Strangers rarely try to help in medical emergencies, and assistance may be especially rare when patients happen to be black, a U.S. study suggests.

Researchers analyzed data on more than 22,000 patients aided by emergency medical services (EMS) and found bystanders on public streets only intervened about 3 percent of the time. Black patients, however, were 58 percent less likely to get help from strangers at the scene than white people, the study found.

“We don’t know whether there were bystanders present at these medical emergencies – bystanders there who didn’t help – or whether there weren’t any people around,” said study co-author Erin York Cornwell, a sociology researcher at Cornell University in Ithaca, New York.

“However, we observe low rates of bystander support even in very densely populated areas, which suggests that the low rates of bystander support also reflect situations where bystanders are present but don’t directly engage with the patient,” York Cornwell added by email.

Each year, people in the U.S. make approximately 36 million calls to EMS, York Cornwell and co-author Alex Currit note in the American Journal of Public Health.

Bystanders often arrive at the scene before EMS. Even without formal medical training, bystanders can significantly increase patients’ survival odds when, for example, they perform cardiopulmonary resuscitation (CPR) during cardiac arrest, put pressure on wounds to slow severe bleeding or do the Heimlich maneuver for choking.

But when researchers analyzed 2011 data from the National Emergency Medical Services Information System, they found that bystanders rarely got involved. Patients in the poorest counties across the country were least likely to receive help from strangers, the study found.

Among both blacks and whites, patients in counties with low population densities were most likely to receive help from bystanders. But within areas with the same density, whites were more likely to receive help than blacks.

Segregation may play a role, York Cornwell said.

Previous research has found that black patients suffering cardiac arrest are less likely to get CPR from a bystander than white patients. Often, this is attributed to a racial gap in CPR training, with black people less likely to live in communities where many residents get CPR or first aid training and more likely to feel unqualified to help in an emergency.

“But I think this goes beyond training, because we’re looking at such a huge range of conditions,” York Cornwell said. “Some of the most common forms of helping that we see are simple things like providing a blanket or water, putting pressure on a wound, stabilizing a patient, or helping with medications like aspirin.”

The study also didn’t address the race of the bystanders, and prejudice may influence people’s decisions to get involved, said Dr. Rachael Sharman, a psychology researcher at the University of the Sunshine Coast in Queensland, Australia who wasn’t involved in the study.

When people see an emergency on the street, they may consider their own safety before they decide to intervene, Sharman added by email. “There exists a stereotype that a black person may represent a greater threat,” Sharman said.

Because people decide whether to help based on what may be a biased perception about safety, addressing the lack of bystander intervention is a hard problem to fix, Sharman noted.

“Bigger cities, with greater social problems will always see a low level of bystander intervention,” Sharman said. “Fixing racial disparities comes back to integration, where race eventually becomes irrelevant in how a person is perceived.”