Charles Kinsey was trying to calm down an autistic patient who had run away from his North Miami assisted living facility when police arrived. A bystander filmed their interactions on July 18, before Kinsey was shot. (Courtesy Hilton Napoleon)

Black and Hispanic people are more likely to be injured by police than whites are because police are more likely to stop them, according to a new study.

The research, published Monday by the journal Injury Prevention, did not, however, find that police were more likely to injure blacks and Hispanics than whites after they were stopped.

The study contributes to a continuing debate among researchers about racial bias in the use of force by police. Other recent research suggests that police are more likely to push, shove or handcuff black and Hispanic civilians after stopping them, or to use a baton or pepper spray on them.

The new study used a federal sample of hospital records from 2011 and 2012. It is one of the first studies to comprehensively examine not only the lethal use of force but also injuries inflicted by police that do not result in death, which are far more common than fatalities.

Last week, for example, an officer shot and wounded an unarmed man in North Miami, Fla., as he lay on his back in the street with his hands in the air. The man, a behavioral therapist named Charles Kinsey, is black, and many observers saw evidence of police bias in the videos of the incident that circulated online.

Injuries inflicted by police accounted for 3.3 percent of all injuries resulting from an attack by another person, and the deaths by police accounted for about 6.5 percent of all homicides, according to one of the authors of the new study.

Roughly 55,400 people were injured or killed in interactions with police officers in 2012. About 32 out of every 10,000 stops or arrests resulted in a visit to the emergency room. Another 1.7 encounters per 10,000 resulted in a hospitalization, and an additional 0.7 incidents ended in death. The data on police killings came from the Guardian and The Washington Post.

The rates of hospitalization and death per stop and arrest did not vary by race. However, police were twice as likely to stop or arrest Latinos compared to whites and three times as likely to stop or arrest African Americans, so members of those groups were also more likely to be seriously hurt or killed.


The racial disparities in stops and arrests could indicate bias on the part of police officers, said Ted Miller, an author of the study and an economist at the health-research organization Pacific Institute for Research and Evaluation, headquartered in Calverton, Md. The disparities also could reflect that police spend more time in places where crime is more common, which are often poorer neighborhoods where more people of color live.

The data did not include cases in which patients neither were admitted to nor went to an emergency room. ​It could be that police allow racial bias to influence decisions that involve less thought, such as whether to push, shove or handcuff a civilian,​ but that officers are more judicious and think carefully before using the the kinds of weapons and hand-to-hand maneuvers that can cause injury.

Miller and his colleagues compared hospital records on incidents involving police to those for cases of assault in general. Injuries resulting from assault were more severe than the injuries that law enforcement inflicted, and victims of assault were more likely to be admitted to the hospital.

"For the most part, police who have to inflict an injury on somebody are trying to inflict as little injury as they can while controlling the situation and protecting their own safety," Miller said.

The exceptions were the comparatively rare cases of firearm violence. Forty percent of gunshot wounds inflicted by law enforcement were fatal, compared to just 26 percent of gunshot wounds in general.

In unpublished research, Miller and his colleagues also compared their data from the United States to data on police interactions in Australia.

The number of injuries per arrest was similar in both countries, suggesting that police in Australia do not interact with people who are inherently less violent. Yet Australian police killed almost no one. Miller attributed the difference to Australia's strict gun-control laws, which allow the police to avoid using lethal force while still protecting themselves and the public from harm.

"That makes a difference in the ability of those situations to escalate to the point of being deadly," he said.