(Bloomberg) -- A first of its kind study assessing language used in medical records at a Chicago medical center finds health-care providers more often describe Black patients with “negative descriptors,” such as “non-compliant” and “agitated,” compared to White people in the sample.
Black people were 2.54 more likely to have one or more “negative descriptors” in their medical records, even when adjusted for sociodemographic and health characteristics, a study to be published in the February issue of Health Affairs finds.
Doctors traditionally use terms like “non-adherent” and “non-compliant” when patients don’t follow their advice. However, medical providers could use other, more neutral descriptions, such as that a patient “declined” treatment, said Michael Sun, a third-year medical student at the University of Chicago Pritzker School of Medicine and the lead author of the paper.
Black patients have long reported unfair treatment from health-care providers. This study suggests there may be implicit bias among individual providers, such as doctors and nurses and broader bias in health-care systems, the authors write, which “has the potential to stigmatize Black patients and possibly compromise their care.”
The study’s aim is “not to police language,” said Sun, but “a signal to think about the underlying reasons why we’re using these words.” He hopes its findings encourage providers “to describe patients more compassionately in order to provide more equitable care,” he added.
The Covid-19 pandemic has sparked intense scrutiny of health care disparities and possible biases that could contribute to them. Black people were more likely to die of the virus, particularly early on in the pandemic. Skepticism of the health-care industry also kept some from seeking life-saving vaccinations.
The American Medical Association last June published new guidelines addressing systemic racism in medicine, including discrimination, bias, abuse and microaggressions. The National Institutes of Health also last year launched an initiative to examine and end structural racism and racial inequities in biomedical research.
For the new study, researchers used machine learning to analyze the language in 40,113 history and physical notes made by medical providers, such as doctors and nurses, about 18,459 patients. The notes were made at a large urban academic medical center in Chicago between January 1, 2019 and October 1, 2020. More than 60% of patients in the sample were Black, nearly 30% were White, 6.2% were Hispanic or Latino, and 3.5% were characterized as “other.”
It also found that patients with Medicaid or Medicare insurance had higher odds of a negative descriptor compared to those with private or employer insurance. Unmarried patients also had higher adjusted odds of a negative descriptor compared with married ones.
The analysis did not examine whether rates of non-compliance varied by race or insurance type.
Other studies and surveys have shown disparities in medical treatment in the U.S. Between 2005 to 2013, more than 12% of Black respondents reported racial discrimination in health care compared with 2.3% of White people in a study from the Centers for Disease Control and Prevention.
©2022 Bloomberg L.P.