More Uncertainty About Reasonable Degree of Medical Certainty

For reasons that are not clear, the Reporters for the American Law Institute’s most recent torts Restatement decided to tackle an expert witness issue, the meaning of “reasonable degree of medical certainty” (RDMC). The discussion is tucked away in a comment and the accompanying note, and might easily be missed by readers interested in the restated principles of tort law.  Restatement (Third) of Torts: Liability for Physical and Emotional Harm § 28, cmt. e (2010).  The phrase RDMC is usually incanted by an expert witness, or in a leading question by an examining lawyer, to qualify the witness’s opinion about a diagnosis, a causal attribution, a prognosis, or some other medical or scientific issue.  Indeed, the phrase is usually modified for non-medical expert witnesses, for whom it takes the form of “reasonable degree of scientific certainty.”  In any event, the phrase is used in tort cases, but also in criminal, contract, equity, trusts and estate, and other cases.  There is nothing unique about the usage in torts cases; nor is there any substantive relationship between the phrase RDMC and the law of torts. 

The Reporters for the Tort Restatement, however, found a “troubling inconsistency” between the standard of certitude for expert witnesses and the standard for the overall burden of proof in most torts cases, the so-called preponderance of evidence.  To resolve the apparent inconsistency, the Reporters suggest that courts adopt a different standard for expert witnesses to use, a “more likely than not” standard.  For most jurisdictions, this standard will be quite novel. The Restatement’s suggested standard has been the subject of previous critiques.  See James Beck & Mark Hermann, “A Critique of ALI’s Proposal to Bar Expert Testimony to a ‘Reasonable Degree of Professional Certainty’,” 2 West’s Medical Malpractice Law Report 1 (May 11, 2007); James Beck, “ALI Draft Would Abolish ‘Reasonable Degree of Professional Certainty’ Requirement” (April 12, 2007); John Day, “ALI Rejects ‘Reasonable Degree of Medical Certainity [sic]’” (May 18, 2007).

The inconsistency perceived by the Reporters, however, is non-existent.  The standards assess two very different measures – one assesses the level of certainty that an expert witness possesses about an opinion that is necessary to the case, and the other assesses the overall quantity of evidence that the party with the burden of proof has presented for each element of every claim that make up his case.  The two standards are not even close to measuring something that can or would conflict.  The independence of the standards can easily be seen when one realizes that expert witnesses in tort cases testify about issues that carry a burden of proof of clear and convincing evidence, as is often the case for tort cases involving fraud, conspiracy, or punitive damage claims.  The law of crimes typically requires proof beyond a reasonable doubt, but when expert witnesses testify about cause of death or of harm, the law does not conflate the burden of proof with the standard for expert witness certainty, and exclude experts who cannot opine about their conclusions beyond a reasonable doubt. 

Rather than creating a “troubling inconsistency” by interpreting RDMC as different from preponderance of evidence, courts are avoiding a conflation of different standards that would lead to a much deeper, more widespread “troubling confusion” throughout all areas of the law that involve expert witness opinion testimony.  One of the cases cited by the Reporters in their note to Comment e illustrates the error.  Bara v. Clarksville Mem. Hosp. Sys., Inc., 104 S.W.3d 1 (Tenn. App. 2002).  The Reporters’ explanatory parenthetical identifies this case as “holding instruction to the jury requiring plaintiff to prove causation to reasonable degree of medical certainty was erroneous and required reversal.”  This point is exactly the error made by the Reporters; the trial court involved had confused, just as the Reporters have, the standard for the jury with the standard of certitude for the expert witness.  Adopting a “more likely than not” standard for expert witness certitude would be a wholesale change in the law, and would not avoid any conflict.  Comment e is based upon a basic category mistake.

 The Restatement acknowledges that its proposal is at odds with the law of several states, such as Pennsylvania, where expert witnesses must opine to a “reasonable degree of professional certainty.”  McMahon v. Young, 442 Pa. 484, 485, 276 A.2d 534, 535 (1971).  Accord Menarde v. Philadelphia Transp. Co., 376 Pa. 497, 501, 103 A.2d 681 (1954); Nestor v. George, 354 Pa. 19, 46 A.2d 469, 472 (1946) (medical opinions of possible, or even probable, causes are incompetent to establish causation); Vorbnoff v. Mesta Machine Co., 286 Pa. 199, 206, 133 A. 256, 258 (1926) (a “less direct expression of opinion would fall below the required standard of proof, and therefore would not constitute legally competent evidence”); Griffin v. University of Pittsburgh Medical Center-Braddock Hosp., 950 A.2d 996, 2008 Pa. Super. 104, * (2008), alloc. denied, 601 Pa. 680; 970 A.2d 431 (2009); Corrado v. Thomas Jefferson Univ. Hosp., 790 A.2d 1022, 1027, 1031 (Pa. Super. 2001); McCann v. Amy Joy Donut Shops, 325 Pa. Super. 340, 472 A.2d 1194 (1984); Schwartz v. Feldman, 196 Pa. Super. 492, 175 A.2d 153 (1961).  See also Beezer v. Baltimore & Ohio RR, 107 F.Supp. 361, 362 (W.D.Pa. 1952), aff’d, 203 F.2d 954, 954 (3d Cir. 1953) (per curiam).  Pennsylvania courts have made clear that this level of certainty is more than a mere probability, and this requirement is especially important in Pennsylvania, where the reliability standard is either unclear or not confidently applied by its courts.

Pennsylvania is certainly not alone in interpreting “reasonable degree of medical certainty” to mean more than a bare probability.  The law of Kansas, like that of Pennsylvania, requires causation to be proven by expert testimony to a reasonable degree of medical certainty.  Johnston v. United States, 597 F.Supp. 374, 412 (D. Kan. 1984).  As the court noted in Johnston, a statistical method that attributes a greater than 50 percent probability to two events being causally related does not satisfy the reasonably certain level of proof.  Id.  Statistical risk calculations used to show causation require a “healthy degree of skepticism,” and should not be relied upon to give a deceptively easy answer to a complex issue.  Id. at 394 – 95.

The Reporters criticize the traditional RDMC standard because it fails to assure the quality of expert witnesses’ “qualifications, expertise, investigation, methodology, or reasoning.”  This is, of course, true but irrelevant.  The RDMC standard is not the only standard that applies to expert witnesses, and there is no reason to believe that it ever was used to as the sole guarantee of adequacy of every aspect of expert witness opinion testimony.  Indeed, as numerous cases have pointed out, the standards of Federal Rules of Evidence 702 and 703 are totally independent of witness qualifications.  Some very expert expert witnesses have been precluded from testifying to dubious opinions.

Comment e also suffers from a temporal incoherence.  The Restatement urges the adoption of “the same preponderance standard that is universally applied in civil cases.  Direct and cross-examination can be employed to flesh out the degree of certainty with which an expert’s opinion is held and identify opinions that are speculative and therefore inadmissible.” As noted above, the preponderance standard is not universally applied in civil cases, and certainly not in criminal cases.  More important, making a standard of admissibility turn on the outcome of direct and cross-examination is incoherent because the opinion must be first admissible before it can be the subject of these examinations.  Leaving the issue to motions to strike would unrealistically require courts and juries to “unring the bell,” which all have heard in the courtroom.

 The Restatement further criticizes RDMC because it “suggests reliance on medical or scientific standards of proof,” and “seems to impose a high threshold for the opinion to be admissible.”  Id. at Comment e.  These criticisms further reveal that the Restatement has wondered into the field of evidence and away from the subject of torts.  The concern implicit in these criticisms, however, curiously suggests that physicians and scientists should not rely upon medical or scientific standards in the courtroom.  (Scientists would rarely make the mistake lawyers make of referring to evidence as “proof.”)  The Restatement, having wondered into the field of expert evidence, and having waived off medical and scientific standards, begs the question as to what standards should be employed by expert witnesses. 

The Restatement’s suggestion that the expert witnesses should be held to give opinions that are “more likely than not” true fails to resolve the problems that the Reporters perceive in the RDMC standard.  Even in civil cases, subject to a preponderance standard for the burden of proof, the Restatement’s comment e would be unworkable. 

First, the certainty requirement applies to all expert testimony, regardless whether proffered by the party with the burden of proof on the issue.  If we were to follow the Reporters’ reasoning, we would have to abandon any standard for expert witnesses presented by the party without the burden of proof, to ensure a concordance between the two standards.  We would also have to raise the burden on expert witnesses when they offer opinions that are essential to satisfying elements of a claim that requires more evidence than a mere preponderance.

Second, expert witness opinion testimony is often based upon assumptions, which the trier of fact may find are not established or which are themselves subject to some level of uncertainty.  The expert witness’s opinion, expressed to a mere probability, will often then fail to be sufficiently weighty to support a verdict because of the uncertainty in the witness’s assumptions, premises, or personal credibility.  For example, a jury, in evaluating an expert witness’s opinion that was barely “more likely than not,” would have to find that the witness was absolutely credible, and that the bases of his opinion were impeccable if they were to vote for the party with the burden of proof.  Similarly, trial courts might well find themselves granting judgments notwithstanding the verdict if expert witnesses gave opinions that barely met the preponderance standard because any doubt about the credibility or witness’s premises would unravel the sufficiency of the case given to the jury.

Third, there are even more far-reaching problems with simply substituting “more likely than not” for RDMC as a threshold requirement of expert witness testimony.  The Restatement Comment implies that expert witnesses can discern the difference between an opinion that they believe is “more likely than not” and another which is “as likely as not.”  On some occasions, there may be opinions that derive from quantitative reasoning, for which an expert witness could truly say, with some level of certainty, that his or her opinion is “more likely than not.”  On most occasions, an expert witness’s degree of certainty is a qualitative opinion that simply does not admit of a quantitative characterization.  The Restatement’s comment perpetuates this confusion by casting the RDMC standard as a bare probability.

Fourth, the Restatement’s Comment suggests that expert witnesses are themselves expert in their own level of certainty, and that they have the training and experience to distinguish an opinion that is 50.1% likely from another that is only 50% likely.  The assignment of precise mathematical probabilities to personal, subjective beliefs is a doubtful exercise, at best.  See, e.g., Daniel Kahneman and Amos Tversky, “Judgment under Uncertainty:  Heuristics and Biases,” 185 Science 1124 (1974). 

Given all the vagaries of “more likely than not,” the Restatement’s novel standard is not likely to bring new clarity to this aspect of expert witness opinion testimony.  Ultimately, Comment e, to Section 28, is a frolic and detour in the law of torts.  Although professionals who find themselves on the witness stand may not recognize the legalistic locution of RDMC, they immediately recognize that there are some opinions that are not sufficiently strongly to act upon in a professional context.   The courts that impose a RDMC standard similarly recognize that the mere conjunction of expertise and opinion is an insufficient warrant to permit a jury to receive the opinion.  See Schachtman, “Reasonable Degree of Medical Certainty” (Dec. 3, 2010).