Peter Neufeld is a co-founder of the Innocence Project. Keith Findley is a professor at the University of Wisconsin Law School. Dean Strang is a criminal defense lawyer and law professor at Loyola University Chicago. Findley and Strang are also co-founders of the Center for Integrity in Forensic Sciences.

A new study from the University of Washington concludes that, over nearly 40 years, medical examiners and coroners undercounted killings by U.S. police by more than half. During that time, these officials missed or covered up more than 17,000 police killings between 1980 and 2018.

The study — published in the Lancet, a respected medical journal — raises important questions about the objectivity and reliability of physicians who work as government medical examiners and coroners. The findings add to an accumulating list of grave concerns about racial bias in forensic pathology and policing, adding evidence that police kill Black people at a rate 3.5 times greater than White people, and that medical officials underreported Black deaths at a higher rate than White deaths.

Although many forensic pathologists carefully follow scientific principles, the frequency of police killings overlooked or incorrectly diagnosed means the problem is systemic. Yet, because most medical examiners are specialized doctors (although some coroners are not), Americans may assume that they are the gold standard in forensic evidence.

One systemic feature that appears to produce error, however, is that medical examiners long have been close allies of police and prosecutors — frequently partisans, not neutrals. Many allow the police, but not others, to observe autopsies and to influence critical steps in death investigation. They often talk freely to prosecutors, but only grudgingly — if at all — to defense lawyers. In a 2011 survey, 22 percent of medical examiners and coroners reported pressure from government officials to change the cause or manner of death on a certificate.

Medical examiners claim, in essence, the dubious right to diagnose (or rule out) not a medical condition, but a crime. Worse, many react fiercely to questions about their objectivity, biases or ethics. They deny — irrationally and unscientifically — that they are subject to universal human cognitive biases at all.

Like other physicians, medical examiners often rely appropriately on patient medical history to assist in diagnosis. But they also routinely use nonmedical background information that can produce serious contextual bias. They often accept police suspicions, prior criminal records, police theories of criminal liability, tropes about drug or alcohol use, police concerns about their own culpability, and race in assessing cause and manner of death. The absence of rigorous science leads to false negatives — such as when the pathologist ignores the role of excessive police restraints — and to false positives, such as charging a civilian with homicide when the death was accidental (or natural).

Another recent study in the Journal of Forensic Sciences found that, in actual case work, pathologists who examined 10 years of children’s death certificates in Nevada were much more likely to declare a child’s death a homicide rather than an accident when the child was Black. Corroborating these findings, the researchers presented pathologists with a hypothetical postmortem examination of a young child. The pathologists were many times more likely to find the child’s death to be a homicide than an accident, based on identical medical evidence, if told that the child was Black and the caretaker was the mother’s boyfriend than when told that the child was White and the caretaker was the child’s grandmother.

Responding to the study, vocal forensic pathologists sought to bury the research and the researchers. They first demanded that the journal retract the peer-reviewed article, then asked the journal’s publisher to censor it, and finally filed ethics complaints against the forensic pathologists and neuroscientist who co-authored the article. The efforts failed, but the effect was chilling.

This reaction was telling. In a letter lashing out against the study, 74 pathologists wrote, “Manner determination [e.g., homicide, suicide, accident, natural or undetermined] is not a ‘scientific’ determination. It is a cultural determination that places a death in a social context for the purpose of public health statistics.” They elaborated that “there is no ‘right’ answer in many manner determinations” and “the goal is consistency rather than some nonexistent criteria for correctness.”

Such subjective determinations are especially vulnerable to cognitive and, in particular, racial biases. The shocking likelihood that medical examiners and coroners have missed (whether deliberately or through unexamined biases) more than half of police killings over almost four decades is a sharp reminder of these tendencies and their dangers.

Medical examiners and other doctors have an important role to play in making and explaining objective physical findings that laypeople would miss or misunderstand. When constrained by scientific rigor and objectivity, their courtroom contributions are valuable. But as a group, their eagerness to embrace contextual bias that distorts reliability and their willingness to act as partisans aiding the prosecution, raise real concerns. We now can point to about 17,000 more people who would second these concerns, if they could.