Like rest of society, doctors implicitly favor whites over blacks

"This supports speculation that subtle race bias may affect health care, but does not imply that it will," said University of Washington researcher Janice Sabin, who presented the study in San Diego Tuesday (Oct. 28) at the American Public Health Association's annual meeting. "This research is too preliminary to know if there is a direct relationship between physicians' implicit, or unconscious, racial attitudes and the quality of medical care."

Sabin, who is an acting assistant professor of medical education and biomedical informatics, said: "Our findings fit with previous research showing bias is common in the general population. But we have to remember people are not racist if they hold an implicit bias."

Data from the study were drawn from a sample of more than 400,000 anonymous people who took the race attitude Implicit Association Test online during a 28-month period between 2004 and 2006. More than 2,500 of those test-takers identified themselves as doctors. The race attitude and a variety of other tests that examine various forms of bias are found on the Web site of Project Implicit (https://implicit.h … du/implicit/ ) operated by the University of Washington, University of Virginia and Harvard University.

The Implicit Association Test captures subconscious, or implicit, bias by asking people to quickly associate positive or negative words with a series of photographs. In the race test the photographs are of black and white faces.

"We don't call what these tests show prejudice. We talk about it as hidden bias or unconscious bias, something that most people are unaware they even possess," said Anthony Greenwald, a UW psychology professor who created the Implicit Association Test and was part of the research team on the new study.

The people who took the test were predominantly American. Overall, 86 percent of participants said they lived in the United States. Of 2,535 physicians taking the test, 76 percent identified themselves as U.S. residents. Test-takers were predominantly white -- 69 percent of the entire sample and 66 percent of the doctors. There were smaller samples of blacks, Asians and Hispanics.

The majority of physicians in all racial and ethnic groups showed an implicit preference for white Americans compared to black Americans except for black physicians, who on average did not favor either group.

The researchers also looked at the factors of gender and other advanced college education and found virtually the same results. Women exhibited slightly lower, but still significant, implicit preference for whites. The story was the same for more than 6,100 people who said they had law degrees and more than 7,900 individuals who said they had doctoral degrees.

"The implicit bias effect among all the test-takers is very strong," said Sabin. "People who report they have a medical education are not different from other people, and this kind of unconscious bias is a common phenomenon.

"Although it may be a common assumption that high levels of education expose people to diverse ideas and ways of not being biased, subtle biased associations about race are nevertheless common. We have come a long way, but people may be unaware of these common unconscious biases. In medical education, we need to include examination of personal attitudes and beliefs in diversity training and continuing medical education to increase self-awareness so that doctors are alert to hidden attitudes and beliefs that may affect care for minority populations."

Other members of the research team are Brian Nosek, an assistant psychology professor at the University of Virginia, and Dr. Frederick Rivara, a UW professor of pediatrics. The research was funded by the Agency for Healthcare Research and Quality, the National Institute of Mental Health and a UW Magnuson Health Scholars Award.

The Implicit Association Test was developed a decade ago to measure the unconscious roots of people's thinking and feeling. Since it was created, more than 6 million people have taken versions of the test that have measured unconscious attitudes about such topics as race, gender, sexual orientation, age and various ethnic groups.

Provided by University of Washington