Uncounted COVID Deaths Reveal a Troubling Truth About Official Death Records
by Jo Ellen Nott
In a study led by Andrew Stokes of Boston University, researchers developed algorithms trained on hospital-verified data to audit out-of-hospital deaths during the COVID-19 pandemic. The results, published in Science Advances, point to “critical gaps in the U.S. death investigation system.”
The study’s scope was narrower than all pandemic-era deaths. Researchers analyzed adults 25 and older who died from natural causes between March 2020 and December 2021, excluding deaths from external causes such as drug poisoning and homicide. They treated a death as officially reported as COVID-19 only when ICD-10 code U07.1 appeared as an underlying or contributing cause of death.
Between March 2020 and December 2021, researchers estimated that 155,536 COVID-19 deaths were not officially recorded as such, with a 95% uncertainty interval of 150,062 to 161,112. That means the study estimated 19% more COVID-19 deaths than federal records reflected. The gap was especially stark in homes, where COVID-19 deaths were 160% higher than official records reflected, suggesting 111,245 unrecognized COVID-19 deaths occurred in residences.
The study highlights a distinct demographic and geographic bias in these “hidden” deaths. Racial inequities were evident, as undercounts were significantly higher among Black, Hispanic, Asian, and American Indian/Alaska Native populations. Geographic bias was also visible in regional disparities uncovered by the study. The South saw the greatest inaccuracies; Alabama, for instance, had 67% more predicted COVID-19 deaths than were officially recorded. Socioeconomics also played a role. Males, individuals without a high school education, and those in lower-income counties were more likely to have COVID-19 deaths go uncounted. Many of the predicted uncounted deaths were attributed to underlying causes such as Alzheimer’s disease and related dementia, cardiovascular disease, and diabetes.
The implications for the medicolegal community are twofold. First, although the study did not directly examine the specific reasons why many death certifications list inaccurate causes of death, other sources point to COVID-era challenges such as inadequate staffing to conduct postmortem COVID-19 testing, a lack of standardized training and protocol for death investigators, and partisan beliefs that may cloud investigators’ judgment, particularly for county coroners who are not required to have medical backgrounds. For the criminal justice system, the lesson is direct. Official death records are not neutral simply because they are official. When government death-investigation systems are politically influenced or medically untrained, the resulting data can obscure the deaths of the very communities most dependent on accurate public accountability.
The study also cautions that its machine-learning approach rests on important assumptions. The model treated inpatient hospital COVID-19 death certifications as reliable training data because hospital testing was far more consistent, then applied those patterns to out-of-hospital deaths. The researchers acknowledged that the assumption that an inpatient-trained model can be applied to out-of-hospital deaths is empirically untestable and represents the largest source of potential bias.
Second, the study suggests that machine learning may be useful for identifying gaps in cause-of-death data, although the researchers cautioned that algorithms cannot replace broader systemic reform. They note that similar methods could potentially be adapted to other high-stakes areas where cause-of-death data is incomplete or suspected to be biased, specifically drug overdoses and deaths in police custody.
For investigators, this means official cause-of-death data may increasingly be subject to digital audits that can reveal incomplete or biased reporting. For prisoners, detainees, families of people who die in police custody, and communities historically underserved by government institutions, the study reinforces a familiar warning that the official record may be the beginning of the inquiry, not the end of it. To address these gaps, the researchers recommend increased funding and training for death investigators, as well as increased hiring of medical examiners.
Source: Forensic Mag; Science Advances.
As a digital subscriber to Criminal Legal News, you can access full text and downloads for this and other premium content.
Already a subscriber? Login
