Autopsy Reports and the Confrontation Clause: A Presumption of Admissibility

by and

Vol. 99 No. 3 (2015) | Fixing Discovery | Download PDF Version of Article
Cartoon of an autopsy

In 2004, the Supreme Court, in Crawford v. Washington,1 restored the “original meaning” of the Sixth Amendment’s Confrontation Clause. The framers of that clause — which guarantees a criminal defendant the right “to be confronted with the witnesses against him” — meant to outlaw the old-world practice of condemning men through ghost accusers who couldn’t be cross-examined at trial.

Crawford firmed up the right in favor of criminal defendants but it raised as many questions as it answered. One of the most important is whether the Confrontation Clause applies with full rigor to autopsy reports offered for their truth. There are two views. On one side is the argument that autopsy reports are prepared by neutral pathologists — highly trained specialists who are effectively separate from law enforcement, working under a statutory duty to determine the cause of unusual deaths. Their reports can appear in prosecutions, but the vast majority do not. To require these impartial participants to testify imposes a massive, pointless burden on them and serves to bar or undermine just prosecutions because autopsy evidence is soon lost and often impossible to recreate.

On the other side is the argument that autopsy reports are a formal record, created sometimes at police behest, by state agents who practically function as an arm of law enforcement. Autopsies, far from being a reading on some machine, are the product of human skill and judgment. The defendant, as with any other formalized testimony, should be able to test for fraud or incompetence. Pathologists are “witnesses” against the accused.

The issue usually arises when an autopsy report is offered in evidence or testified to by a colleague who was not its author. If the report is “testimonial,” it cannot be admitted into evidence unless the author testifies (or did so previously, under cross-examination). If the examiner dies or retires or moves away, the answer to this question often determines whether the case goes on. We think autopsy reports can be nontestimonial — and presumptively are.

The State of Confrontation Clause Jurisprudence

Crawford (2004). Crawford is what is usually referred to as a “landmark” decision. That term once referred to a conspicuous object that guided wayfarers and ships at sea. For the intrepid adventurers at the bar, however, the more prominent theme since Crawford has been confusion. Supreme Court Justice Antonin Scalia wrote Crawford but a few years later pronounced Confrontation Clause jurisprudence “in a shambles.” In Crawford, the Court found that the Confrontation Clause “is a procedural rather than a substantive guarantee. It commands, not that evidence be reliable, but that reliability be assessed in a particular manner: by testing in the crucible of cross-examination.” Only “testimonial” hearsay triggered the Clause’s application — this was the key. Justice Scalia continued:

We leave for another day any effort to spell out a comprehensive definition of “testimonial.” Whatever else the term covers, it applies at a minimum to prior testimony at a preliminary hearing, before a grand jury, or at a former trial; and to police interrogations. These are the modern practices with closest kinship to the abuses at which the Confrontation Clause was directed.2 Melendez-Diaz (2009). The applicability of the Crawford regime to forensic reports was addressed in Melendez-Diaz v. Massachusetts.3 Could Massachusetts introduce three “certificates of analysis” from a state lab, created at police request, establishing that a trafficker’s seized substances were in fact cocaine? The answer, wrote Justice Scalia, was “No”:

The documents at issue here, while denominated by Massachusetts law “certificates,” are quite plainly affidavits: declaration[s] of facts written down and sworn to by the declarant before an officer authorized to administer oaths.” . . . [They] are functionally identical to live, in-court testimony, doing precisely what a witness does on direct examination.4 The certificates did not “directly accuse petitioner of wrongdoing,” but that was irrelevant. What mattered was that they “provided testimony against petitioner, proving one fact necessary for his conviction — that the substance he possessed was cocaine.”5 The Court was not swayed by the claim that the analysts were not “typical” of the witnesses that most acutely concerned the framers. The questions of autopsies came up. Justice Scalia stated that “whatever the status of coroner’s reports at common law in England, they were not accorded any special status in American practice.”6 Carolyn Zabrycki, now a California prosecutor, claimed in an article written four years after Crawford that, despite the confusion created by the decision, “one type of statement has, so far, garnered consensus: autopsy reports.”7 Melendez- Diaz disrupted all that. A number of federal and state courts have since found autopsy reports testimonial, usually reasoning, as did the U.S. Court of Appeals for the Eleventh Circuit, that the reports do “precisely what a witness does on direct examination.”8

Williams (2012) and Clark (2015). In 2012 came Williams v. Illinois,9 a long, confusing exhibition involving a state expert who referred at trial to a DNA “profile” created by the private lab Cellmark that allowed her to match up defendant Sandy Williams’s blood and semen samples. A plurality led by Justice Samuel Alito, with Chief Justice John Roberts and Justices Anthony Kennedy and Stephen Breyer, held that the Cellmark “statements” weren’t the “sort of extrajudicial statements” that the Clause barred. The statements were “sought not for the purpose of obtaining evidence to be used against petitioner, who was not even under suspicion at the time, but for the purpose of finding a rapist who was on the loose.”10 (The Court also ruled that the Cellmark statements were related by the expert “solely for the purpose of explaining the assumptions on which that opinion rests,” and so were “not offered for their truth.”11 This article argues that autopsy reports themselves are usually admissible under the Confrontation Clause without resort to the “not-for-truth” device.)

Justice Breyer, in a separate concurring opinion, said he would adhere to the dissent in Melendez-Diaz. The Confrontation Clause worked to disallow ex parte accusations; the need for cross-examination is “considerably diminished” with a statement made by an “accredited laboratory employee operating at a remove from the investigation in the ordinary course of professional work.” So anxious was Justice Breyer about the looming question of autopsies that he felt obliged to address it. The majority’s rule, he said, could bar “reliable case-specific technical information like autopsy reports”:

Autopsies, like the DNA report in this case, are often conducted when it is not yet clear whether there is a particular suspect or whether the facts found in the autopsy will ultimately prove relevant in a criminal trial. Autopsies are typically conducted soon after death. And when, say, a victim’s body has decomposed, repetition of the autopsy may not be possible. What is to happen if the medical examiner dies before trial? Is the Confrontation Clause “‘effectively’” to function “‘as a statute of limitations for murder’”?12

A dissenting Justice Elena Kagan, joined by Justices Scalia, Ruth Bader Ginsburg, and Sonia Sotomayor, argued that the Confrontation Clause, plain and simple, “applies with full force to forensic evidence of the kind involved” in the case. After all, “[c]ross-examination of the analyst is especially likely to reveal whether vials have been switched, samples contaminated, tests incompetently run, or results inaccurately recorded.”13

Finally, the Court’s recent opinion in Ohio v. Clark14 evidences an intent to apply the “primary-motive” test flexibly according to the circumstances. The Clark Court also emphasized that the presence of law enforcement involvement is critical in finding a statement to be testimonial; the Court didn’t say that a statement could never be testimonial in the absence of law enforcement involvement, but it did in effect say “hardly ever.”

Why Autopsies are Different

The Centrality of “Primary Purpose”

After Williams and Clark, all nine Justices agree on using some sort of “primary-purpose” test to determine testimoniality, but they split over what the statement’s primary purpose must be. The Alito plurality in Williams says the primary purpose, to qualify as testimonial, must be “accusing a targeted individual.” The Kagan dissenters insist that a statement is testimonial when it “establish[es] past events potentially relevant to later criminal prosecution.” Justice Thomas dislikes the primary/nonprimary distinction altogether, but agrees that a testimonial declarant “must primarily intend to establish some fact with the understanding that his statement may be used in a criminal prosecution.”

In Michigan v. Bryant,15 the Court stated that the primary-purpose test must be applied in light of all the circumstances of both the speaker and the person spoken to, and that the test is objective rather than subjective. “[C]ourts making a ‘primary-purpose’ assessment,” we were told, “should not be unjustifiably restrained from consulting all relevant information.” Nontestimonial primary purposes so far recognized include seeking medical attention (Bryant), requesting aid in a 911 call,16 catching a dangerous rapist of unknown identity,17 statements furthering a conspiracy,18 or simply shooting the breeze with an accomplice.19

Declarants may have multiple purposes, too. A perceptive statement of the primary-motive test was expressed by Judge Robert Sack in United States v. James20: A “statement triggers the protections of the Confrontation Clause when it is made with the primary purpose of creating a record for use a later criminal trial.”21

The Supreme Court has struggled to work out a definition of “testimonial,” but it has given us a way of making that determination. Under that inquiry, modern autopsy reports, in our view, are usually nontestimonial. Our conclusion is not that there is an “autopsy exception,” but rather that when an autopsy report is written under conditions like those outlined below, it simply does not come within the prohibition. It’s not a matter of “indicia of reliability” or the evidence’s importance. It’s about the reasons we perform autopsies: The primary purpose is ordinarily not to create a record for use at a later criminal trial.

Applying the “Primary-Purpose” Test to Autopsies

In July 2013, San Mateo County Coroner Robert Foucrault announced that a 16-year-old girl on an Asiana Airlines flight that crashed in San Francisco died from blunt-injury trauma. She was hit by a fire truck. It seems safe to conclude that Foucrault was not animated by a desire to flesh out a D.A.’s case for criminal negligence against firefighters or to supply facts for a federal air-safety indictment against the pilots. He was motivated by a duty he has under a California statute to determine cause of death. Police did not instigate his report and it may never be used in a criminal trial. Most courts have held that police involvement is a prerequisite to testimoniality. So generally how involved are police with autopsies?

Pathologists today operate under statutes setting out their responsibilities. In Florida, for instance, 12 situations legally trigger autopsies, among them “criminal violence,” “accident,” “suicide,” death occurring “[s]uddenly, when in apparent good health,” or “disease constituting a threat to public health.”22 The New York City Office of the Chief Medical Examiner “performs autopsies where people died in unexpected circumstances, unnatural deaths.” California’s code adds “unattended deaths” and enumerates modes of demise like a grim book of fate: “deaths due to drowning, fire, hanging, gunshot, stabbing, cutting, exposure, starvation, acute alcoholism, drug addiction, strangulation, aspiration, or . . . sudden infant death syndrome.”23

Autopsies, to be sure, can be the crux of a murder prosecution. In People v. Dungo,24 Reynaldo Dungo claimed he strangled his girlfriend in the “heat of passion” — voluntary manslaughter at most, argued his lawyer. But the autopsy revealed that she was asphyxiated for “more than two minutes.” The jury knew it was no sudden impulse. But most autopsies do not lead to criminal investigations. New York City’s medical examiner performs an average of 5,500 autopsies a year, but in 2010 only 533 city residents had homicide as their cause of death.25 Not every homicide leads to a criminal trial, so less than 10 percent of autopsy reports could possibly appear in a prosecution. In 2004 the Los Angeles Medical Examiner’s office took 9,465 cases and found that 1,121 died from homicide, 709 from suicide, 3,090 from accidents, and 4,256 from natural causes. In other words, some 90 percent of autopsies involved causes other than homicide. In a small suburban county like Marin County, California, 289 investigations reported only two homicides.26

Autopsies are associated in the American mind with criminal investigations — think “Law & Order,” “Bones,” “NCIS,” “Quincy” (for people of a certain age), etc. — and judicial discussion of autopsies is often in the context of a murder trial. But autopsies have significant purposes besides punishment. In 2007 the Centers for Disease Control analyzed autopsies in 47 states and the District of Columbia and found them essential to monitor infant mortality; to gather statistics about Alzheimer’s, meningitis, diabetes, or cirrhosis; to track prevalence of death from noxious fumes, allergies, or gun accidents. Autopsies help us effectively direct clinical-research funds. They taught us that HIV patients who died in hospitals could have been given antibiotics that would have extended their lives. They proved that prostate cancer is best detected by early screening. State laws that obligate autopsies after deaths in prisons, orphanages, or nursing homes serve to protect the vulnerable. Pathologists are the unsung heroes of consumer safety; they revealed that polyethylene bags suffocate children and that cyanide is a fatal fumigant. And before there was Vitamin Water, there was Radithor, the “radioactive water” that sold wildly until examiners weighed in. (“The Radium Water Worked Fine Until His Jaw Came Off” ran a newspaper headline.)

Autopsies established that perhaps as many as 20 percent of hospital patients die each year from misdiagnoses — and help reduce that percentage by teaching doctors that, say, what they thought was a gastric ulcer was in fact a stomach infection with sepsis. Autopsies identify dangerous new street drugs — from “wood” alcohol in 1918–19 to “bath salts” in 2011. Dr. Milton Helpern, the legendary New York City chief examiner, proved that, contrary to popular belief, heroin addicts in the mid-1930s were dying in epidemic proportions not from the opiate itself but from malaria-infected syringes. In the 1950s, his office discovered that the subtle poison of household carbon monoxide was leaking from cooking ranges and refrigerators — a design flaw that caused many wrongful murder prosecutions — and saved hundreds of lives. Autopsies revealed that fatal cocktails took Heath Ledger and Cory Monteith, and these sorts of overdose reports almost never figure in a trial against a drug dealer. Nor does a quest for indictment instigate autopsies after deadly outbreaks of salmonella or E. coli.

The point is that these types of autopsies are self-evidently nontestimonial. We should not testimonialize the work of the same neutral examiner in the same examining room conducting the same objective procedure because his subject appears to have been killed by a human being instead of a bacterium or the poison of a meth lab. The CDC survey showed that autopsies were done in 15,388 “apparent homicide” cases — equal to the number of suicides — out of a total of 173,745 autopsies performed that year.27 This means that over 90 percent of autopsy reports lacked even the possibility of use in a criminal trial. Creating prosecution evidence is not the primary purpose of autopsies in America.

A proper autopsy can never itself establish someone’s guilt. An ancient physician may have found that Julius Caesar suffered 23 bodily wounds, but only an eyewitness or confession could prove tyrannicide. From the impartial examiner’s view, the task is always the same: to show that a human being died from a particular cause. There is an impressive Sherlockian specificity here: An examiner might be able to show strangulation from harm to “neck organs consistent with fingertips,” “pinpoint hemorrhages in her eyes” indicating lack of oxygen, and self-inflicted tongue biting. Another pathologist might state that the “amount of pressure required to stop the flow of blood from the brain is ‘about 4.4 pounds’” and that death resulted when this force was kept up for “three to six minutes.” These discoveries disclose a great deal — but never the perpetrator’s identity. The report, moreover, can be used by both sides. Indeed, pathologists’ work also often terminates a prosecution by, say, establishing a time of death that matches a suspect’s alibi or by allowing the defense to show that the cause of death was a “ruptured congenital brain aneurysm” and that a fistfight “was not a contributing cause.”

Each autopsy report must be considered individually, but most autopsies fall short of testimoniality as defined by the Supreme Court. Consider Melendez–Diaz where “the sole purpose of the affidavits was to provide ‘prima facie evidence of the composition, quality, and the net weight’ of the analyzed substance.”28 Surely if a man bleeding to death in a parking lot can identify his shooter to police without “testifying” — the facts of Bryant — a pathologist can likewise relate conclusions about cause of death (not identity of suspect) without an intent to accuse. The pathologist, moreover, can deliver his statement without the medium of any police officer. The drug certificates in Melendez-Diaz were testimonial because the labs tested the powder or blood for one reason only: enforcing criminal laws on drugs. Other people may have wanted the evidence for civil suits (e.g., the K-Mart that employed Mr. Melendez-Diaz), but prosecution is the state’s reason, and the drug test was specifically done at police request. So, too, in the case with forensic disciplines like fingerprinting, ballistics, and arson analysis — designed, one and all, to prove criminality. But as shown, autopsies are mostly not conducted with the primary motivation of generating evidence for a criminal trial.

Autopsies are Generally Neutrally Performed

Forensic evidence sways juries because it seems neutral and scientific. This is why flawed or misleadingly used forensic evidence often lies behind a false conviction. Judges know this. The Wall Street Journal reported last year that recent court decisions and law-enforcement policies increasingly cast doubt on evidentiary “staples” like “hair samples, burn patterns, bite marks, ballistics evidence and handwriting analysis.”29 So are autopsies any better?

Yes — and the chief difference is that the pathologist who performs an autopsy is not an arm of law enforcement but a doctor under a civil-statutory duty to investigate mysterious deaths. Most examiners are actually private M.D.s under contract. Yet Crawford’s concern was the “involvement of government officers in the production of testimonial evidence.” When Justice Scalia blasted the notion of “neutral” government officials, he was taking aim at Washington State’s claim about neutral police officers. Being on the government payroll — like a National Weather Service forecaster or Amtrak conductor — does not make you adversarial to criminals. Crawford had in mind officials with an “investigative and prosecutorial function.” But a pathologist’s natural colleagues are not crime-lab technicians but dentists, radiologists, neurologists, and anatomists. Even routine field-written statements of Border Patrol agents can be nontestimonial, regardless of whether they later are offered at a smuggling trial.

Unlike Crawford’s interrogators, pathologists do not ask leading questions or interpret vague answers. They have no stake in the competitive enterprise of ferreting out crime. They are not praised for successful prosecutions or blamed for acquittals. They do not carry guns or badges or deceive or cajole. They conclude on the conditions of a body, not on who bears guilt for it. The National Association of Medical Examiners states that the “[p]erformance of a forensic autopsy is the practice of medicine.” A routine autopsy report — cool, impartial, precise — is akin to a careful hospital record.

Certainly the typical testimonial infirmities are absent. No issues of perception — foggy? dark? no glasses? — exist. Concerns about deteriorating memory vanish because examiners dictate or take notes while they (as they put it) “cut their case.” Verbal ambiguity is rarely a problem when speaking of “drowning due to the effects of atherosclerotic heart disease and cocaine use” (Whitney Houston) or a “[b]ullet wound of entrance at the level of the 6th cervical vertebra 5 cm. to the right of the midline” (John Dillinger). Is there a risk of fabrication? Prof. Paul Giannelli wrote a paper on crime-lab error and fraud and offered precisely one example of a pathologist’s falsification. If Bryant could say that people in mortal distress are unlikely to “fabricat[e],” we might observe that board-certified pathologists, too, have other things on their mind — namely, accuracy — and no inherent motive to lie. As Lt. Bowers of the Alameda County Coroner’s Bureau told us, a pathologist’s livelihood is premised on “credibility,” and a “tainted” doctor will struggle to find a job in county offices or lucrative defense work.

The best claim for cross-examination is to test competence. Pathologists may train for years but they are still humans exercising judgment. Mistakes can be made and conclusions at times are subjective. Yet unlike a good deal of evidence at criminal trials, autopsy reports are carefully substantiated, allowing review by others inside an examiner’s office or opposing experts. A pathologist’s tools are not interrogations but scalpels and microscopes. Even when discretion is required the work of an examiner is still almost entirely a matter of clinical recordation.

Pathologists do not become part of the prosecution as even a neutral witness does. A bystander to a crime may only want to relate what she saw, but by the time she is questioned by detectives and handed to the district attorney to be prepared for the stand, the risk of tilted testimony or one-side-only elicitations is obvious. Not with pathologists. Police and prosecutors — when kept appropriately separate from the doctor, as discussed below — cannot sway the report’s substance, or create a favorable record through suggestive questioning, or see to the omission of defendant-friendly evidence. There is no risk that the pathologist, preparing to make his “Y” incision in the body, will tell only one side of the story, because there is only one side: cause of death. “We are not interested in whodunit,” said Dr. Helpern. “All we want to know is what did it.”30

Pathologists Today Are Not the Coroners of the Common Law

A footnote in Crawford claimed “several early American authorities flatly rejected any special status for coroner statements.”31 It cited two antebellum decisions and the treatise of the great Michigan Justice Thomas Cooley. What Cooley actually wrote was closer to the reverse: He said there are “exceptions” to the rule that witnesses can be confronted in criminal cases, one being where a “witness was sworn . . . before a coroner.”32 The example reminds us that pathologists, unlike coroners, do not “swear” anyone or take evidence from any place other than their examining table. In the cases cited by Justice Scalia (respectively, from 1844 and 1858) the “coroner statements” were statements to a coroner, during a deposition, which the coroner submitted directly to the court. This is classic ex parte stuff — in a word, an inquest (which has the same root as “inquisition”). It is the opposite of the practice of the modern clinical pathologist, who with at least eight years of medical training starts with the premise that “‘I’m going to use my eyes, and I’m going to use my hands to figure out what caused the death.’”33 This is true to the etymology of “autopsy,” a mid-17th-century derivation of the Greek autopsia (“seeing with one’s own eyes”), which first appears in Westlaw’s annals only in 1843.

Crawford said, quite rightly, that applying a constitutional clause to a “phenomenon that did not exist at the time of its adoption . . . involves some degree of estimation.” Doesn’t that require us to look into what a coroner did when John Marshall strode the earth? James Wilson, the wisest framer when it came to questions of criminal procedure, described coroners as elected laymen, complementary to sheriffs, whose duty it was to summon juries and accumulate evidence. No founding-era coroner, wrote Dr. Helpern, actually “knew anything about the medical aspects of a case,” and when they bothered at all, their medical judgments were nothing more than crude layman’s guesses. Fast-forward to the Louisiana Supreme Court describing the coroner’s duties at common law, circa 1852: to “hold an inquest on the body;” to “require, at the public expense, the services of physicians, to give their opinion on the subject”; to “institute a public prosecution against the supposed perpetrator of the deed”; and to “cause [the guilty] to be arrested.”34 This is all a far cry from the work of the medical examiner today.

In 1840 Charles Dickens was part of a coroner’s inquest into an infant’s death. A beadle assembled 12 men, brought them to a morgue, and, with the coroner (a surgeon and ex-member of parliament), exhibited the body. Could the difference from modern-day practice be greater? A jury able to converse with a member of the prosecutorial apparatus and no defense presence to speak of? Jurors personally confronting (and recoiling at) ghastly evidence? An English treatise from 1883 (the year the Brooklyn Bridge opened), cited by Justice Thomas in Williams, states that coroners were “charged with investigating suspicious deaths by asking local citizens if they knew ‘who [was] culpable either of the act or of the force.’”35 In The Great Gatsby, set in 1925 New York, Fitzgerald describes a coroner brought in by police; he shows a corpse to a witness and takes her sworn statement that she did not know Jay Gatsby — not exactly medical testimony. It was only around this time New York City began to replace coroners with full-time pathologists — after the scandalizing 1915 Wallstein Report revealed that coroners, most of them bribe-hungry political hacks, were guilty of all the ineptitude one might expect of plumbers and saloonkeepers — literally — given the task of sophisticated medical evaluation.

The triumph of science made the advancement possible. In the early 1800s, we still tested for poison by feeding animals a victim’s last meal. Coroners still exist today, but they are largely elected officials who never undertake actual medical work. The word “coroner,” some three centuries older than the word “autopsy,” comes from the Anglo-French corouner, or keeper of the Crown’s pleas. In olde England, only the king examined corpses, just as he was the only man with knights enough to enforce the law. (Hence the two meanings of “court.”) A common-law coroner was an inquisitor. It was a different office in a different age.

Policy Considerations

Our argument seeks only to apply Supreme Court precedent to autopsy reports. But many important policy considerations nonetheless loom, revolving around the notion (as one court wrote) that it is “against society’s interests to permit the unavailability of the medical examiner who prepared the report to preclude the prosecution of a homicide case,” especially with such reliable, nonaccusatory evidence. The Williams plurality added another: A rule that operated to exclude neutral lab evidence would “encourage prosecutors to forgo DNA testing and rely instead on older forms of evidence, such as eyewitness identification, that are less reliable.”36

Years can pass between an autopsy and a prosecution. In one Illinois case, four gang members killed three teenagers in 1979: Two were convicted soon thereafter; one was arrested in California in 1988 and pleaded guilty; the last was only convicted in 1992. We should look doubtfully on a misinterpreted right of confrontation that allows murderers to escape justice by avoiding arrest or delaying trial long enough. This would effectively impose a “statute of limitations” on one of the few crimes that knows no such time limit.

Autopsy reports, unlike drug substance tests, cannot be replicated. Disinterment or cold storage is sometimes an option but another complete autopsy never is. It is true that evidence can also be lost forever with, say, an eyewitness who never testifies and then dies. The difference is that with most witnesses testimonial infirmities are usually important, whereas with autopsies, they almost never are. Melendez-Diaz said that the “prospect of confrontation” would “deter fraudulent analysis,” but fraud is a problem quite separate from testimoniality.37 A wealthy county like Marin in the San Francisco Bay Area may have two homicides a year, but across the water in Alameda County — home to Oakland — examiners might perform one or two homicide autopsies a day. Dr. Thomas Beaver, that county’s chief pathologist, estimates that he is under subpoena to appear in court every single day. He told us that a rule finding autopsy reports to be testimonial would force examiners’ offices like his to choose between time on their work and letting prosecutions fail.

Finally, the rule of the Williams dissenters — a statement is testimonial if made primarily to prove events “potentially relevant to later criminal prosecution” — is simply too diffuse. What wouldn’t be “potentially” relevant? (This article, we hope, is potentially relevant to prosecutions.) Melendez- Diaz affirmed that business records are ordinarily admissible without confrontation, even though potentially relevant to a later criminal prosecution. For Fortune 500 companies, millions of business records are potentially relevant to criminal charges; multinational corporations keep due-diligence inquiries for FCPA subpoenas and hedge funds catalog emails to fend off insider-trading charges. Or a pharmacist may know that her legally mandated logs of pseudoephedrine purchases will be used against meth dealers; the logs are still nontestimonial business records. The dissenters’ test really turns on the “primary purpose” and not the “potentially relevant” part. The inquiry is into the totality of circumstances under which the record was prepared.

When an Autopsy Report Is Nontestimonial

Formal Separation

Before Melendez-Diaz, courts regularly held that autopsy reports were admissible as nontestimonial business or public records. Since that decision the most important factor for judges undertaking the primary-purpose inquiry with autopsy reports has been the degree of police involvement in the report’s creation. For that reason the best way to avoid a confrontation problem is to ensure that an examiner’s work is maximally independent of police and prosecutorial influence. We looked at federal circuit and state supreme courts that ruled on this issue since Melendez- Diaz. Most of them declined, properly in our view, to set out a categorical rule about whether autopsies are testimonial.

The then-chief D.C. medical examiner testified about autopsy reports placed into evidence. The D.C. Circuit saw testimonial reports, “document[s] created solely for an evidentiary purpose made in aid of a police investigation.”38

For instance, observed the court:

  • The Office of the Medical Examiner was “required” by the D.C. Code to investigate deaths when requested by the Metropolitan Police Department or U.S. Attorney’s Office.
  • “Law enforcement officers . . . not only observed the autopsies, a fact that would have signaled to the medical examiner that the autopsy might bear on a criminal investigation, they participated in the creation of reports,” i.e., they were “present” at several examinations and they supplemented one report with a “crime diagram” and wrote in another: “Should have indictment re: John Raynor for this murder.”

In United States v. Ignasiak,39 a doctor was convicted for overprescribing deadly pain medications. The Eleventh Circuit found the reports testimonial. They were prepared “for use at trial” under a “statutory framework” in which “medical examiners worked closely with law enforcement”:

  • “Under Florida law, the Medical Examiners Commission was created and exists within the Department of Law Enforcement. Fla. Stat. § 406.02.”
  • The Commission “must include one member who is a state attorney, one member who is a public defender, one member who is sheriff, and one member who is the attorney general or his designee, in addition to five other noncriminal justice members.”
  • The examiner “relied upon information collected by ‘deputies on the scene.’”

In United States v. James,40 conviction in a creepy conspiracy to murder for insurance cash turned on toxicology reports and autopsies that showed whether the deaths were accidental or caused by malicious poisoning. The Second Circuit found the reports nontestimonial: the “circumstances under which the analysis was prepared” didn’t “establish that the primary purpose of a reasonable analyst in the declarant’s position would have been to create a record for use at a later criminal trial.” The “key,” said the court, was the “particular relationship between [the medical examiner’s office] and law enforcement”:

  • “[N]either the government nor defense counsel elicited any information suggesting that law enforcement was ever notified that Somaipersaud’s death was suspicious, or that any medical examiner expected a criminal investigation to result from it.”
  • “There is no indication in Brijmohan’s testimony or elsewhere in the record that a criminal investigation was contemplated during the inquiry into the cause of Sewnanan’s death,” especially since the facts at the time suggested “accidental ingestion or suicide.”

In State v. Kennedy,41 an autopsy report showed that the victim’s head had been bashed in. Prosecutors actually conceded that the report was testimonial, but the West Virginia Supreme Court of Appeals nonetheless proceeded to apply the “primary purpose” test to find testimoniality:

  • “[M]ost compellingly, the autopsy and required report’s use in judicial proceedings is one of its statutorily defined purposes.” The examiner is obligated to assist in the “formulation of conclusions, opinions or testimony in judicial proceedings.”
  • “Kennedy was under suspicion and in fact, in custody, when the autopsy was conducted and therefore the autopsy report could arguably be said to have been prepared to ‘accuse a targeted individual.’”
  • “Dr. Sabet testified that law enforcement officers [were] present during the autopsy, providing a ‘detailed history’ and engaging in a dialogue with the medical examiner about cause of death,” which “suggests a collaborative investigative effort in making the case against a suspect.”

In People v. Leach,42 a husband strangled his wife to death. The Illinois Supreme Court found the autopsy report nontestimonial: It was neither “prepared for the primary purpose of accusing a targeted individual” nor “for the primary purpose of providing evidence in a criminal case.”

  • “[A]lthough the police discovered the body and arranged for transport, there is no evidence that the autopsy was done at the specific request of the police. The medical examiner’s office performed the autopsy pursuant to state law, just as it would have if the police had arranged to transport the body of an accident victim.”
  • “Although [Dr. Choi] was aware that the victim’s husband was in custody and that he had admitted to ‘choking’ her, his examination could have either incriminated or exonerated him, depending on what the body revealed about the cause of death . . . . Dr. Choi was not acting as an agent of law enforcement, but as one charged with protecting the public health by determining the cause of a sudden death that might have been ‘suicidal, homicidal or accidental.’”
  • “Unlike a DNA test which might identify a defendant as the perpetrator of a particular crime, the autopsy finding of homicide did not directly accuse defendant. Only when the autopsy findings are viewed in light of defendant’s own statement to the police is he linked to the crime. In short, the autopsy sought to determine how the victim died, not who was responsible, and, thus, Dr. Choi was not defendant’s accuser.”

In People v. Dungo, Dungo admitted to choking his girlfriend after a fight. The California Supreme Court held that the expert’s testimony about the autopsy report did not require confrontation of the report’s author.

  • “Criminal investigation . . . [is] only one of several purposes” for autopsies: “the decedent’s relatives may use an autopsy report in determining whether to file an action for wrongful death. And an insurance company may use an autopsy report in determining whether a particular death is covered by one of its policies . . . . Also, in certain cases an autopsy report may satisfy the public’s interest in knowing the cause of death, particularly when (as here) the death was reported in the local media. In addition, an autopsy report may provide answers to grieving family members.”43
  • “The presence of a detective at the autopsy and the fact that the detective told the pathologist about defendant’s confession” did not make the report testimonial because the report itself was “simply an official explanation of an unusual death.”

Finally, in State v. Navarette,44 a man was shot from a car. At issue was whether the shooter was the car’s driver or passenger Navarette. Dr. Zumwalt, New Mexico’s chief medical examiner, testified using a colleague’s report that the bullet wound and its lack of soot were consistent with Navarette’s position in the car. This was testimonial, said the New Mexico’s Supreme Court:

  • Dr. Zumwalt “conceded that it was immediately clear that this autopsy was part of a homicide investigation” and said two police officers had attended the autopsy.
  • Examiners were under a statutory duty to report about individuals who “die suddenly and unexpectedly,” so there was “no reason” an examiner should not “anticipate[] that criminal litigation would result.”

The court concluded that the pathologist’s findings “went to the issues of whether Reynaldo’s death was a homicide and, if so, who shot him. These issues reflected directly on Navarette’s guilt or innocence.”

But if an examiner has a decedent and nothing more, how could her findings possibly reflect directly on Navarette’s guilt or innocence? How would the pathologist know who Navarette even was? A drowning could be a crime or a poolside tragedy; a heart attack could be caused by obesity or arsenic. It is only when a cause-of-death finding is linked to evidence extraneous to the report that a conviction happens.

Melendez–Diaz noted that the drug analyst’s job existed “under Massachusetts law.”45 Courts have followed suit in examining the terms of autopsy-authorizing statutes, which vary considerably. California, for instance, has three different models among its counties, and a statute providing that a pathologist’s “[i]nquiry . . . does not include those investigative functions usually performed by other law enforcement agencies.”46 In Kansas, a pathologist can obtain “law enforcement background information” or perform an “examination of the scene of the cause of death.”47 Statutory provisions are just one element in the totality inquiry and probably lack the significance courts ascribe to them. A separate examiner’s office could be muscled by a sheriff, while an examiner with a lab in a police basement could still maintain perfect neutral integrity. Statutes say little about what actually happens, such as the extent to which a pathologist confers with police or family members to get the facts before an exam.

Melendez-Diaz observed that the “majority of [labs producing forensic evidence] are administered by law enforcement agencies.”48 Not so with pathologist operations: 43 percent of Americans are served by independent coroner or examiner offices and another 14 percent by offices within health departments. Some medical examiners are even part of a university’s school of medicine. Most autopsies occur in mortuaries or hospital pathology wings.

The third or so of pathologists who work within law-enforcement bailiwicks do so not because their work is primarily related to law enforcement, but for administrative reasons. Many rural or suburban counties simply can’t afford to separately fund or house examiners and law enforcement. Mortuaries are costly; insurers don’t cover autopsies. Marin County saved $500,000 a year by merging its coroner and sheriff’s offices. The fact is that when corpses are involved, both law enforcement and examiners must be, too. Federal judges might observe that for similar administrative reasons their courthouses also host U.S. attorneys, ATF agents, or federal marshals without compromising the judiciary’s integrity.

We hesitate to suggest that examiners should have no contact with law enforcement. Pathologists want all available information. This can mean acquiring police reports — or the reports of paramedics or firemen, or medical histories and hospital records. Sometimes it means a pre-autopsy conference with police or a talk with the victim’s family. New York’s Dr. Helpern — who estimated that he performed some 20,000 autopsies and supervised 60,000 more over 45 years — wrote that in cracking the famous Coppolino murders a witness’s tip that a victim had been injected with succinylcholine, a nearly undetectable muscle relaxant, was essential. “Had I been doing this autopsy without knowing the history of the case,” he wrote, he might have missed the “tiny pink spot” on the left buttock that marked the needle’s point of entry. His resourceful toxicologist then proceeded to invent a method to trace the substance in the victim’s organs. Confrontation Clause jurisprudence cannot be so hyper-technical as to impose a rule that might make medical examiners’ reports less thorough and reliable.

In another case, Dr. Helpern explained why he believed Ms. Carolee Biddy, in her day a noted murder defendant, was wrongly convicted. Her stepdaughter had gotten into a powerful drain cleaner. The pathologist, unaware of this fact, gave the cause of death as asphyxia, which it was. But the jury saw it as Ms. Biddy’s doing, when Dr. Helpern, after studying photos of the girl’s epiglottis, was sure that her throat had swollen shut from the chemical. A wall between examiner and the case’s known facts, besides being pointless in noncriminal cases, will allow murderers to escape and innocents to suffer. If there remains a concern about law-enforcement involvement, a solution is to require pathologists to keep a record of contact with police, so that defense counsel can later look for improper influence or misrepresentation. But one is more likely to find a pathologist influencing law enforcement — especially in invalidating a theory of the detectives or prosecutors — than the other way around.

To assure that autopsy reports avoid the Confrontation Clause’s prohibition, statutes and protocols should provide that pathologists receive no guidance from police beyond the receipt of basic facts and no specifics about the identity of possible perpetrators. A report might properly reference a “subarachnoid hemorrhage,” as one did, but it should not have mentioned the beating at the parking lot. (If a reference to outside facts creeps in, redact it.) Pathologists should be cautious about visiting a murder scene — uncommon anyway once “medical investigators” assumed this role — an act that risks police-doctor contact. Reports should be nonaccusatory and devoid of legal conclusions. As put by the National Association of Medical Examiners, the task is to produce a “neutral and objective medical assessment of the cause and manner of death.”

Expert Testimony

If an autopsy report is testimonial, can one doctor testify using the work of another? This practice is a real problem for defendants. In Ignasiak, for instance, one doctor testified about autopsies performed by another, but “Dr. Minyard indicated she lacked enough information to agree or disagree with Dr. Kelly’s conclusion that patient S.P.’s death was a suicide” and “could not testify from direct knowledge about the condition of a particular patient’s heart, lungs, or brain and, as a result, whether that patient may have actually died from a heart attack, stroke, or some cause other than drug overdose.”49 A nonautopsying expert will be able speak to procedure, highlighting an office’s diligence and expertise, but not about the one-off errors and oversights that are precisely what the defense seeks to uncover.

Justice-counting in Williams leads to the conclusion that unadmitted autopsy reports, if testimonial, cannot serve as the basis for the opinion of an expert who played no role in the autopsy, even if the testimony is the expert’s own independent conclusion and he can be cross-examined about it. The New Mexico Supreme Court believes this was decided by Justices Thomas and Kagan in Williams, the latter rejecting such testimony as a “neat trick.” Precedents already disallow “surrogate” testimony or testimony that is a “mere conduit” for inadmissible evidence. The logic, per Melendez-Diaz, is that confrontation lets a defendant test “honesty, proficiency, and methodology” — even when examiners boast the “scientific acumen of Mme. Curie and the veracity of Mother Theresa.”50

Justices Alito, Kennedy, and Breyer, and Chief Justice Roberts disagree: Expert statements made “solely for the purpose of explaining the assumptions on which that opinion rests are not offered for their truth.”51 That’s only four votes. The better path is to admit autopsy reports directly, as nontestimonial reports under the “primary-motive” test, after assuring that law-enforcement involvement is not so pervasive as to prevent a finding that the report was prepared for purposes unrelated to use as evidence in a criminal trial.

Conclusion: The Proper Test

There is no “autopsy” exception to the Confrontation Clause. We think, rather, that the vast majority of autopsy reports are just outside the clause’s scope — presumptively nontestimonial in light of a medical examiner’s actual work. At least 90 percent of autopsy reports in fact do not have a primary purpose in furnishing evidence for a prosecution. The primary purpose, the predominating purpose, is public health. Even in the fraction of cases where a report is eventually used in a prosecution, that doesn’t mean the report was prepared for such a purpose.

Courts should presume that autopsy reports are nontestimonial because they are written independently by neutral doctors concerned with accuracy, not police officers seeking conviction. As Clark so recently emphasized, active involvement by the prosecution is virtually required in order to find that a statement is testimonial, so our proposed presumption is quite consistent with the Court’s recent take on the Confrontation Clause. The presumption isn’t overcome by the fact that the examiner and the police might be administratively conjoined. The proper test for the Confrontation Clause, fairly applied, is:

Has there been specific and pervasive involvement by law enforcement in the preparation of the autopsy report, such as to change the basic character of the document from one serving pathological purposes to one primarily serving prosecutorial purposes?

Only when that line is crossed does a medical examiner become a “witness” against the accused. Defense allegations that this happened should be litigated in light of a record of contacts kept by the medical examiner.

The state cannot generate evidence against the accused without the right of confrontation. But to demand confrontation of every autopsy report in a prosecution would be to misinterpret a noble principle and would very likely subvert justice before promoting it. This is an exceedingly unstable area of law. The proper application of the Confrontation Clause does not command a majority in the Supreme Court or consensus in the states. Which means there is still time to do the right thing.


  1. 541 U.S. 36 (2004).
  2. Id. at 86.
  3. 557 U.S. 305 (2009).
  4. Id. at 310–11.
  5. Id. at 313 (alteration in original).
  6.  Id. at 322.
  7. Carolyn Zabrycki, Toward a Definition of Testimonial: How Autopsy Reports Do Not Embody the Qualities of a Testimonial Statement, 96 Cal. L. Rev. 1093, 1099 (2008).
  8. United States v. Ignasiak, 667 F.3d 1217, 1230 (11th Cir. 2012).
  9. 132 S. Ct. 2221 (2012).
  10. Id. at 2228.
  11. Id.
  12. Id. at 2251 (Breyer, J., concurring) (internal citations omitted).
  13. Id. at 2264–2265 (Kagan, J., dissenting).
  14. 135 S. Ct. 2173 (2015).
  15. 562 U.S. 344 (2011).
  16. Davis v. Washington, 547 U.S. 813 (2006).
  17. Williams v. Illinois, 132 S. Ct. 2221 (2012).
  18.  United States v. Grasso, 724 F.3d 1077 (9th Cir. 2013).
  19. United States v. Cazares, 788 F.3d 956 (9th Cir. 2015) (“a conversation between two gang members about the journey of their burned gun is not testimonial”).
  20. 712 F.3d 79 (2nd Cir. 2013).
  21. Id. at 97.
  22. Fla. Stat. § 406.11(1)(a)-(c) (West 2006). Most states seem to have variations on this.
  23. Cal. Gov. Code § 27491 (West 1969).
  24. 286 P.3d 442 (Cal. 2012).
  25. James, 712 F.3d at 99 n.10 (citing OCME, General Information Booklet, http://www. nyc.gov/html/ocme/downloads/pdf/General% 20Information/OCME% 20General% 20Information% 20Booklet.pdf; Deaths and Death Rates by Selected Causes New York City— 2010, http://www.health.ny.gov/statistics/ vital_statistics/2010/table33c.htm).
  26. Marin County Sheriff’s Office, Coroner Division Annual Report, 2011, 6, available at http://www.marincounty.org/main/board-actions/2013/may/may- 14/~/media/Files/MarinGov/Board%20ctions/2013/20130514SOCoroner2011.pdf.
  27. See, e.g., Donna L. Hoyert, Hsiang-Ching Kung, & Jiaquan Xu, Autopsy Patterns in 2003, 20(32) Vital and Health Statistics, March 2007.
  28. Melendez-Diaz v. Massachusetts, 557 U.S. 305, 311 (2009).
  29. National Association of Medical Examiners, Forensic Autopsy Performances Standards 4 (2006), available at http://www.mtf.org/pdf/ name_standards_2006.pdf.
  30. Milton Helpern with Bernard Knight, Autopsy: The Memoirs of the Milton Helpern, the World’s Greatest Medical Detective 66–71, 175–76 (1977).
  31. Crawford, 541 U.S. at 47 n.2.
  32. Thomas M. Cooley, Constitutional Limitations 318 (1868).
  33. United States v. Ignasiak, 667 F.3d 1217, 1232 n.19 (11th Cir. 2012).
  34. State v. Parker, 7 La. Ann. 83, 84 (1952).
  35. Williams, 132 S. Ct. at 2262 (Thomas, J., concurring).
  36. Id. at 2228 (Alito, J.).
  37. Melendez-Diaz v. Massachusetts, 557 U.S. 305, 318–19 (2009).
  38. 651 F.3d 30 (D.C. Cir. 2011).
  39. 667 F.3d 1217 (11th Cir. 2012).
  40. 712 F.3d 79, 85 (2d Cir. 2013).
  41. 735 S.E.2d 905 (W.Va. 2012).
  42. 980 N.E.2d 570 (Ill. 2012).
  43. People v. Dungo, 286 P.3d 442 (Cal. 2012).
  44. 294 P.3d 435, 436 (N.M. 2013).
  45. Melendez-Diaz v. Massachusetts, 557 U.S. 305, 311 (2009).
  46. Cal. Gov’t Code § 27491 (2012).
  47. Kan. Stat. Ann. § 22a-231 (2000).
  48. Melendez-Diaz, 557 U.S. at 318.
  49. Ignasiak, 667 F.3d at 1234.
  50. Melendez-Diaz, 557 U.S. at 319 n.6.
  51. Williams v. Illinois, 132 S. Ct. 2221, 2228 (2012)