The legal standards of causation.
Burdens of proof, the relative-risk 2.0 threshold, and the admissibility of expert testimony.
The science of causation produces probabilities; the law requires decisions. This section explains the legal standards that translate one into the other — the burden of proof, the relative-risk threshold courts use to define causation, and the rules that govern whether an expert's causation opinion may even be heard. It is written for physicians, attorneys, and claims professionals, and favors no party.
Two worlds, one question.
Science and the law approach causation differently, and understanding the gap between them prevents most errors at the boundary. Science describes populations and speaks in probabilities; it can rarely say that a given exposure caused a given individual's disease with certainty. The law, by contrast, must reach a decision — yes or no, compensable or not — and it has built standards to convert scientific probability into a usable legal answer.
A causation analysis that is carefully reasoned in scientific terms can still be rejected by a court, and a legally sufficient opinion can rest on thin science. The remedy is not to choose one world over the other but to understand both: to perform the science rigorously and then to express it in the legal terms the forum requires. The standards below are the vocabulary of that translation.
“More likely than not.”
In most workers' compensation, civil tort, and other legal causation contexts, the claimant must prove causation by a standard usually summarized as “more likely than not” — greater than 50% probability. The precise statutory wording varies, and the subtle differences can matter, but the general concept is consistent across the common formulations:
| Standard | Abbrev. | Meaning |
|---|---|---|
| More probable than not | MPTN | The causal link is more likely than not — a probability greater than 50%. |
| Preponderance of the evidence | POTE | The greater weight of the evidence favors causation; functionally “more likely than not.” |
| Reasonable degree of medical probability | RDOMP | The medical opinion is held to a probability standard — more likely than not. |
| Reasonable degree of medical certainty | RDOMC | A stricter-sounding phrase, but generally interpreted as at least a preponderance — not absolute certainty. |
The four common threshold formulations. Different words; the same underlying “more likely than not” concept.
The phrase “reasonable degree of medical certainty” often misleads. It does not require certainty in the ordinary sense; courts generally interpret it to mean at least a preponderance of the evidence — more likely than not — and explicitly not absolute or unequivocal proof. Whatever the wording, the evaluator's task is the same: to distinguish what is probable from what is merely possible, and to opine only to the level the evidence supports.
Variations that change the answer.
Beyond the common threshold, jurisdictions add requirements that can change the outcome for identical facts. A competent analysis identifies the controlling standard before forming a conclusion.
Prevailing factor and major contributing cause
A number of states require that the work exposure be more than merely a contributing cause — it must be the primary or prevailing factor. Missouri's law defines the “prevailing factor” as the primary factor, in relation to any other factor, causing both the resulting medical condition and the disability. Kansas adopted a similar prevailing-factor test in 2011, and Florida requires the work injury to be the major contributing cause (greater than 50% contributory) with significant objective findings. Oklahoma's reform tied its “major cause” requirement directly to Daubert-based proof. These standards demand apportionment-grade analysis, not a simple yes or no.
Aggravation versus exacerbation
The distinction is consequential and jurisdiction-specific. In Texas, for example, an aggravation that permanently worsens an underlying condition may be a compensable, work-related injury, whereas an exacerbation — a temporary worsening of symptoms such as pain that does not change the underlying condition — may support temporary treatment but, because there is no permanent change, no permanent indemnity. Using these terms precisely is part of a credible opinion.
Any-contribution standards and claimant presumptions
Some systems apply a lower bar, under which any work contribution may be compensable, typically with apportionment. Others build in presumptions favoring the claimant: under the Longshore and Harbor Workers' Compensation Act, if the evidence is balanced, the presumption favors the claimant; Hawaii presumes compensability unless the employer rebuts it by substantial evidence. None of this changes the science — it changes the legal threshold the science must clear.
The constant: standards are in flux
Legal thresholds for causation change frequently through legislation and case law. The same state may use one phrase one year and another the next, and a newly enacted standard often awaits clarification through appellate decisions. An evaluator should confirm the current controlling standard rather than rely on memory or precedent from another matter.
Where epidemiology meets the 50% line.
The clearest bridge between epidemiology and the more-likely-than-not standard is the relative-risk doubling rule. Relative risk compares the rate of disease in an exposed group with the rate in an appropriate unexposed or reference group. At relative risk (RR) 2.0, the attributable fraction reaches 50%; above RR 2.0, it exceeds the more-likely-than-not threshold. This is why an RR >2.0 has been used by courts to establish “legal” causation.
The logic is the excess fraction, (RR − 1) / RR, which gives the probability that an exposed person's disease is attributable to the exposure:
| Relative risk | Excess fraction (RR−1)/RR | Meets “more likely than not”? |
|---|---|---|
| 1.1 | ≈ 9% | No |
| 1.5 | ≈ 33% | No |
| 2.0 | 50% (the pivot) | At the threshold |
| 2.5 | 60% | Yes |
| 3.0 | ≈ 67% | Yes |
| 4.0 | 75% | Yes |
Below RR 2.0, the exposure more likely than not did not cause the individual's disease; at and above 2.0, it more likely than not did.
A legal line, not a scientific one. The RR ≥ 2.0 threshold was selected from common law, not derived from epidemiology.
Epidemiologists read it cautiously. Many regard an RR below 3 as “weak” evidence, especially from case-control studies. RR 2.0 is a legal bright line, not a scientific certainty.
Below 2.0 can still matter. In any-contribution jurisdictions, a factor with RR < 2.0 may still receive a small apportionment — “minimal causation,” much less than 50%.
The case lineage behind the rule
The 2.0 threshold did not arise in a single decision; it accreted across decades of toxic-tort and product-liability litigation in which courts wrestled with how epidemiologic risk translates into individual causation. The AMA Guides Causation text catalogs the supporting cases (Table 4-1) across several domains:
| Litigation domain | Representative cases |
|---|---|
| Radiation | Johnston v. United States (D. Kan. 1984); In re Hanford Nuclear Reservation Litigation (E.D. Wash. 1998) |
| Swine flu – Guillain-Barré | Cook v. United States (N.D. Cal. 1982); Padgett v. United States (W.D. Tex. 1982); Manko v. United States (W.D. Mo. 1986) |
| IUD – pelvic inflammatory disease | Marder v. G.D. Searle & Co. (D. Md. 1986) |
| Bendectin | DeLuca v. Merrell Dow (3d Cir. 1990); Daubert v. Merrell Dow (9th Cir. 1995); Merrell Dow v. Havner (Tex. 1997) |
| Silicone breast implants | Hall v. Baxter Healthcare (D. Ore. 1996); Allison v. McGhan Medical (11th Cir. 1999); Norris v. Baxter Healthcare (10th Cir. 2005) |
| Asbestos | Landrigan v. Celotex Corp. (N.J. 1992); Caterinicchio v. Pittsburgh Corning (N.J. 1992) |
Selected cases from Table 4-1 of the AMA Guides to the Evaluation of Disease and Injury Causation. 2nd ed. Chicago, IL: American Medical Association; 2014. The lineage is illustrative of the doubling-of-risk concept, not an exhaustive or current statement of law in any jurisdiction.
Whether the opinion is even heard.
Before a court weighs a causation opinion, it must decide whether to admit it at all. Two traditions govern that gatekeeping decision, and which one applies depends on the forum.
In Frye v. United States, a federal appeals court held that expert evidence is admissible only if the principle or technique on which it rests has gained general acceptance in the relevant scientific community. Frye locates the decision not in the court but in the scientific field itself: the question is whether the method is generally accepted, not whether the judge finds it persuasive. A number of states still follow Frye or a Frye-derived standard.
In Daubert v. Merrell Dow Pharmaceuticals — itself a Bendectin birth-defect case — the U.S. Supreme Court held that, under Federal Rule of Evidence 702, the trial judge serves as a gatekeeper who must ensure that expert testimony both rests on a reliable scientific foundation and is relevant to the issue (the “fit” requirement). Daubert governs the federal courts and has been adopted by many states; the analysis is more demanding than Frye because the judge must engage with the methodology itself.
The Court offered an illustrative, non-exhaustive list of factors for assessing reliability:
Testability. Whether the theory or technique can be (and has been) tested.
Peer review and publication. Whether it has been subjected to peer review and publication.
Known or potential error rate. Whether the error rate of the technique is known and acceptable.
General acceptance. Whether it is generally accepted in the relevant scientific community — Frye's question, preserved as one factor among several.
Beyond the four listed factors, the Daubert courts added a pivotal consideration: whether the expert's research and methods were developed independently of the litigation, or instead generated specifically for the case at hand. Opinions built on reasoning created only for the lawsuit — never published, never independently reviewed — are viewed with particular skepticism. This is one reason a causation analysis grounded in the pre-existing, published scientific literature is far more durable than one improvised for the case.
Frye and Daubert, side by side.
| Frye (1923) | Daubert (1993) | |
|---|---|---|
| Core question | Is the method generally accepted in the field? | Is the method reliable and relevant (does it “fit”)? |
| Who decides | The scientific community, in effect | The trial judge, as gatekeeper |
| Key authority | Frye v. United States | FRE 702; Daubert v. Merrell Dow |
| Reliability factors | General acceptance only | Testability, peer review, error rate, general acceptance, + litigation independence |
| Where it applies | Some states | Federal courts and many states |
Many jurisdictions have moved from Frye to Daubert; some retain Frye. The applicable test is determined by the forum.
Two later Supreme Court decisions are commonly grouped with Daubert as the “Daubert trilogy”: General Electric Co. v. Joiner (review of admissibility decisions for abuse of discretion) and Kumho Tire Co. v. Carmichael (extending the gatekeeping function to all expert testimony, not only the strictly scientific). Together they reinforce that the methodology behind an opinion — not merely the credentials of the witness — determines whether it is heard.
Why a disciplined methodology survives Daubert.
The structured causation method set out in the base reference is, in effect, engineered to satisfy these admissibility standards. It is not a coincidence; a method built to be transparent and reproducible is also a method built to withstand gatekeeping.
Testable and tested. The six-step framework and the evidence-rating system are explicit procedures that can be examined, applied by others, and checked.
Peer-reviewed and published. The method derives from NIOSH and ACOEM work and from the published, peer-reviewed scientific literature — not from reasoning invented for a particular case.
Error rate acknowledged. The strength-of-evidence scheme states its limits openly, distinguishing strong, some, insufficient, conflicted, and not-studied evidence rather than projecting false certainty.
Generally accepted. The Bradford Hill viewpoints and the epidemiologic methods underlying the analysis are the accepted framework of the field.
Independent of the litigation. Because conclusions rest on pre-existing published evidence applied through a standard method, they are not vulnerable to the central Daubert objection that the science was manufactured for the case.
An opinion produced this way can be expressed in whatever legal threshold the forum requires — more likely than not, prevailing factor, major contributing cause — while remaining anchored to the same underlying science. That is the practical payoff of method: it is both better science and more durable law.
The vocabulary, precisely.
| Term | What it means here |
|---|---|
| Possibility vs. probability | Possible = the exposure could have contributed; probable = more likely than not (>50%). Only probability meets the standard. |
| Medical certainty vs. probability | “Reasonable degree of medical certainty” is generally interpreted as a preponderance — not literal certainty. |
| General vs. specific causation | General = the exposure can cause the condition (population science); specific = it did cause it in this person (the individual case). |
| Aggravation vs. exacerbation | Aggravation = a permanent worsening of the underlying condition; exacerbation = a temporary worsening of symptoms. Compensability differs by jurisdiction. |
| Prevailing / major contributing cause | Standards requiring the work exposure to be the primary cause (>50%), not merely a contributor. |
| Gatekeeping | The judge's threshold decision, under Frye or Daubert, on whether expert testimony is admissible at all. |
Melhorn JM, Talmage JB, Ackerman WE III, Hyman MH, eds. AMA Guides to the Evaluation of Disease and Injury Causation. 2nd ed. Chicago, IL: American Medical Association; 2014. (Chapter 2, Understanding Work-Relatedness; Chapter 4, Methodology, incl. Table 4-1.)
Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579 (1993); Fed. R. Evid. 702.
Frye v. United States, 293 F. 1013 (D.C. Cir. 1923).
General Electric Co. v. Joiner, 522 U.S. 136 (1997); Kumho Tire Co. v. Carmichael, 526 U.S. 137 (1999).
Independence and disclaimer. OpenCausation.orgSM is an independent, open, noncommercial educational project. It is not affiliated with, endorsed by, or a product of the American Medical Association, and it provides no legal advice. Legal standards vary by jurisdiction and change over time; the controlling law of the relevant forum governs any actual matter.