Guides · Any expert

Differential Diagnosis as a Causation Method: What Courts Require

Differential diagnosis is how physicians identify a disease: list the possible causes of the patient's findings, then eliminate them until one remains. In litigation the same reasoning, often called differential etiology, is used to prove what caused an injury. Courts accept it as a method in principle, and many causation opinions in malpractice, toxic tort and injury cases rest on it. But the label does not carry the opinion. Courts look at both halves of the method, ruling in and ruling out, and exclude opinions where either half is missing or rests on assumption. This guide covers what a reliable differential requires and the recurring ways it fails.

Guide 14 of 14 · November 17, 2026 · 6 min read
On this page
  1. Two steps
  2. Ruling in
  3. Ruling out
  4. Untestable theories
  5. Defense alternatives
  6. What to prepare

The method in two steps

A differential has two steps. First, the expert rules in: identifies the causes capable of producing the injury. A cause cannot be ruled in unless there is a reliable basis to say it can produce this kind of injury at all, which is a general causation question. Second, the expert rules out: eliminates the alternatives until the cause at issue is the most likely one. Both steps must rest on the facts of the case, not on assumptions the record does not support.

When it is done properly, courts treat the process as reliable and leave disagreements about the result to the jury. In Cockayne v. Bristol Hospital (Conn. App. Ct. 2022), a patient's rectum was found perforated the day after nurses administered three enemas. The defense argued the plaintiffs had no evidence of when the perforation occurred or whether the enema could cause it. The Appellate Court upheld the verdict: there was no perforation before the hospitalization, the most likely cause was a foreign object, and of four possible causation events an expert used differential diagnosis to eliminate three and opined that an enema during the hospitalization caused the injury. The court noted that experts need not disprove every other possible explanation, only show that their opinion rests on reasonable probability, and that a physician's deduction by eliminating other causes can establish causation.

Ruling in requires a reliable basis

The most common failure is ruling in a cause that has not been shown capable of producing the injury. A differential cannot create general causation; it presupposes it.

Scottoline v. Women First (Del. 2025) is a recent example from the birth injury setting. A newborn suffered hypoxic-ischemic encephalopathy, and the plaintiffs' pediatric neurologist opined that it caused the child's later neurological and neurodevelopmental disabilities, including a behavioral syndrome within the autism spectrum. Autism spectrum disorder has no known cause. The trial court excluded the opinion under Rule 702 and the Delaware Supreme Court affirmed, holding that the opinion rested on the expert's say-so rather than a reliable scientific basis connecting the birth injury to those conditions. One justice dissented.

The same failure appears when the theory itself lacks support. In Swanigan v. Avenues Healthcare (Utah Ct. App. 2023), a nursing home resident developed foot infections and later died of a heart attack. The plaintiffs' physician opined that the infection caused inflammation, which caused the clot. He acknowledged that the idea was not something generally promulgated by the medical community and that the literature on infection causing thrombotic events was thin. The Utah Court of Appeals upheld exclusion: the theory was untested, not generally accepted, and not shown to rest on sufficient facts or data.

Ruling out requires engagement with the alternatives

The second step fails when the expert does not actually eliminate the alternatives. An expert need not rule out every conceivable cause, but must at least address the obvious ones. Scottoline made the point directly: although a differential need not eliminate all other causes, it must exclude obvious alternatives, and a later report that tried to rule them out could not rescue an opinion already found unreliable.

The library's specific causation cases show the same pattern in toxic tort. In C.W. ex rel. Wood v. Textron (7th Cir. 2015), the Seventh Circuit affirmed exclusion of physicians' opinions that vinyl chloride in well water caused children's illnesses, in part because one ruled out other causes by relying on what other doctors had concluded rather than doing the work, and both lacked a reliable basis to rule in the chemical. In Lancaster v. BNSF Railway (8th Cir. 2023), a physician's differential etiology for a railroad worker's lung cancer was excluded because it assumed exposures to asbestos and diesel fumes that the record did not support. A differential is only as good as the facts fed into it.

Untestable mechanisms are not automatically unreliable

Defendants often argue that a causal mechanism that cannot be tested cannot be reliable. Courts do not always agree, particularly where testing would be impossible or unethical.

In Trujillo v. Vail Clinic (Colo. App. 2020), the plaintiffs' experts testified that cranial compression during labor caused a newborn's ischemic brain injury, a theory known as cranial compression ischemic encephalopathy. The trial court excluded the testimony as unreliable and granted summary judgment. The Court of Appeals reversed. Following Colorado precedent that a theory impossible or unethical to test is not unreliable for that reason alone, it held the theory reasonably reliable on the totality of the circumstances, noting that the trial court had exceeded its gatekeeping role. It then held that the expert's application to the child, through a review of the records and a differential diagnosis, was also reasonably reliable, and that cross-examination and contrary evidence would test it at trial.

The contrast with Swanigan is instructive. In Trujillo the experts pointed to a body of literature and to other courts' treatment of the theory; in Swanigan the expert conceded his theory lacked acceptance and offered little support beyond his own belief.

Alternative causes offered by the defense

The defense can use the same method. A defense expert who proposes alternative causes is not required to prove one of them; the purpose is often to show that the plaintiff's differential failed to rule them out. Courts generally admit such testimony and leave disputes over its factual basis to cross-examination. In Damgaard v. Avera Health (D. Minn. 2015), a case alleging that negligent labor and newborn care caused hypoxic-ischemic encephalopathy and cerebral palsy, the court denied the plaintiff's motion to exclude defense experts who offered alternative causes, including prenatal exposure to a medication. The plaintiff disputed the evidence of that exposure, but the court held that the factual basis of an opinion goes to its weight, and that published studies are not required in every case for a medical expert to reach a reliable causation opinion.

What to prepare

  • Before the differential, establish general causation: a reliable basis that each ruled-in cause can produce this injury.
  • List the obvious alternative causes for this patient, and explain how each was eliminated, with reference to the record.
  • Do the elimination yourself; do not rely on other physicians' conclusions without analysis.
  • Make sure every exposure, event or fact the differential assumes is supported by evidence in the record.
  • Where the mechanism cannot be tested, assemble the other indicia of reliability: literature, other courts' treatment, clinical acceptance.
  • Be cautious about linking an injury to a condition with no known cause.
  • When defending, identify the alternatives the plaintiff's expert did not address, and the assumptions the record does not support.

Supreme Mind builds Expert Witness AI Agents for litigation, for any expert on any matter, from case intake to settlement. A lawyer points an AI Agent at a matter and picks a class of expert, and one run returns a brief in four sections: the expert's likely opinion, ranked cross-examination weaknesses, the Daubert / Rule 702 attack surface, and a directional settlement range. Experts are de-identified composites of a class, never a named individual; supporting quotations are checked verbatim against the public record, authorities carry their check result, and nothing it produces is filed or testifies. A complete sample brief is free, and the method explains how it works.

Sources

This guide is general information, not legal advice.

How each court of appeals applies amended Rule 702: see the Rule 702 tracker.

Run this analysis against the expert on your own matter.Start freeRead a sample brief →

The series

Related: the expert agent library, a complete sample brief, and how it works.

GatekeeperExpert AgentsReport BuilderOne record, from intake to settlement
Book a demo