Causation,
determined by science.
Few questions in medicine carry as much consequence: did this event or exposure actually cause this condition? This resource answers it the way it deserves — with evidence, sound method, and integrity, not assumption or opinion.
Few questions carry as much consequence for an individual, an employer, an insurer, and society. The answer determines who pays, who is held responsible, what care is delivered, and whether justice is served. A question of that weight deserves a disciplined, transparent, reproducible method.
A scientifically sound analysis is reproducible: a second qualified evaluator, reviewing the same records and the same science, should reach the same conclusion. That is the standard this resource exists to advance.
Evidence first: fact, not opinion.
This is the essential distinction. Science does not ask who believes something; it asks whether the finding can be observed, verified, and reproduced. A reliable causation conclusion should rest on evidence and method, not on the confidence or credentials of the person offering it. The method is the discipline placed around clinical judgment so the conclusion is anchored in facts rather than authority.
A person of science uses a standardized, objective, fact-based approach. The question is not who said it, but whether it can be verified. If a finding is supported by the record, reproducible, and methodologically sound, the identity of the person who introduced it should not control its value.
An opinion-driven expert relies primarily on personal experience, reputation, and persuasion rather than a reproducible method. In court and administrative systems, expert opinion can sometimes be treated as more important than independently verifiable fact. When that happens, confidence, credentials, convenience, or presentation may be mistaken for accuracy.
Core principles.
More likely than not.
In many workers' compensation and civil tort contexts, medical causation opinions are generally expressed to a reasonable degree of medical probability. The evaluator must distinguish what is probable from what is merely possible. A factor that is biologically possible, or weakly associated, does not meet the threshold simply because it cannot be excluded.
At relative risk (RR) 2.0, the attributable fraction reaches 50%; above RR 2.0, it exceeds the more-likely-than-not threshold. Move the slider to see how the attributable fraction changes.
Because it is structured, transparent, and reproducible, the method is built to withstand Daubert scrutiny — the legal tests for admissible scientific testimony.
See the legal standards →Four common patterns that distort causation opinions.
See common pitfalls →{{ t.title }}
{{ t.desc }}
A structured method.
AMA six-steps, in sequence.
Developed by NIOSH, adapted by ACOEM, and incorporated into the AMA Guides, the steps are sequential. The analysis proceeds only as far as the evidence allows; if a required step fails, the causal conclusion cannot be supported.
{{ s.title }}
{{ s.desc }}
Nine considerations — weighed, never tallied.
Bradford Hill's nine viewpoints help assess whether an association may be causal. They are not a checklist or scoring system. Causation is assessed along a continuum, with each viewpoint weighed in context. In occupational analysis, temporality, strength of association, and dose-response often carry the greatest practical weight.
No causal hypothesis is proven with absolute certainty. Scientific analysis tests proposed explanations against the evidence and is often strongest when excluding unsupported explanations. Sometimes the most defensible conclusion is not a positive causal finding, but a disciplined exclusion: the evidence does not support the proposed explanation.
See all nine viewpoints →Same method. Three different conclusions.
A neutral method should not predictably favor one side or one outcome. When applied consistently across three different conditions, the same framework may support relatedness, fail to support it, or leave the question unresolved. That variation is not inconsistency — it is the expected result of a method that follows the evidence wherever it leads.
Same neutral method, three different conclusions. That contrast is the point. The method is neutral because it is applied the same way in each case — not because it produces the same result.
These examples are provided for general educational and illustrative purposes only. They are not intended to describe, evaluate, or determine causation in any specific individual case. Any case-specific causation analysis must be based on the particular diagnosis, exposure history, medical record, scientific literature, and applicable legal or administrative standard.
Built for everyone who relies on the answer.
Integrity means applying the method consistently and reporting the results without regard to the requesting party. The science is the science regardless of who asks for the analysis; the conclusion must follow the evidence, not the interests of the person or party requesting the review.
This method advocates for no party and favors no predetermined outcome. Its commitment is to method alone: a structured, evidence-based process applied the same way regardless of who performs the analysis, who requested it, or which party the conclusion may favor. The science is the science. The conclusion must follow the evidence, not the interests of the requesting party.
Adapted from The Science of Causation: Principles and Best Practices for Determining Injury and Disease Relatedness and The Science of Causation: An Introduction to Biostatistics — foundational educational source documents for OpenCausation.orgSM — by J. Mark Melhorn, MD.
Causation analysis, grounded in evidence.
OpenCausation.org provides free access to the principles, method, and science behind credible causation analysis — structured, transparent, and built to be trusted.