More than half of police killings not officially documented on US death certificates, study finds

In contrast, a database from the London-based Guardian newspaper captured a large majority of these deaths, the study found.

The study is the first to measure undercounting of police-related deaths in nationwide death certificate data and in a news media-based database and provides the most accurate count to date of the number of people killed by police in the U.S. It will be published online October 10, 2017 in PLoS Medicine.

"To effectively address the problem of law enforcement-related deaths, the public needs better data about who is being killed, where, and under what circumstances," said Justin Feldman, doctoral student at Harvard Chan School and lead author of the study. "But we also found that a different approach -- compiling data from media reports -- can help solve this problem."

Because there is no system that reliably tracks all law enforcement-related deaths in the U.S., the researchers compared two incomplete lists. One list was from the U.S. Centers for Disease Control's National Vital Statistics System (NVSS), which is based on state death certificate data. The other list, The Guardian's dataset called "The Counted," draws from news stories and crowdsourced information.

Using a statistical technique that looked at the number of people who appeared on the Guardian list only, the NVSS list only, and both lists, researchers looked at the degree of overlap to estimate that there were 1,166 law-enforcement related deaths in 2015. There were 599 deaths reported in The Guardian only, 36 reported in the NVSS only, 487 reported in both lists. Statistical analysis suggested that an estimated 44 deaths weren't reported in either source.

The NVSS documented 44.9% of the total number of deaths and The Counted documented 93.1%.

Other findings:

• Misclassification rates for police-related deaths topped 60% among several groups: people under age 18, blacks, people killed by something other than a firearm (particularly Tasers, which accounted for 46 deaths), and people killed in low-income counties.
• Undercounting of police-related deaths varied widely across states. For instance, nearly all of the 17 police-related deaths in Oregon were counted, but none of the 36 such deaths in Oklahoma were.

In most cases, it appeared that police-related deaths were not counted because the coroner or medical examiner failed to mention police involvement on the death certificate, according to the authors.

Laws requiring that police report all law-enforcement related deaths to state public health authorities, along with increased use of news media reports by state officials responsible for identifying deaths, could help government agencies more accurately track these incidents, the authors said.

"As with any public health outcome or exposure, the only way to understand the magnitude of the problem, and whether it is getting better or worse, requires that data be uniformly, validly, and reliably obtained throughout the U.S.," said Nancy Krieger, professor of social epidemiology at Harvard Chan School and senior author of the study. "Our results show our country is falling short of accurately monitoring deaths due to law enforcement and work is needed to remedy this problem."

Co-authors included Brent Coull, professor of biostatistics at Harvard Chan School, and Sofia Gruskin, professor of preventive medicine and professor of law at University of Southern California.

Funding for the study came from the Open Society Foundations.