The Drowning Gap: When Medical Examiners Can't Tell Murder from Accident

South Side, Last Call

Dakota James left Cupka’s Cafe 2 on East Carson Street sometime around 11:30 PM on January 25, 2017. He was 23 years old, a University of Pittsburgh graduate working at a consulting firm downtown, the kind of person whose friends would describe him as careful even when he’d been drinking. South Side is Pittsburgh’s bar district, a strip of converted rowhouses running parallel to the Monongahela River where weekend foot traffic spills from one establishment to the next until the sidewalks thin out near the Birmingham Bridge.

Security cameras captured James walking east on Carson Street. Then the footage stopped. No camera recorded him entering the water. No witness saw him fall. No one heard a shout or a splash in the 28-degree January air. He was simply present on the street, and then he was not.

His family reported him missing the following day. Pittsburgh police initiated a search of the riverbanks and the waters near South Side. Volunteer dive teams joined. Days passed, then weeks. The Monongahela carried whatever it carried downstream into the Ohio River, where the current broadens and the banks give way to industrial flats and rail yards west of the city.

Forty days after Dakota James vanished from a bar district sidewalk, a kayaker spotted a body floating face-up in the Ohio River near McKees Rocks, roughly eleven miles from South Side. It was James. His wallet was in his pocket.

The Allegheny County Medical Examiner’s Office ruled the death accidental drowning.

The Forensic Record

Dr. Cyril Wecht, the former Allegheny County coroner and one of the most prominent forensic pathologists in the United States, reviewed photographs and documentation from the Dakota James case at the family’s request. Wecht’s findings contradicted the official ruling.

Wecht identified ligature marks on James’s neck consistent with the application of a cord or similar binding. He documented bilateral fingernail bed discoloration, a finding that in forensic literature correlates with asphyxiation caused by oxygen deprivation prior to death. The discoloration was present on both hands. In Wecht’s assessment, these findings indicated manual or ligature strangulation rather than drowning.

Wecht did not perform the original autopsy. He reviewed photographic evidence and the available case file. This distinction matters, because critics rightly note that photo review carries limitations that hands-on examination does not. Lighting, resolution, and angle can produce artifacts. Wecht acknowledged the limitations of his review while maintaining that the visible findings warranted a full criminal investigation.

In 2018, the Allegheny County District Attorney’s office convened a meeting that included representatives from the FBI and the Secret Service. According to reporting by local journalists who obtained records of the meeting, the DA’s office acknowledged the presence of strangulation evidence in the James case. The discussion addressed whether additional forensic resources should be deployed.

No criminal investigation followed. The manner of death remained accidental. The body had already been released to the family.

What the DA’s office conceded in that 2018 meeting sits in the public record like an unexploded device. Law enforcement acknowledged forensic evidence of strangulation in a death ruled accidental, then decided not to investigate. The question was never whether the evidence existed. The question was whether anyone would act on it.

Five Men, Three Rivers, Twelve Years

Dakota James was not the first young man to vanish from Pittsburgh’s waterfront bar district and surface in one of the city’s rivers weeks later. He was not the last.

Jimmy Slack, August 2011. Slack, 29, left a bar on the South Side and was not seen again. His body was recovered from the Monongahela River roughly two weeks later. The medical examiner ruled accidental drowning. Slack’s wallet was recovered with the body.

Paul Kochu, December 2014. Kochu, 22, a Carnegie Mellon University student, was last seen near the waterfront area of the Strip District in the early morning hours. His body was recovered from the Allegheny River approximately six weeks later. Cause of death: accidental drowning. Kochu had been celebrating the end of the fall semester. His wallet was on his person when recovered.

Dakota James, January 2017. The details above. Forty days in the water. Ligature marks identified by independent review. Ruled accidental.

Thomas Hughes, January 2021. Hughes, 26, disappeared after leaving a South Side bar in the early morning hours of a January night. His body was recovered from the Monongahela River approximately five weeks later. Accidental drowning.

Brandon Pfeiffer-Davis, October 2023. Pfeiffer-Davis, 22, was last seen near the waterfront in the South Side area. His body was pulled from the river weeks later. Accidental drowning. Wallet intact.

The similarities arrange themselves without commentary. Young men between 22 and 29. Waterfront bar districts. Late-night or early-morning disappearances during cold months. Extended periods of river submersion before recovery, typically two to six weeks. Wallets found on the bodies. Identical cause-of-death rulings from the same medical examiner’s office.

Five cases over twelve years is a small number. Pittsburgh sits at the confluence of three rivers with an active bar district on the banks of one of them. Young men drink too much and fall into water every year in every city with a waterfront. That is the skeptical case, and it requires serious consideration rather than dismissal.

But the skeptical case does not explain the forensic findings in the James case. It does not explain why law enforcement acknowledged strangulation evidence and then closed the file. It does not address whether the remaining four deaths received the kind of investigation that could distinguish accident from homicide when the body has spent weeks decomposing in cold water.

The answer, in every case, is that they did not.

Tommy Booth: The Body That Would Not Cooperate

The strongest forensic case in this category of deaths did not occur in Pittsburgh. It occurred in the Philadelphia metropolitan area in 2008, and the details make the institutional failure almost impossible to explain away.

Thomas “Tommy” Booth was 22 years old when he disappeared on February 19, 2008. His body was recovered from a body of water fourteen days later. In most water-recovery cases, two weeks of submersion produces significant decomposition, enough to obscure external injuries and complicate cause-of-death determination. Booth’s body did not decompose.

The forensic findings were extraordinary. Full rigor mortis was present at recovery, a condition that in normal circumstances resolves within 36 to 72 hours after death. Rigor persisting at 14 days suggested either extreme cold preservation or that Booth had not been in the water for the entire period between disappearance and recovery.

Lividity, the settling of blood due to gravity after death, was present on Booth’s back. This means his body was supine, face-up, in a fixed position for a sustained period after death. A person who drowns after falling face-first into water does not develop lividity on the posterior surface of the body. The lividity pattern indicated Booth died on his back, on a flat surface, before entering the water.

No water was found in his lungs. A drowning victim aspirates water in the final moments of life, producing characteristic findings in the pulmonary tissue. The absence of water in Booth’s lungs meant he did not drown. He was dead before he entered the water.

According to investigators who worked the case, the medical examiner privately stated the case was 99% likely a homicide. The constellation of findings, no decomposition, persistent rigor, posterior lividity, dry lungs, pointed almost exclusively to a person who was killed elsewhere, stored in a cold environment, and placed in the water days or weeks after death.

The official manner of death was ruled undetermined.

Not homicide. Not accident. Undetermined. The medical examiner who privately expressed near-certainty that this was a killing would not commit that assessment to paper. The gap between what the examiner said in conversation and what the examiner wrote on the death certificate is the gap this entire article is about.

What Maryland Proved

In May 2025, the state of Maryland completed an independent audit of water-recovery deaths that had been classified by county medical examiner offices across the state. The audit was prompted by legislative pressure following a series of investigative reports documenting cases where families had challenged drowning rulings.

The audit examined 87 deaths. An independent panel of forensic pathologists, none of whom had been involved in the original rulings, reviewed autopsy records, scene photographs, toxicology reports, and law enforcement files for each case. The panel operated on consensus: reclassification required unanimous agreement among all reviewing pathologists.

Of the 87 deaths reviewed, the panel unanimously reclassified 36 as homicides.

Thirty-six of eighty-seven. Not a marginal correction. Not a handful of edge cases. Forty-one percent of the deaths originally classified as accidental drowning were, upon independent review, unanimously determined to be killings.

The reclassified cases shared characteristics with the Pittsburgh pattern and the Booth case. Bodies recovered from water after extended submersion. Original autopsies that found no competing cause of death and defaulted to drowning. Scene evidence and documentary records that, when examined with fresh eyes, revealed indicators the original examiner had missed, minimized, or declined to pursue.

The Maryland audit did not suggest that medical examiners were corrupt. It demonstrated something more systemic: the structure of death investigation in the United States produces predictable errors in water-recovery cases. The errors flow in one direction. They classify homicides as accidents, not the reverse.

No comparable audit has been conducted in Pennsylvania. No one has applied the Maryland methodology to Allegheny County’s water-recovery deaths. No one has asked the question at an institutional level, and the absence of the question is itself an answer.

The Machinery of Misclassification

Drowning is the only common cause of death that is diagnosed by exclusion. There is no definitive autopsy finding that confirms drowning. Unlike a gunshot wound, a stab wound, or blunt force trauma to the skull, drowning leaves no signature. The forensic pathologist examines the body, rules out every identifiable cause of death, and if nothing else explains the death of a person recovered from water, records drowning as the cause.

This diagnostic framework works reasonably well when the body is recovered quickly. Soft tissue remains intact. External injuries are visible and photographable. Ligature marks, petechial hemorrhaging, defensive wounds, injection sites: all of these can be identified within hours of death.

The framework collapses when the body has been in water for weeks.

Cold water slows decomposition but does not stop it. Aquatic organisms feed on soft tissue, beginning with the face, hands, and genitals, the areas with the thinnest skin. Lividity becomes unreliable as gases from decomposition redistribute blood through the vascular system. Skin slips from the hands, destroying evidence under fingernails. Bruising fades or becomes indistinguishable from postmortem color changes. Ligature marks on the neck can be confused with the natural creasing of decomposing tissue.

A forensic pathologist examining a body recovered after four weeks in a 38-degree river is working with degraded evidence in every category. The honest answer in many of these cases is that the cause of death cannot be determined. But medical examiners face institutional pressure not to write “undetermined” on death certificates. Insurance companies dispute undetermined rulings. Law enforcement treats undetermined as an open case requiring resources. Families demand answers. Undetermined is a bureaucratic headache for every entity that touches the file.

Accidental drowning closes the loop. Insurance pays. Police close the case. The family gets a death certificate that allows them to settle estates and move through probate. The medical examiner clears the file from the queue and moves to the next case.

Allegheny County’s medical examiner office, like most in the United States, operates under budget constraints that limit the number of pathologists on staff, the time available per case, and the ability to conduct specialized forensic testing. The National Association of Medical Examiners recommends a maximum caseload of 250 autopsies per pathologist per year. Many offices exceed this number by 50% or more. When you are running 375 cases a year and a body pulled from the Monongahela shows no obvious competing cause of death, the path of least resistance is accidental drowning.

None of this requires malice. None of it requires conspiracy. It requires only an underfunded system handling decomposed remains with a diagnostic framework that defaults to accident when the evidence degrades past the point of certainty. The system does not fail because bad people run it. The system fails because it was designed for fresh bodies examined within hours, and it encounters bodies that spent six weeks in a river.

The Strongest Skeptical Case

Intellectual honesty requires addressing the statistical argument against foul play, because it is a strong argument.

Approximately 4,300 Americans die from unintentional drowning each year. Seventy-six percent of drowning victims are male. Alcohol is a factor in roughly 70% of water-related fatalities among adults. The demographic most likely to drown accidentally is young men between 18 and 34 who have been drinking, which is an exact description of every Pittsburgh case.

Pittsburgh has three rivers converging at its downtown core. The South Side bar district runs within a few hundred feet of the Monongahela. The riverwalk paths are accessible and often unfenced. In winter, when temperatures drop below freezing and riverside surfaces become slick, a person who has been drinking heavily faces genuine risk of an accidental fall into water that will kill within minutes through cold water shock.

Five deaths over twelve years in a three-river city with major waterfront nightlife. Run the base rates and you might expect more, not fewer. A Bayesian analysis that begins with the prior probability of accidental drowning in this exact demographic and this exact geography will produce a posterior estimate that heavily favors accident in any individual case.

Wecht reviewed photographs, not the physical body. He was retained by the James family, which creates an adversarial posture that a neutral reviewer might not share. The published forensic literature documents cases where cold-water immersion produces minimal decomposition in water temperatures between 34 and 42 degrees Fahrenheit, which complicates the timeline arguments in the Booth case without requiring a refrigerated storage theory.

These are not trivial objections. Anyone who cares about getting the answer right, rather than getting the answer that confirms a preferred theory, has to sit with the statistical argument long enough to feel its weight.

But statistics describe populations. They do not examine individual bodies. The question is not whether young men drown accidentally in Pittsburgh. They do. The question is whether these specific men drowned accidentally, and whether the investigations conducted were sufficient to make that determination.

The ligature marks on Dakota James’s neck are not explained by base rates. The DA’s acknowledgment of strangulation evidence is not resolved by population-level drowning statistics. Tommy Booth’s posterior lividity and dry lungs cannot be averaged away. The Maryland audit did not reclassify 41% of its cases because the original examiners were working with impossible evidence. It reclassified them because the original examiners were not looking hard enough.

The skeptical case explains why we should not assume these were homicides. It does not explain why we should assume they were properly investigated. Those are different questions, and conflating them is the error that allows the system to persist unchallenged.

What Accountability Would Require

The Maryland audit provides a template. It is not a radical proposal. It is a procedural mechanism that other states could replicate if the political will existed.

Step one is straightforward. File FOIA requests with the Allegheny County Medical Examiner’s Office for the complete autopsy records, scene photographs, toxicology reports, and investigative notes for every water-recovery death classified as accidental drowning within the county over the past two decades. Pennsylvania’s Right-to-Know Law provides a legal mechanism for this, though medical examiner offices routinely resist disclosure by citing ongoing investigation exemptions, even when no investigation is active.

Step two requires an independent forensic panel. Not the same pathologists who made the original rulings. Not pathologists employed by the same county. An external panel with no institutional relationship to Allegheny County, applying the same consensus methodology Maryland used: review the full file, make an independent determination, require unanimity for reclassification.

Step three is the difficult one. When the panel identifies cases where the manner of death should have been classified differently, those findings need to trigger actual criminal investigation. Reclassification without investigation is an academic exercise. In Maryland, the reclassified cases were referred to law enforcement for potential reopening. Whether those referrals produce investigations, indictments, or convictions remains to be seen, but the referral itself breaks the closed loop where a death ruled accidental stays accidental regardless of new evidence.

Step four addresses the structural problem. ME offices need caseload limits that are actually enforced, not recommended. They need dedicated forensic resources for water-recovery cases, including consultation protocols that trigger automatically when a body has been submerged for more than 72 hours. They need quality review processes where a percentage of cases are audited annually by external reviewers, the same mechanism that hospitals use for surgical outcomes and that labs use for diagnostic accuracy.

None of this is exotic. It is standard institutional accountability applied to an institution that has historically operated without it. Medical examiner offices in the United States function with less external oversight than the average restaurant kitchen. The health department inspects the walk-in cooler at your local sandwich shop more frequently than anyone inspects the determinations coming out of most county morgues.

In Allegheny County, the families of Dakota James, Paul Kochu, Thomas Hughes, and Brandon Pfeiffer-Davis received death certificates that said their sons drowned accidentally. Those certificates may be correct. Some or all of these young men may have fallen into the river after drinking too much on a freezing night. That is a real possibility and not one to be discounted.

But the families did not receive investigations capable of distinguishing that possibility from the alternative. They received a default. The system gave them the answer it gives everyone when a body comes out of the water too late and too decomposed for easy determination. It gave them the answer that closes the file.

In Maryland, when someone finally asked whether those files should have been closed, 41% of them opened back up as homicides. The question has never been asked in Allegheny County. Until it is, the drowning gap stays open.

Sources

• Allegheny County Medical Examiner’s Office — Dakota James case file (2017)
• Dr. Cyril Wecht — Forensic Pathology Consultation — Forensic review commissioned by James family (2017-2018)
• Pittsburgh Post-Gazette — DA Meeting with FBI/Secret Service (2018)
• FBI — Statement on “Smiley Face Killer” Theory (2008) — “No evidence of organized serial killing”
• Maryland Office of the Chief Medical Examiner — Independent Audit Results (2025) — 87 water-recovery deaths reviewed, 36 reclassified as homicides
• National Association of Medical Examiners — Caseload Standards — Recommended maximum 250 autopsies per pathologist per year
• CDC — National Center for Health Statistics, Unintentional Drowning Data — Approximately 4,300 deaths annually, 76% male
• CDC — Alcohol and Drowning Factsheet — Approximately 70% of adult drowning fatalities involve alcohol
• Pittsburgh Post-Gazette — South Side Drowning Cases (2011-2023)
• KDKA — Pittsburgh River Deaths Coverage
• Pennsylvania Right-to-Know Law, 65 P.S. § 67.101 et seq.