Probability values. Id. at 1152-1158.
Curiously, there has been little if any discussion or analysis of the use of the 0.5 Prior Probability in reported New York State paternity cases, at either the trial or appellate level. However, the appellate courts of a number of other states have noted the issue, some analyzing it at great length. See, e.g. Cole v. Cole, 74 N.C.App. 247, 328 S.E.2d 446, affd. 314 N.C. 660, 335 S.E.2d 897 (Supreme Court of North Carolina, 1985) (rejecting 95.98% Probability of Paternity via conventional formula because of strong evidence of sterility due to vasectomy); Plemel v. Walter, 303 Or. 262, 735 P.2d 1209 (Supreme Court of Oregon, 1987) (expert cannot report single figure as the Probability of Paternity but must present figures based on assumed Prior Probabilities ranging from 0 to 100%, i.e., the “chart approach”); In re the Paternity of M.J.B., 144 Wis.2d 638, 425 N.W.2d 404 (Supreme Court of Wisconsin, 1988) (50% Prior Probability not fact sensitive; Probability of Paternity statistic can only be used if there is competent evidence of sexual intercourse during conceptive period; Prior Probability assumption can be addressed by expert witnesses); State of New Jersey v. Spann, 130 N.J. 484, 617 A.2d 247 (Supreme Court of New Jersey, 1993) (in criminal context, expert must report and explain effect of varying range of Prior Probabilities, citing Plemel, supra; contra: State v. Skipper, 228 Conn. 610, 637 A.2d 1101 [Supreme Court of Connecticut, 1994; assumption of 0.5 Prior Probability inconsistent with presumption of innocence] ); Commonwealth v. Beausoleil, 397 Mass. 206, 490 N.E.2d 788 (Supreme Court of Massachusetts, 1986) (although Probability of Paternity calculation has a “basic flaw” due to arbitrariness of 50% Prior Probability assumption, “chart approach” and other alternatives rejected as “unduly complicated”; defense counsel can address issues and problems of a single Probability of Paternity estimate on cross examination).
In 1994, the Legislature (see Laws of 1994, ch. 170, eff. June 15, 1994) amended FCA § 532, creating a rebuttable presumption of paternity if the blood test results “indicate at least a ninety-five percent probability of paternity”. Although § 532 contains no further specificity concerning the derivation of the Probability of Paternity, and makes no mention of the term “Prior Probability of Paternity”, it is more reasonable than not to assume that the Legislature contemplated the laboratory report itself utilizing a Prior Probability assumption of 0.5. The reasons are several: First, the amendment was passed in the context of a long history, both state-wide and nationally, of qualified paternity testing laboratories utilizing the 0.5 statistical convention. See Schatkin, supra at 72; joint AMA-ABA guidelines: Present Status of Serologic Testing in Problems of Disputed Parentage, 10 Fam.L.Quart. No. 3, 247, 263 (1976). Second, the amendment refers to the blood test results themselves and not the other (non-genetic) evidence in the case. Because, by mathematical definition, a Prior Probability assumption is necessary to derive a statistical “Probability of Paternity”, and because the laboratories' information and expertise is and should be confined to the genetic testing itself, it is appropriate for the laboratory to use a 0.5 Prior Probability which is most neutral concerning the non-genetic evidence. Third, by setting a ninety-five percent Probability of Paternity figure in the statute, the legislature implicitly understood the mathematics of the conventional formula, see note 4, supra and accompanying text. As noted in Cole v. Cole, supra at 254, 328 S.E.2d 446: “․ Where 50% is used as the prior probability, the Bayes' Theorem ensures that every alleged father is “probably” the father, i.e., the blood tests results only improve upon the 50% prior probability of paternity. Thus, the probability of paternity is generally probative, at best, where it is 90%, or as some have suggested 95%”. Fourth, by creating a “rebuttable presumption” the Legislature effectively codified existing case law which provided that paternity blood test results should be given strong but not conclusive weight by the trier of fact (see, e.g. Social Servs. v. Klaus D., 188 A.D.2d 381, 591 N.Y.S.2d 388; Allison M. v. James P., 122 A.D.2d 270, 505 N.Y.S.2d 181).
Accordingly, this court rejects respondent counsel's closing argument (which is an extreme extension of Dr. Bing's testimony) that “․ the blood tests, if accepted as valid based on the 0.5 prior probability, violate state and federal constitutional principles [to due process] that mandate consideration of a litigant's unique situation”. This court holds that pursuant to FCA § 532 it is appropriate for a certified laboratory to utilize a Prior Probability value of 0.5 in calculating the Probability of Paternity and for that result to be admitted in evidence. By creating a rebuttable presumption, the 1994 amendment to FCA § 532 permits the court to consider the laboratory's reported Probability of Paternity figure along with the non-genetic evidence in the case. The rebuttable presumption also allows the court to consider questions of interpretation or qualification to the laboratory's reported Probability of Paternity figure, including their use of the 0.5 Prior Probability assumption, through the mechanism of expert witness testimony. See Commonwealth v. Beausoleil, supra. Such expert witness testimony was utilized in this case, and permits the court to consider additional explications of the Prior Probability assumption including, but not limited, to the “chart approach”, noted above.
The flaw in Dr. Bing's testimony is not that he tried to further explicate the meaning of the 0.5 Prior Probability assumption, but that he did so improperly. It was proper to note that a significant piece of non-genetic evidence-such as a fertility problem-should be considered in evaluating ultimate paternity. Where respondent's expert erred was in asking the court to qualify or limit the significance of the 0.5 Prior Probability assumption based on the fertility issue alone. Since the Prior Probability value refers to all the non-genetic evidence in a case, the fertility issue cannot be considered in isolation. Rather, it has to be considered in light of other critical non-genetic evidence introduced in this case, in particular, uncontroverted evidence that respondent had sexual intercourse with the petitioner, including penetration, during the period of conception of this child. In the absence of credible or any evidence that anyone other than respondent had access to petitioner during the period of conception, this fact, as Dr. Bing ultimately acknowledged, would raise, considerably, any assigned Prior Probability value.
The testimony of petitioner's expert, Dr. Harris, was far more probative. He correctly utilized the “chart approach” (see above) in showing the effect of a range of Prior Probability values on the actual genetic test results in this case, namely, the 49,111/1 Combined Paternity Index. When this is done it becomes clear that the Probability of Paternity remains over 99% even with Prior Probability values close to zero. This is not surprising because it has been noted that, by mathematical definition, when the genetic evidence corresponds to a Paternity Index of over 100 to 1, the effect of the non-genetic evidence is small, and when the genetic evidence is very strong (as in this case) it makes little difference what the non-genetic evidence is in calculating the overall Probability of Paternity. See Schatkin, supra at 73. The genetic and non-genetic facts of this case are totally consistent with Dr. Harris's exposition: The genetic evidence is extremely strong in pinpointing respondent as the probable father because the supplement of DNA testing can far more precisely than other systems alone genetically identify a parent. The non-genetic evidence also strongly points to respondent as the father in that non-disputed credible evidence indicates exclusive sexual intercourse by respondent during the period of conception; this evidence is only slightly weakened by respondent's asserted fertility problem since he was not infertile and was still capable of fathering a child. Compare, Cole v. Cole, supra, involving strong evidence of complete sterility.
In sum, petitioner's very credible testimony of exclusive access with respondent during the period of conception; the absence of evidence that petitioner had sexual intercourse with anyone else during the period of conception; respondent's own admission of having had sexual intercourse, including penetration, with petitioner during the period of conception; evidence that although the respondent had a fertility problem he was not completely infertile; and the highly positive blood test result of over 99% Probability of Paternity, all combine to provide evidence that is clear, convincing, and entirely satisfactory that respondent is the father of the subject child and that respondent has failed to rebut the applicable statutory presumption that he is the father.
Therefore, the clerk of the court is directed to enter an order of filiation, and the matter is further set down for a hearing on the issue of support.
[Portions of opinion omitted for purposes of publication.]
1. A 1979 semen analysis indicated a 42 million sperm count (10 million per ml.); 40% good motility (movement); 49% good morphology (shape). An additional specimen indicated a 73 million sperm count (26 million per ml.), with similar motility and morphology.A 1993 semen analysis indicated a 16 million sperm count (1 million per ml.); 40% good motility; 51% good morphology.
2. A Prior Probability value of zero is the equivalent of saying that, as a factual determination, for example, the respondent is sterile; or that sexual intercourse never occurred; or that if sexual intercourse occurred, it was not during the time of possible conception, etc. See Schatkin, Disputed Paternity Proceedings, Vol. 1, § 11A.11 (4th rev. ed.).
3. See note 1, supra.
4. Bayes' Theorem, in regard to calculating paternity, is:E27Probability of Paternity = E37Paternity Index x Prior ProbabilityPaternity Index x Prior Probability + (1-Prior Probability)Schatkin, id. at 72; see also Social Services v. Bart D., 121 Misc.2d 425, 431, 467 N.Y.S.2d 1001 (Fam.Ct., Kings Cty, 1983).
GEORGE L. JUROW, Judge.