Not one, but two, of our readers emailed us about a recent Sixth Circuit decision, Best v. Lowe’s Home Centers Inc., 2009 WL 1010883 (6th Cir. April 16, 2009), when it came down. Unfortunately, we had other things going on – and Best wasn’t a drug/device case – so we didn’t get around to examining that decision until now. As the title of this post indicates, Best had to do with the validity of “differential diagnosis” under Daubert (for you non-lawyers, that’s the name of the most important Supreme Court case on admissibility of expert testimony as trial evidence) as a diagnostic tool when the technique is employed by a plaintiff’s expert as a basis for a causation opinion. As the title also indicates, we’re not particularly pleased with the outcome.
Best involved chemical with the whopper of a name: “2-Propen-1-aminium, N, N-dimethyl-N-2-propenyl-chloride.” We can’t tell you what any of that stuff means, but apparently, it’s used to treat water in swimming pools. Plaintiff claimed he was splashed in the face with this stuff by the negligent actions of the defendant’s employee. As a result of the exposure, plaintiff alleged he permanently lost his sense of smell (the technical term is “anosmia”). 2009 WL 1010883, at *1.
The Daubert issue arose when the plaintiff’s expert, who was board-certified in otolaryngology (and thus undisputedly qualified), used “differential diagnosis” to arrive at a causation opinion. Differential diagnosis has been defined as “A physician’s consideration of alternative diagnoses that may explain a patient’s condition.” Fed. Jud. Center, Reference Manual on Scientific Evidence (2d), “Reference Guide on Toxicology,” at 433. Or, at more length:
Expert witnesses and courts, however, frequently use the term “differential diagnosis” to describe the process by which causes of the patient’s condition are identified, particularly causes external to the patient. Additionally, courts sometimes characterize causal reasoning as “differential etiology,” a term not used in medical practice, but one that more closely suggests the determination of cause. For the sake of clarity and consistency, this reference guide uses the term “differential diagnosis” in its traditional medical sense, that is, referring to the diagnosis of disease, and refers to the process of identifying external causes of diseases and conditions as “determining cause,” “determining external cause,” or some similar phrase, as the circumstances warrant.
Id. “Reference Guide on Medical Testimony, at 443-44 (footnotes omitted). If anybody’s interested in following the distinction between “differential diagnosis” and “differential etiology” further, see McClain v. Metabolife International, Inc., 401 F.3d 1233, 1252 (11th Cir. 2005), otherwise, follow us.
In Best, there were no applicable epidemiological studies – not that helped the plaintiff, anyway – so the plaintiff’s expert (who was also the original treating physician, which helps) reached the conclusion he was paid to reach via differential diagnosis. There was pretty good temporal evidence, although not directly related to loss of smell: “immediately [after the exposure] he had suffered from irritation and burning of his skin, irritation to his nasal passages and mouth, dizziness, and shortness of breath. 2009 WL 1010883, at *2. Four months after the incident, plaintiff first saw the expert (as a treater), reporting “clear drainage from his nose following the spill and that he eventually lost his sense of smell completely.” Id.
Exposure data, on the other hand, was downright lousy – nobody ever quantified the exposure, and it was of a fleeting, as opposed to chronic, duration. Id. at *3. The expert was also unable to “to determine the threshold level of exposure that could cause harm.” Id.
The expert selected a number of alternative possible explanations for anosmia, including use of a almost a dozen prescription drugs, and proceeded to “rule out” all but one of them. Id. at *2.
The other half of differential diagnosis – “ruling in” the target chemical – was also rather weak. The substance was “ “irritating to the mucous membrane and upper respiratory tract” and “[m]ay be harmful if inhaled,” but its “acute inhalation toxicity” was “undetermined,” and there was no data linking it to anosmia. Id. The expert speculated that the chemical, which he called a “chlorine derivative,” caused some sort of chemical burn to the relevant nerve endings. Id. But he had failed to examine the allegedly affected area when he had the chance. Id. at *1.
The district court threw out as “speculative” the differential diagnosis under Daubert, because it was based on temporal association and lack of any other obvious alternative. 2009 WL 1010883, at *3. The Sixth Circuit reversed, finding the differential diagnosis sufficient.
The Sixth Circuit faulted the district court’s reliance upon Moore v. Ashland Chemical Inc., 151 F.3d 269 (5th Cir. 1998) (en banc), which it interpreted as “implicitly” rejecting differential diagnosis altogether as a basis for a causation opinion. 2009 WL 1010883, at *9. In the Sixth Circuit, a properly conducted differential diagnosis was proper “methodology” under Daubert. Best, 2009 WL 1010883, at *6. Because the trial court used the wrong standard, and “did not recognize that differential diagnosis is a valid technique,” the appellate court avoided the abuse of discretion standard that Daubert determinations usually receive on review. 2009 WL 1010883, at *6.
Instead, the court proceeded with “de novo” review – not only with respect to whether any differential diagnosis was OK, but also as to the validity of this expert’s specific opinion. Id. We think the Sixth Circuit went overboard there, and that error on whether differential diagnosis can ever be a proper Daubert technique (if error there was) does not justify also reducing the appellate deference to be given to the trial court’s evaluation of the employment of that technique in the particular case.
Evaluating the case-specific differential diagnosis on a step-by-step basis, Best first held that the expert determined, as well as anyone could, that the plaintiff was not lying about having lost his sense of smell. 2009 WL 1010883, at *7-8. We don’t have any real problem with that. The expert ran the necessary tests. Although the results were close to “malingering,” in the absence of incontrovertible physical proof, it’s a jury’s role to conclude that a plaintiff is fabricating an injury.
We have more problems with the “ruling in” aspect of the Sixth Circuit’s opinion. The court’s reliance upon the expert’s treatment of “a list of possible causes for the injury, including virus, accident, brain tumor, brain surgery, exposure to chemicals, medications, or an idiopathic (unknown) cause,” 2009 WL 1010883, at *8, isn’t really on point. That’s ruling out, not ruling in. The court admits that there’s no “published material” linking the chemical to anosmia. At best, the safety data for the chemical was inconclusive, with mention of “irritation” but nothing about loss of smell. The Sixth Circuit instead relied upon the expert’s experience with other patients who were exposed to other “chlorine derivatives” – not even the same chemical. Id.
We question any differential diagnosis that relies primarily upon analogies to some other purportedly similar chemical. As the court stated in McClain:
[A]nother methodological problem undermines [the expert’s] analogical approach. . . . [E]ven minor deviations in chemical structure can radically change a particular substance’s properties and propensities. [The expert] failed to show that the [] analogy is valid or that the differences in chemical structure between [the two chemicals] make no difference. He simply assumes its validity without offering any scientific evidence. As he said, one presumes the same effect by drugs in the same class until proven otherwise. Such presumptions do not make for reliable opinions in toxic tort cases.
401 F.3d at 1246 (citations and quotation marks omitted). “[U]sing this ‘guilt by association’ inference in their methodology is of questionable scientific reliability.” Caraker v. Sandoz Pharmaceuticals Corp., 188 F. Supp.2d 1026 (S.D. Ill. 2001); accord, e.g., Rider v. Sandoz Pharmaceuticals Corp., 295 F.3d 1194, 1201 (11th Cir. 2002); Hollander v. Sandoz Pharmaceuticals Corp., 289 F.3d 1193, 1207 (10th Cir. 2002); Glastetter v. Novartis Pharmaceuticals Corp., 252 F.3d 986, 990 (8th Cir. 2001); Hans v. Matrixx Initiatives, Inc., 2006 WL 5229820, at *5-6 (W.D. Ky. Sep. 29, 2006) (rejecting chemical analogy in another anosmia case); Newton v. Roche Laboratories, Inc., 243 F.Supp.2d 672, 677 (W.D. Tex 2002).
The Sixth Circuit’s reliance (2009 WL 1010883, at *8) upon McCullock v. H.B. Fuller Co., 61 F.3d 1038 (2d Cir. 1995), is also shaky, because: (1) the expert in McCullock relied on published literature (“reference to various scientific and medical treatises,” 61 F.3d at 1044), and (2) McCullock affirmed admission of expert testimony under an abuse of discretion standard, id.).
Under Daubert, a valid differential diagnosis passes muster only if the expert first shows general toxicity in some reliable fashion:
The first step in the diagnostic process is to compile a comprehensive list of hypotheses that might explain the set of salient clinical findings under consideration. . . . The issue at this point in the process is which of the competing causes are generally capable of causing the patient’s symptoms or mortality. Expert testimony that rules in a potential cause that is not so capable is unreliable. . . . It is important to realize that a fundamental assumption underlying differential diagnosis is that the final, suspected “cause” must actually be capable of causing the injury.
Clausen v. M/V New Carissa, 339 F.3d 1049, 1057-58 (9th Cir. 2003) (citations and quotation marks omitted). Just tossing “differential diagnosis” around doesn’t cut it:
No one doubts the utility of medical histories in general or the process by which doctors rule out some known causes of disease in order to finalize a diagnosis. But such general rules must . . . be applied fact-specifically in each case. The underlying predicates of any cause-and-effect medical testimony are that medical science understands the physiological process by which a particular disease or syndrome develops and knows what factors cause the process to occur. Based on such predicate knowledge, it may then be possible to fasten legal liability for a person’s disease or injury.
Black v. Food Lion, Inc., 171 F.3d 308, 314 (5th Cir. 1999). Given Black, we question the Sixth Circuit’s conclusion that Moore represented an across-the-board rejection of differential diagnosis as a technique by the Fifth Circuit.
Thus a plaintiff’s expert cannot assume what s/he is trying to prove in a differential diagnosis. McClain, 401 F.3d at 1253. “Where an expert employs differential diagnosis to ‘rule out’ other potential causes for the injury at issue, he must also ‘rule in’ the suspected cause, and do so using scientifically valid methodology.” Ruggiero v. Warner-Lambert Co., 424 F.3d 249, 254 (2d Cir. 2005). The first step is proof that the substance is capable of causing the injury at issue. “In the absence of such a foundation for a differential diagnosis analysis, a differential diagnosis generally may not serve as a reliable basis for an expert opinion on causation in a toxic tort case. McClain, 401 F.3d at 1253.
Looking just at the evidence discussed in the Sixth Circuit’s opinion, we don’t think there was enough “ruling in” in Best to justify reversing a trial court’s exclusion ruling. We don’t think the Sixth Circuit thought there was either – which explains the sleight-of-hand with the standard of review.
The Best court also reviewed the expert’s “ruling out” methodology. 2009 WL 1010883, at *9. We don’t have much problem with the main rationale that the court used. The defendant argued that the expert only ruled out nine of ten possible drugs. Id. But as to the other one, the court stated that the defendant “presented no evidence that [the drug] might cause anosmia.” Id. Fair enough. Assuming that was the case, there’s no protest from us. We tend to agree that, for an alternative cause to defeat a differential diagnosis, there must be at least some evidence that it in fact causes the injury at issue – that it is, in fact, a potential alternative cause.
Where we do have more problem is the Sixth Circuit’s one-line blow off of idiopathic causation. 2009 WL 1010883, at *9 (the expert “concluded, based on his own experience, that an idiopathic anosmia would not appear over such a short period of time”). Where did that “short period of time” excuse come from? If it’s an idiopathic condition, then by definition it’s of unknown origin – and thus could have been in the offing throughout the plaintiff’s life – not just from the date of the incident. Maybe the defendant didn’t press the point, since there’s no discussion in Best as to the extent of idiopathic causes for anosmia. But we think it’s important.
In considering idiopathic causes, it’s critical that, if there’s a large idiopathic component, that the defendant get evidence of that affirmatively into the record. See Bland v. Verizon Wireless LLC, 538 F.3d 893, 897 (8th Cir. 2008) (differential diagnosis invalid where causation of a disease is “is unknown” in “the majority of cases”); Redfoot v. B.F. Ascher & Co., 2007 WL 1593239, at *11 (N.D. Cal. June 1, 2007) (“differential diagnosis is faulty because [the expert] failed to consider. . .that the cause of autism is not known today”); Doe v. Ortho-Clinical Diagnostics, Inc., 440 F. Supp.2d 465, 478 (M.D.N.C. 2006) (strong likelihood of unknown causes “serves to negate. . .use of the differential diagnosis technique”); Valentine v. PPG Industries, Inc., 821 N.E.2d 580, 599-600 (Ohio App. 2004) (“differential diagnosis is not a reliable technique for identifying causation” of a disease because “medical science does not enable physicians and other scientists to pinpoint a cause. . . . [T]he present state of scientific knowledge on the cause of [the disease] precludes reliability in this context”), aff’d, 850 N.E.2d 683 (Ohio 2006).
The court in Best relied heavily on the Third Circuit’s analysis of differential diagnosis in In re Paoli Railroad Yard PCB Litigation, 35 F.3d 717 (3d Cir. 1994), which it found “instructive.” 2009 WL 1010883, at *6-7. Paoli Yard makes quite clear, though, that differential diagnosis becomes unreliable where an expert witness fails to articulate a reason for ruling out alternative causes that appear in a plaintiff’s medical records, history, or examination. Thus, “a requirement that experts at least consider alternative causes” is “at the core of differential diagnosis.” 35 F.3d at 759 (3d Cir. 1994). For that reason:
[P]art of differential diagnosis is using these standard techniques to rule out alternative causes – thus, where a defendant points to a plausible alternative cause and the doctor offers no explanation for why he or she has concluded that was not the sole cause, that doctor’s methodology is unreliable.
Id. at 759 n.27. Cases following Paoli Yard agree. While a medical expert need not “rule out categorically all other possible causes” for an injury, “[o]bvious alternative causes need to be ruled out.” Heller v. Shaw Industries, Inc., 167 F.3d 146, 156 (3d Cir. 1999) (citation and quotation marks omitted). See, e.g., Kannankeril v. Terminix International, Inc., 128 F.3d 802, 808 (3d Cir. 1997) (“the defendant may point to a plausible cause of the plaintiff’s illness other than the defendant’s actions. It then becomes necessary for the plaintiff’s expert to offer a good explanation as to why his or her conclusion remains reliable”); Wicker v. Consolidated Rail Corp., 371 F. Supp. 2d 702, 715 (W.D. Pa. 2005) (an “expert applying differential diagnosis [must] explain why he did not find a specific alternative cause when challenged to do so by the opposing party; failure to reasonably explain why a conclusion of such alternative cause was not reached would therefore result in a finding of unreliability”); Bart v. Certainteed Products, Inc., 2005 WL 182711, at *3 (E.D. Pa. Jan. 27, 2005) (excluding differential diagnosis due to “the short shrift paid to possible alternative causes”); Soldo v. Sandoz Pharmaceuticals Corp., 244 F. Supp.2d 434, 567 (W.D. Pa. 2003) (excluding differential diagnosis that “does not reliably rule out reasonable alternative causes”).
So do non-Third Circuit courts. Differential diagnosis is viewed as “a diagnosis based upon ruling out all other causes.” Courtaulds Fibers, Inc. v. Long, 779 So.2d 198, 202 (Ala. 2000). The expert employing it must “attempt to consider all possible causes, or to exclude each potential cause until only one remained.” Turner v. Iowa Fire Equipment Co., 229 F.3d 1202, 1208 (8th Cir. 2000) (excluding opinion for failing to rule out alternative causes). The New Jersey Supreme Court – perhaps the most liberal court in the country on expert causation opinions – has held:
[A]fter the expert rules in plausible causes, the expert then must rule out those causes that did not produce the patient’s condition by engaging in a process of elimination, eliminating hypotheses on the basis of a continuing examination of the evidence so as to reach a conclusion as to the most likely cause of the findings in that particular case . . . . In rejecting the alternative hypotheses, the expert must use scientific methods and procedures and justify an elimination on more than subjective beliefs or unsupported speculation.
Creanga v. Jardal, 886 A.2d 633, 639-40 (N.J. 2005) (citations and quotation marks omitted). Thus, “if the expert completely fails to consider a cause that could explain the patient’s symptoms, the differential diagnosis is not reliable.” Carlson v. Okerstrom, 675 N.W.2d 89, 105 (Neb. 2004). See also Christian v. Gray, 65 P.3d 591, 604-05 (Okla. 2003) (“[a] reliable differential diagnosis. . .generally is accomplished by determining the possible causes for the patient’s symptoms and then eliminating each of these potential causes until reaching one that cannot be ruled out”); Cooper v. Smith & Nephew, Inc., 259 F.3d 194, 202 (4th Cir. 2001) (“if an expert utterly fails to consider alternative causes. . .a district court is justified in excluding the expert’s testimony”); Cavallo v. Star Enterprise, 892 F. Supp. 756, 771 (E.D. Va. 1995) (where “other possible causes of an injury cannot be ruled out, or at least the probability of their contribution minimized, then the ‘more likely than not’ threshold for proving causation may not be met”), aff’d in pertinent part, rev’d on other grounds, 100 F.3d 1150 (4th Cir. 1996).
Finally, there’s a most revealing aside at the very end of Best:
We further note that, even without [the expert’s] testimony, summary judgment might be inappropriate in this case in light of this court’s recent decision in Gass v. Marriott Hotel Services, 558 F.3d 419, 434 (6th Cir. 2009) (holding that expert testimony was not required to prove the causation element of a negligence case where the plaintiffs were allegedly exposed to pesticides and immediately developed respiratory injuries). Because we conclude that [the] opinion is admissible, however, we have no need to decide whether the holding in Gass is applicable to the present case.
Best, 2009 WL 1010883, at *11 (emphasis added). This more or less confirms what we thought about the opinion all along – given the temporal relationship between the purported injuries and the exposure, nothing else really mattered to the court.
If that’s true, then it puts the Sixth Circuit at odds with a great deal of Daubert law holding that timing isn’t everything. “[T]emporal connection between exposure to chemicals and an onset of symptoms, standing alone, is entitled to little weight in determining causation.” Bland, 538 F.3d at 898-99 (quoting Moore, 151 F.3d at 278); Ervin v. Johnson & Johnson, Inc., 492 F.3d 901, 904-05 (7th Cir. 2007) (“mere existence of a temporal relationship between taking a medication and the onset of symptoms does not show a sufficient causal relationship”); McClain, 401 F.3d at 1243 (“[t]he post hoc ergo propter hoc fallacy assumes causality from temporal sequence”); Meister v. Medical Engineering Corp., 267 F.3d 1123, 1129 (D.C. Cir. 2001) (the expert “failed to show any nexus between [plaintiff’s] symptoms and [the product]; the mere simultaneous existence of the two clearly is not an appropriate methodology”); Glastettner, 252 F.3d at 990 (plaintiff “demonstrate[d] a temporal association” but “that association is not scientifically valid proof of causation”); Terran v. Secretary HHS, 195 F.3d 1302, 1317 (Fed. Cir. 1999) (plaintiff “points to the temporal association. . ., which has been held to be insufficient proof of causation”); Allison v. McGhan Medical Corp., 184 F.3d 1300, 1321 (11th Cir. 1999) (“mere coincidence of temporality” held a “questionable” basis for an expert opinion); Kennedy v. Collagen Corp., 161 F.3d 1226, 1227 (9th Cir. 1997) (admissible causation opinion must be based upon “other factors in addition to the temporal relationship”); Hodges v. Secretary HHS, 9 F.3d 958, 960 (Fed. Cir. 1993) (“a proximate temporal association alone does not suffice to show a causal link between the vaccination and the injury”); Sakaria v. Trans World Airlines, 8 F.3d 164, 172 (4th Cir. 1993) (“temporal sequence of the two events” without medical causation testimony is insufficient to suggest probability of causation). And these are just the appellate decisions in the federal courts – we finally got tired of researching.
In sum, while we agree with the Sixth Circuit in Best that, when done right, differential diagnosis can be a grounds for a causation opinion under Daubert, we think that the district court was within its discretion to conclude that this particular differential diagnosis was inadequate – both on “rule in” and “rule out” grounds. The key to Best is the standard of review. We don’t think the Sixth Circuit should have employed de novo review to reverse what was a decision that the Supreme Court has held is within the discretionary power of the District Court. Kumho Tire Co. v. Carmichael, 526 U.S. 137, 152 (1999) (adopting abuse of discretion review in Daubert cases).