Underestimated COVID-19 Deaths: AI Reveals Hidden Toll

A latest analysis suggests the official U.S. COVID-19 death toll significantly underestimated the true number of lives lost during the pandemic’s first two years, with as many as 155,000 deaths potentially unrecorded. The study, leveraging machine learning, highlights disparities in which deaths were missed, disproportionately affecting marginalized communities.

The Hidden Deaths: How AI Uncovered the Missing Toll

Researchers, led by epidemiologist Mathew Kiang at Stanford University, utilized a machine learning algorithm to identify deaths likely caused by COVID-19 but recorded under different causes. The study, published in Science Advances, analyzed over 5.7 million death certificates from the Centers for Disease Control and Prevention (CDC) between March 2020 and December 2021.

The algorithm was trained on hospital death records, where COVID-19 testing was prevalent, to understand how the disease manifested in administrative language. It then identified similar patterns in death certificates from locations with less testing, such as homes and care facilities, potentially revealing unrecognized COVID-19 deaths. Researchers estimate that approximately 16% of COVID-19 deaths went uncounted during those years.

Disparities in Uncounted Deaths

The analysis revealed that unrecognized COVID-19 deaths were more common among Hispanic individuals and other people of color. These deaths similarly occurred more frequently in the early months of the pandemic and in states in the South and Southwest, including Alabama, Oklahoma, and South Carolina. The study indicates that systemic inequities in healthcare access contributed to this disparity.

“These vulnerable groups are just taking a higher risk at every step, and the accumulation of all of that is this disparity in COVID mortality at the end,” explains Mathew Kiang.

A Fragmented Death Reporting System

The researchers attribute the underreporting to a fragmented and under-resourced death registration system in the United States, which was overwhelmed during the pandemic. The study highlights the need for improvements in death investigation systems to accurately capture the full impact of public health crises.

Steven Woolf, a researcher at Virginia Commonwealth University, emphasizes that these barriers to care continue to disproportionately affect marginalized communities, leading to higher mortality rates even after the peak of the pandemic. “People on the margins continue to die at disproportionate rates because they can’t access care,” he stated.

Implications for Future Pandemic Preparedness

This research underscores the importance of robust and equitable public health infrastructure. Accurate death reporting is crucial for understanding the true burden of disease and informing effective public health interventions. Addressing systemic inequities in healthcare access is essential to prevent similar disparities in future pandemics.