Stigmatizing Language in Medical Records Affects Future Treatment
MAY 10, 2018
Cecilia Pessoa Gingerich
Anna Goddu, MSA study from Johns Hopkins University School of Medicine found that the presence of stigmatizing language in a patient’s medical records affected a physician’s clinical decision-making later on.
Researchers found that reading a vignette about a hypothetical patient with sickle cell disease that contained stigmatizing language affected how study participants treated that patient’s pain, as compared to participants who read a neutral vignette of the same patient.
Mary Catherine Beach, MD, MPH, an associate professor at Johns Hopkins University School of Medicine, Baltimore, designed the study to investigate whether physicians perpetuate bias against certain patients through language in their medical records.
"This record may be the only source of information a new clinician has about some patients," said Beach. "We have to question the assumption that the medical record always represents an objective space."
Medical students and residents (n=413) at an urban, academic medical center participated in the study and were randomly assigned to read a vignette with either stigmatizing or neutral language. Researchers wrote the vignette with stigmatizing language based on text from records at the same medical center.
The stigmatizing language included specifying that the patient spent time with friends outside a McDonald’s and saying “narcotic” instead of “opioid.” Additionally, the stigmatizing chart said, "he is narcotic dependent and in our ED frequently," while the neutral version said, "he has about 8-10 pain crises a year, for which he typically requires opioid pain medication in the ED."
After reading the first part of the patient chart, residents were asked to choose between a few options of pain medication dosing. They then read an update on the patient’s status and were given the chance to adjust the pain management plan. The choices were combined to yield a score from 2 to 7, with higher scores indicating more aggressive pain management.
Residents who were influenced by the biased chart treated the patient’s pain less aggressively than those who read the neutral chart (4.7 stigmatizing language vs. 5.3 neutral language, p < 0.001).
"There is growing evidence that the language used to communicate in health care reflects and influences clinician attitudes toward their patients," said Anna Goddu, MS, a Johns Hopkins University School of Medicine student who co-authored the study. "Medical records are an important and overlooked pathway by which bias may be propagated from one clinician to another, further entrenching health care disparities."
All study participants completed an abbreviated Positive Attitudes towards Sickle Cell Patients Scale (PASS) that included the 7 PASS items applicable to this situation with a hypothetical patient. Scores varied from 7 to 35, with higher scores representing more positive attitudes.
Resident’s and medical student’s PASS scores showed that participants exposed to the biased language viewed the hypothetical patient with more negative attitudes (20.3 stigmatizing language vs. 25.1 neutral language, p < 0.001).
This effect was evident among both groups, but medical students (21.6 stigmatizing vs. 26.6 neutral, p < 0.001) saw the patient with more positive attitudes than residents (18.7 stigmatizing vs. 23.0 neutral, p < 0.001).
"Attitudes seem to become more negative as trainees progress," Beach said. "It may be that trainees are influenced by negative attitudes and behaviors among their peers and seniors in the clinical setting."
The authors urged clinicians to vigilantly guard against including stigmatizing language in patient charts and to be aware of bias as the read notes written by others.
"I hope our study makes clinicians think twice before including certain, nonessential points about a patient's history or demeanor in the medical record," Goddu said.
The study, “Do Words Matter? Stigmatizing Language and the Transmission of Bias in the Medical Record,” was published in The Journal of General Internal Medicine.