What is the difference between general causation and specific causation in a toxic tort case?
General causation asks whether a substance is capable of causing a disease in human beings at exposures of the kind alleged; specific causation asks whether it caused the disease of the particular claimant in front of the court. The Eleventh Circuit stated the requirement in The Estate of Carol Davis v. Lockheed Martin Corp., Nos. 24-10080 and 24-10416 (11th Cir. 28 July 2026), a case about volatile organic compounds at Lockheed Martin’s Sand Lake Road facility in Orlando and the neurodegenerative diseases the plaintiffs attributed to them: “Plaintiffs in these actions must establish both general and specific causation.” The court affirmed the exclusion of two experts and the summary judgment that followed, writing that “the district court’s orders excluding Kantor and Kendall and granting summary judgment to Lockheed based on those exclusions are AFFIRMED.”
The order between them is not arbitrary. The Reference Manual on Scientific Evidence, Fourth Edition — the Federal Judicial Center and National Academies volume published on 31 December 2025 — states the dependency in its Reference Guide on Epidemiology, quoting the epidemiological literature: “[A]n agent cannot be considered to cause the illness of a specific person unless it is recognized as a cause of that disease in general.” That is why general causation is usually briefed first, and why an exclusion there ends a matter without any inquiry into the individual record. In In re Paraquat Products Liability Litigation, No. 3:21-md-3004-NJR (S.D. Ill. 17 April 2024), Chief Judge Nancy Rosenstengel dismissed four trial selection cases the same day she excluded the plaintiffs’ sole general causation expert, because, in her words, “Dr. Wells’ exclusion mandates the dismissal of the four trial selection cases.”
Two allocations of decision-making sit on top of this and are worth separating. Whether an opinion may be heard at all is an admissibility question decided by the trial judge under Federal Rule of Evidence 702, reviewed on appeal for abuse of discretion — the holding of General Electric Co. v. Joiner, 522 U.S. 136 (1997), decided 15 December 1997: “abuse of discretion is the proper standard by which to review a district court’s decision to admit or exclude scientific evidence.” What causation requires as an element of the claim is supplied by the substantive law of the governing jurisdiction, and it is not uniform across the states. Nothing on this page says what any particular forum requires, or what any expert or claimant should do. The Institute’s orientation on the structure of these showings sits at Causation & Health Effects.
What evidence does general causation require, and does it have to be epidemiology?
General causation is argued from a body of literature rather than a single study, and the disciplines that produce it are distinct. The Reference Manual on Scientific Evidence, Fourth Edition, carries separate reference guides for exposure science, epidemiology and toxicology. Its Reference Guide on Exposure Science and Exposure Assessment describes exposure science as the discipline that studies “individuals and populations and their behaviors related to contact with stressors, the nature and extent of such contact, and the fate of these stressors in the environment and in organisms over space and time,” and says that it “has a complementary role with epidemiology and toxicology, two disciplines devoted to understanding the inherent potential hazard of a given stressor.” The Reference Guide on Toxicology gives the first of that field’s “three central tenets” as “the dose makes the poison”: “This implies that all chemical agents are intrinsically hazardous; whether they cause harm is only a question of dose.”
Two appellate decisions show courts examining how a body of literature was weighed rather than demanding a particular study type. In In re Onglyza (Saxagliptin) and Kombiglyze (Saxagliptin and Metformin) Products Liability Litigation, 93 F.4th 339, No. 22-6078 (6th Cir. 13 February 2024), the Sixth Circuit affirmed exclusion of a general causation expert who had applied the Bradford Hill criteria, on a record in which the expert “cherry-picked data to bolster his case” and in which, quoting In re Lipitor (Atorvastatin Calcium) Marketing, Sales Practices & Products Liability Litigation, 174 F. Supp. 3d 911, 932 (D.S.C. 2016), a failure “to adequately account for contrary evidence is not reliable or scientifically sound.” The Eleventh Circuit in In re Deepwater Horizon BELO Cases (18 October 2024) took the Hill guidelines seriously without treating them as a scoring sheet: “No factor is dispositive. And ‘[n]o algorithm exists for applying the Hill guidelines.’” The inner quotation there comes from the Restatement (Third) of Torts: Liability for Physical and Emotional Harm § 28 cmt. c(3).
Whether human epidemiology is indispensable is a genuine point of division and depends on the forum and on the state of the science for the substance and disease at issue; it is a question for the court applying its own law, not a matter this Institute resolves. What travels across forums is the demand that the analysis engage the studies that cut against it. The Institute’s longer treatment of how that body of evidence is assembled is at General causation & epidemiology, and the interaction with admissibility practice is at Daubert & expert admissibility.
Does a relative risk above 2.0 prove that a chemical caused someone’s disease?
No, and the source most often cited for the proposition says so itself. The Reference Manual on Scientific Evidence, Fourth Edition, describes the reasoning as a judicial one: “Some courts have reasoned that when epidemiologic studies find that exposure to the agent causes an incidence in the exposed group more than twice the incidence in the unexposed group (i.e., a relative risk greater than 2.0), the probability that exposure to the agent caused a similarly situated individual’s disease is greater than 50%.” Those courts, it continues, hold that such evidence “is sufficient to satisfy the plaintiff’s burden of production and permit submission of specific causation to a jury.” Burden of production and submission to a jury are legal sufficiency concepts. They are not a finding that the substance caused anyone’s illness.
The leading American statement of the rule is a state one and carries its own limit. In Merrell Dow Pharmaceuticals, Inc. v. Havner, 953 S.W.2d 706 (Tex. 9 July 1997), No. 95-1036, a Bendectin case, the Supreme Court of Texas surveyed the authorities and recorded that “[o]ther courts have likewise found that the requirement of a more than 50% probability means that epidemiological evidence must show that the risk of an injury or condition in the exposed population was more than double the risk in the unexposed or control population.” It then stated its own limit: “We do not hold, however, that a relative risk of more than 2.0 is a litmus test or that a single epidemiological test is legally sufficient evidence of causation.” Havner also requires the claimant to be like the people in the study: “A claimant must show that he or she is similar to those in the studies. This would include proof that the injured person was exposed to the same substance, that the exposure or dose levels were comparable to or greater than those in the studies, that the exposure occurred before the onset of injury, and that the timing of the onset of injury was consistent with that experienced by those in the study.” That is Texas law, and other jurisdictions treat the threshold differently; which rule governs is for the court applying its own precedent.
The Reference Manual attaches conditions to the doubling argument that are frequently dropped when it is repeated. The first is a valid study: “The propriety of this ‘doubling’ reasoning depends on group studies identifying a genuine causal relationship and a reasonably reliable measure of the increased risk,” which requires attention to random error, bias and confounding. The second is similarity: “Only if the study participants and the plaintiff are similar with respect to relevant risk factors will a risk estimate from a study or studies be valid when applied to an individual.” Its illustration is unforgiving — if the exposed subjects in a smoking study “have smoked half a pack of cigarettes a day for 20 years, the degree of increased incidence of lung cancer among them cannot be extrapolated to someone who smoked two packs of cigarettes for 30 years, without strong assumptions about the dose–response relationship.” A number lifted out of a study without those conditions has left its own premises behind.
How much does dose matter, and what happens when the exposure record is thin?
Dose is the hinge, and the Reference Manual on Scientific Evidence, Fourth Edition, opens its Reference Guide on Exposure Science and Exposure Assessment with “the toxicological concept attributed to Paracelsus that ‘the dose makes the poison.’” The same guide separates the components that make up a dose, describing exposure “in terms of intensity or magnitude (how much), frequency (how often) and duration (how long) of contact at an external boundary.” A substance identified, a receptor identified and a diagnosis identified still leave the quantitative question open, and that question is answered by exposure assessment and dose reconstruction rather than by the medical record. The Institute’s material on that workstream is at Exposure assessment & dose reconstruction.
Circuits differ in how explicitly they convert dose into an element of the expert showing. In In re Deepwater Horizon BELO Cases (11th Cir. 18 October 2024), two cleanup workers attributed chronic sinusitis to crude oil and dispersants, and the Eleventh Circuit affirmed the exclusion of Dr. Michael Freeman and Dr. Gina Solomon, repeating that a toxic-tort plaintiff “must demonstrate the levels of exposure” to the alleged toxin “that are hazardous to human beings generally,” quoting McClain v. Metabolife International, Inc., 401 F.3d 1233, 1241 (11th Cir. 2005). It reviewed the ruling under the Joiner standard — “We review the exclusion of expert testimony for abuse of discretion” — and closed: “We AFFIRM the judgment in favor of BP.” The same court repeated the requirement in 2026 in The Estate of Carol Davis v. Lockheed Martin Corp.: “A plaintiff must demonstrate the level of exposure to the allegedly harmful chemical that is hazardous to a human being.”
New York does not require the same arithmetic, which is why the jurisdiction has to be named before the rule is stated. In Parker v. Mobil Oil Corp., 7 N.Y.3d 434 (N.Y. 17 October 2006), a benzene and acute myelogenous leukemia case brought by a former gasoline station attendant, Judge Ciparick wrote for the Court of Appeals that “it is not always necessary for a plaintiff to quantify exposure levels precisely or use the dose-response relationship, provided that whatever methods an expert uses to establish causation are generally accepted in the scientific community.” New York applies the general acceptance standard of Frye rather than the federal rule, and the court quoted it. The order of the Appellate Division was nonetheless affirmed, because the experts there had not quantified the exposure by any accepted method either.
What have courts said about differential etiology?
Two federal courts of appeals mark the range — one accepting differential etiology as a method, the other excluding a particular application of it. The Fourth Circuit stated the acceptance in Westberry v. Gislaved Gummi AB, 178 F.3d 257 (4th Cir. 20 May 1999), Nos. 98-1540 and 98-1587, a talc exposure case: “Differential diagnosis, or differential etiology, is a standard scientific technique of identifying the cause of a medical problem by eliminating the likely causes until the most probable one is isolated.” The same opinion fixed the outer limit of what the method must accomplish, quoting Heller v. Shaw Industries, Inc., 167 F.3d 146, 156 (3d Cir. 1999): “[A] medical expert’s causation conclusion should not be excluded because he or she has failed to rule out every possible alternative cause of a plaintiff’s illness.”
The Sixth Circuit supplied the counterweight in Tamraz v. Lincoln Electric Co., 620 F.3d 665 (6th Cir. 8 September 2010), Nos. 08-4015 and 08-4016, where a welder attributed parkinsonism to manganese in welding fumes. Writing for the panel, Judge Sutton held that “[s]imply claiming that an expert used the ‘differential diagnosis’ method is not some incantation that opens the Daubert gate,” and that the expert’s “efforts to ‘rule in’ manganese exposure as a possible cause or to ‘rule out’ other possible causes turned on speculation, not a valid methodology.” The court reversed. The distinction the two decisions draw together is between an expert who addressed the alternatives on the individual record and one who named them and moved on.
The Reference Manual on Scientific Evidence, Fourth Edition, endorses the logic while insisting on the vocabulary: “The term differential diagnosis in a clinical context refers to identifying the patient’s medical condition, whereas differential etiology refers to identifying the causal factors involved in an individual’s condition.” It explains why the logic works — “Eliminating other known and competing causes increases the probability that a given individual’s disease was caused by exposure to the agent” — and describes the sequence, in which an expert “first determines other known causes of the disease in question and then attempts to ascertain whether those competing causes can be ruled out as a cause of plaintiff’s disease.” Whether a given application met that standard is a question for the court that has the record. The Institute’s treatment of the individual showing is at Specific causation & differential etiology.
Why is a plausible mechanism not the same as proof in a particular person?
Because plausibility is a hypothesis about what could happen and Rule 702 asks for knowledge about what did. Tamraz v. Lincoln Electric Co. (6th Cir. 2010) put the gap in three sentences: “That is a plausible hypothesis. It may even be right. But it is no more than a hypothesis, and it thus is not ‘knowledge.’” The opinion grounded that in Rule 702’s “knowledge” requirement as Daubert construed it, noting that the “‘knowledge’ requirement of Rule 702 requires ‘more than subjective belief or unsupported speculation,’” quoting Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 590 (1993). A mechanism explains how an effect could occur; it does not establish that it occurred at this dose, in this person, on this timeline.
The Supreme Court described the same distance in General Electric Co. v. Joiner, 522 U.S. 136 (1997), a case in which an electrician attributed lung cancer to polychlorinated biphenyls. Rejecting the argument that a court may examine methodology but not conclusions, the Court wrote that “conclusions and methodology are not entirely distinct from one another,” that “[t]rained experts commonly extrapolate from existing data,” and that “nothing in either Daubert or the Federal Rules of Evidence requires a district court to admit opinion evidence which is connected to existing data only by the ipse dixit of the expert.” A court, it held, “may conclude that there is simply too great an analytical gap between the data and the opinion proffered.” Mechanistic reasoning is one of the places that gap tends to open.
Two practical consequences follow, and neither is a prediction about any case. The first is that mechanism is corroboration inside a weight-of-evidence argument rather than a substitute for the exposure figure and the individual differential. The second is that the same record can support an admissible general causation opinion and an inadmissible specific causation one, because they are separate showings judged separately. Nothing on this page states that any substance caused any person’s disease, or could not have; that is a question for a credentialed expert working from a complete record, and any question about an individual’s own health belongs with that person’s physician. What a loss is worth, if one is established, is a different discipline again, and belongs to our Economic Damages Institute.