Beck & Herrmann on "Reasonable Medical Certainty": Part I

Quite some time ago now, we promised to respond to a Beck & Herrmann post decrying an ALI proposal to abolish any requirement that experts offer their opinions to a “reasonable degree” of medical, professional, or scientific “certainty.” (Call this the “RDC” rule for short.) The ALI proposal would abrogate any RDC requirement and demand only that the expert hold his or her opinion to be more likely true than not -- at least in the context of opinions offered to prove causation in tort cases involving physical harm.

Beck & Herrmann attempt to rebut three arguments offered by the ALI in favor of abandoning any RDC requirement: (1) that the medical and scientific communities have no such “reasonable certainty” standard; (2) that the requirement imposes a more demanding standard for admissibility than the law imposes for satisfaction of the burden of persuasion in civil cases; and (3) that the RDC standard affords no effective guarantee of the soundness of the expert’s analysis.

We’ll respond to Beck & Herrmann’s specific critiques of the three ALI arguments in a second installment. In this first post, we will: (a) discuss how the RDC rule appears to have originated; (b) attempt some description of the demands it may currently impose; and (c) consider some of the functions it might aspirationally perform.

A. Historical Antecedents

The rule is most routinely implicated, of course, in expert testimony from physicians – the paradigmatic evidentiary context on which Beck & Herrmann also focus. They say the rule has been in place for “generations.” To be more precise, its judicial adoption appears to date from the 1960s.

In a 1998 Maryland Law Review article, Jeff L. Lewin (a Pennsylvania attorney and an adjunct professor of law at Widener University) makes a convincing case that the practice of asking medical experts to state their opinions to a “reasonable degree of medical certainty” was first adopted by the Chicago trial bar, sometime before 1930, in specific response to distinct features of the local legal ecology. See generally Jeff L. Lewin, The Genesis and Evolution of Legal Uncertainty About “Reasonable Medical Certainty,” 57 Md. L. Rev. 380 (1998).

In the beginning, according to Lewin, was the “reasonable certainty” rule, which involved the standard of proof for damages attributable to medical conditions from which the plaintiff might suffer in the future. By the end of the nineteenth century, it was a well-established principle under the Illinois precedents, as well as the law of other jurisdictions, that a plaintiff could not recover damages attributable to future medical symptoms or conditions unless the plaintiff could show they were reasonably certain to occur. This was originally treated as a substantive sufficiency requirement but was later adopted, in a series of Illinois decisions dating from early in the twentieth century, as a test for the admissibility of expert physician opinion on future damages.

During the same period, Lewin notes, the Illinois Supreme Court adopted a form of the “ultimate issue” rule that barred physicians from expressing definitive opinions on causation. Testimony that some factor did in fact cause the plaintiff’s injuries was seen as an encroachment on the jury’s ultimate fact-finding prerogative. Physician testimony about what “might have caused” the injury was permissible, but expert opinion that some factor actually did cause the injury was inadmissible. To modern legal ears, such a rule may sound paradoxical. But such was Illinois law, in the Olden Days.

These “reasonable certainty” and “ultimate opinion” rules were obviously in significant tension. Although some of the details remain shrouded behind history’s veil, Lewin offers substantial lexical evidence that the phrase “reasonable medical certainty” originated among personal injury and worker’s compensation counsel in Illinois, in the effort to frame questions compliant with both rules – e.g., “Doctor, have you formed an opinion, based on a reasonable medical certainty, as to what may or could have been the cause of plaintiff’s condition?” By the 1930s, Illinois counsel were using the “reasonable medical certainty” locution with some regularity. Lewin attributes its spread to other jurisdictions largely to a best-selling trial manual published in 1935 by Irving Goldstein, an Illinois lawyer and adjunct professor at Northwestern University Law School. By the 1950s, published decisions in other states started to reflect counsel’s growing use of the Goldsteinian formula, to a point where the phrase began to find a habitat in West’s Keynotes.

In the 1960s, the expression’s incidence in judicial opinions rose exponentially. Many decisions from this period attached no special legal significance to the phrase, which often appeared in quotations from trial testimony. In some jurisdictions, however, courts began to embrace “reasonable medical certainty” as a standard for admissibility. Lewin posits (plausibly in our view) that by then, the phrase had come to enjoy such widespread currency as a term of testimonial art that some lawyers and judges simply assumed it embodied a hallowed legal requirement, often with little explicit reflection on what, precisely, it meant.

However that may be, the standard was not universally accepted – and where accepted, it was not always assigned the same meaning. In Alaska, Nebraska, and Washington, for example, the courts adopted a “reasonable medical certainty” requirement but interpreted it to mean only “more probable than not.” In other jurisdictions, the courts articulated a “reasonable medical certainty or probability” requirement without indicating whether certainty and probability came to the same thing. In a handful of states – Lewin mentions Pennsylvania, South Carolina, and perhaps Tennessee – the standard was interpreted to bar testimony whose truth the expert characterized as probable but less than reasonably medically certain.

B. Current Doctrine

In the wake of the 1975 adoption of the Federal Rules of Evidence, the Supreme Court’s 1993 decision in Daubert, and the adoption of the federal standards for expert testimony in many states, courts do not nowadays lack for methods by which to screen expert testimony. In consequence, the RDC rule plays a less prominent evidentiary role than it may have done at common law. Colorado, for example, recently jettisoned the “reasonable medical certainty/probability” standard altogether. See People v. Ramirez, 155 P.3d 371 (Colo. 2007). What the standard demands, where still in force, remains the subject of substantial variation and considerable uncertainty. At different times and in different places, it has represented a sufficiency requirement, an admissibility criterion, or some vague blend of both. It has sometimes been extended to nonmedical expert testimony, but its applicability in that context generally rests on a less secure footing than with medical experts. A number of courts have held that the standard imposes no requirement for magic language or a specific form of words, yet motions to exclude expert testimony for lack of a formulaic recital are not uncommon, and an expert’s departure from the customary verbal formula, whether on direct or on cross, may increase the risk that the expert’s testimony will be discounted or excluded as mere speculation (as though the law could countenance no middle ground between “certainty” and idle conjecture).

There is considerable ambiguity, in short, about what, precisely, may be the current status and contours of the standard whose potential “abolition” Beck & Herrmann bemoan.¹ This, by itself, might be counted as a mark against a legal standard, but we will not pursue that argument any further for now.

C. The Standard’s Potential Functions

Whatever its actual current meaning in the jurisdictions still adhering to it, the requirement that expert testimony reflect a “reasonable degree” of medical, professional, or scientific “certainty” might hypothetically assume any of multiple forms. The rule might require the witnesses themselves to attest that their opinions are held to a “reasonable degree of certainty.” It might condition admissibility on a judicial determination that their testimony is of RDC caliber. It might make RDC a part of the sufficiency calculus. Or it might combine some or all of those requirements:

The “Reasonable Degree of Certainty” Standard: Seven Potential Versions
	Expert Must Profess RDC	Actual RDC Is a Necessary Condition for Admissibility	Actual RDC Is a Necessary Condition for Sufficiency
1	X	X	X
2	X	X
3	X		X
4	X
5		X	X
6		X
7			X

The appropriate form (if any) for an RDC rule might depend on the functions it is intended to perform. We will focus here on just two such functions, because we think they’re the most plausible candidates. We will paint them with an admittedly broad brush.

1. The Liability-Minimizing Function

One potential function for the rule might be to limit civil liability for certain specific kinds of claim (e.g., suits against health care providers, pharmaceutical companies, and medical device manufacturers) to cases where the specialized or technical analysis supporting the imposition of liability reaches a higher than usual threshold of … well, let’s call it “warranted assertibility.” We say “higher than usual threshold” because the different jurisdictional flavors of Daubert’s reliability calculus (including the analyses followed in several states where the evidentiary assessment is still nominally conducted under Frye but increasingly exhibits Daubertian features) already make substantial demands in this area – demands considerably more stringent than those formerly prevailing at common law.

The rule’s use to perform this liability-limiting function would seemingly represent a policy decision that certain fields of professional and/or commercial endeavor deserve special forms of legal protection (a policy animus also operative, for example, in many state statutes governing the competency of experts to testify on the standard of care in malpractice cases). We do not propose to weigh in on that policy question here. We will instead offer only a few more limited observations.

First, the pursuit of substantive agendas through evidentiary or procedural means has sometimes resulted in debasement of policy discourse (because the strategy is sometimes adopted for the very purpose of presenting substantively controversial policy agendas in misleading terms – i.e., in bland procedural clothing). Subject to that caveat, however, there is nothing instrinsically reprehensible or wrongheaded about pursuing substantive goals through procedural methods, so long as the substantive policy agenda is made explicit. To their credit, Beck & Herrmann are reasonably disclosing about theirs.

Second, this strategy for insulating certain favored classes of defendant from certain forms of liability would be an indirect strategy. Not that there’s anything wrong with that. We do it all the time. We try to protect criminal defendants from wrongful incarceration by giving them the right to confront the witnesses against them. We try to prevent accidents by regulating motor vehicle speed. But whenever we pursue some policy goal through indirection, it is reasonable to ask what incremental contribution the means will make to achieving the end, at what collateral cost, and whether some other means might be better tailored to the task.

Third, although the liability-limiting agenda under discussion would be consistent with several alternative implementations of the RDC rule, it does suggest certain directions for implementation. It suggests, for example, that the requirement might apply specifically to medical testimony, or testimony about the causation of personal injury (or certain types of claim involving personal injury), rather than to expert testimony more generally. It suggests, as well, that the requirement might extend beyond the mere demand that experts employ a certain form of words in describing the confidence with which they hold their opinions (i.e., beyond Version 4). It suggests, that is, that in the course of their admissibility and/or sufficiency decisions, the courts might also participate in evaluating whether the evidence possesses whatever degree of “certainty” the rule requires (as in Versions 1-3). Such a judicial evaluation might even be prescribed in the absence of any requirement for an expert attestation (as in Versions 5-7).

2. The Confidence-Testing Function

A second function the rule might perform would relate more to the goal of promoting the trustworthiness of expert opinion testimony more generally, by adding a layer of screening at the level of the expert’s own intellectual conscience. Once again, such a function would be consistent with multiple alternative implementations of the rule. But once again, some plausible directions for implementation are suggested. Even if it were felt, for example, that existing tests for sufficiency and admissibility provide basically adequate tools for gauging the probativeness and warranted assertibility of expert testimony (or that they would provide adequate tools for those purposes after some judicious tinkering), concern might still arise over the sincerity of expert opinion; its honesty, if you will. Because it is precisely opinion testimony we’re discussing, perjury laws cannot be expected to do a very good job at deterring experts from presenting opinions that may be methodologically defensible, to one degree or another, but in which the experts do not, in their heart of hearts, feel much confidence (i.e., truly believe). No doubt the temptation for experts to bend their testimony in the client’s favor, within whatever limits the rules on reliability may permit, is not one that any plausible evidentiary measure could hope to eliminate fully. But something might be gained from requiring experts to say, under oath, that they hold their opinions at some reasonably high level of confidence. At least some subset of experts -- some whose own consciences would not permit them to make such an attestation, plus some others who would fear for their reputations if they did – would balk.

This second vision of the “reasonable degree of certainty” rule suggests that forms of words should be important. No weaseling: are you reasonably certain, or not? Version 4 represents the purest implementation of this vision.

Our throat now cleared, we will turn, in the next installment, to the specifics of the Beck & Herrmann argument.

_________________________________

¹ The very decisions that Beck & Herrmann cite as exemplary embodiments of the RDC standard serve as vivid illustrations of the standard’s ambiguity. They say they are “very comfortable,” for example, with the standard articulated in State v. Jackson, 92 Ohio St. 3d 436, 448, 751 N.E.2d 946, 961 (2001). In the course of rejecting an objection that testimony should have been expressed as “probable” rather than “reasonably certain,” the Jackson court held that “reasonable certainty” is synonymous with “probability.” To us, that sounds suspiciously like the “coin-flip” test that Beck & Herrmann decry. The hermeneutic trail, however, does not end there. What, after all, does “probability” really mean to the Ohio Supreme Court? The reader might hope to find enlightenment in the Jackson opinion’s citation of State v. Benner, 40 Ohio St. 3d 301, 313, 533 N.E.2d 701, 714 (1988). In Benner, a coroner testified that the cause of death was “more than likely strangulation or asphyxia.” On appeal, the defendant protested that this testimony did not satisfy the RDC standard. The Benner court rejected that objection, and/or held that any error in admitting the testimony was harmless. The Benner opinion first equated “reasonable certainty” with “probability.” It then offered a lengthy if ultimately inconclusive exegesis of the latter term, drawing heavily on standard and legal dictionaries. In the course of that exegesis, it noted: (1) that one of its earlier decisions had equated “probable” with “extremely likely”; (2) that the wisdom of another earlier decision, in which the court had posited that something “likely” is less certain than something “probable,” was questionable; (3) that the OED defines “probable” and “likely” in basically equivalent terms; (4) that Black’s Law Dictionary defines “likelihood” as “probability,” and also as connoting something less than reasonable certainty; (5) that numerous decisions in other jurisdictions have equated “probable” and “likely”; (6) that the coroner’s characterization of his opinion as “more than likely” placed his testimony on a point located somewhere on the continuum between “likely” and “extremely likely”; and (7) that any error was harmless. Are we clear now? Not in the estimation of the Ohio Court of Appeals. See State v. Riley, 2007-Ohio-879, at ¶¶ 26-28 (Ct. App. Wood County 2007) (Benner inapposite because decided prior to the 1994 amendments to Ohio Evid. R. 702), appeal dism’d in part, 866 N.E.2d 1092 (Ohio 2007). Indeed, Ohio’s intermediate appellate courts appear to question whether the RDC rule remains viable following Ohio’s adoption of a Daubert regime. See, e.g., Coe v. Young, 145 Ohio App. 3d 499, 502-04 763 N.E.2d 652, 655-56 (Ashtabula County 2001) (in light of Daubert’s focus on methods rather than the conclusions they generate, “there is no requirement that an expert utter any ‘magic language’; i.e. that his opinion was within the reasonable degree of certainty or reasonable degree of certainty within the particular knowledge of his professional experience [sic]”).

Beck & Herrmann are also “very comfortable” with the very different version of the RDC standard espoused in McMahon v. Young, 442 Pa. 484, 485-86, 276 A.2d 534, 535 (1971) (mere probability insufficient for medical testimony), and Corrado v. Thomas Jefferson Univ. Hosp., 790 A.2d 1022, 1031 (Pa. Super. 2001) (even “very highly probable” would be insufficient). Pennsylvania has not adopted Daubert and continues to adhere to Frye. In particular, Pennsylvania adheres to the standard variant of Frye under which general acceptance is a precondition for the admissibility of only “novel” scientific evidence. The state’s peculiarly strict version of the RDC rule thus functions, to some degree, as a surrogate or substitute for Daubertian gatekeeping – a highly imperfect one, some might say.

Both the Ohio and Pennsylvania courts have sometimes characterized the RDC standard as going to “competency.” This poses potential issues under Fed. R. Evid. 601 about the standard’s applicability in diversity proceedings, adding another layer of uncertainty.