The language used to describe patients in electronic health records (EHR) may be perpetuating racial bias and other negative stereotypes in health care, according to a study recently published in 

 A multi- disciplinary team of investigators from the University of Chicago used machine learning techniques to analyze potentially stigmatizing language across more than 40,000 history and physical notes in EHRs made up of 18,459 adult patients and 33,142 unique encounters at an academic medical center in Chicago, Illinois, between January 2019 and October 2020. They determined Black patients were 2.54 times more likely than White patients to be described in negative terms (95% CI, 1.99-3.24). Negative terms included refused (n = 1461), (not) adherent (n = 605), (not) compliant (n = 561), and agitated (n = 409). Findings were adjusted for sociodemographic and health characteristics.

Moreover, the differences were not solely based on race. For example, Medicare and Medicaid beneficiaries were more likely to have negative descriptors applied to them than were patients with commercial- or employer-based insurance (Medicare: adjusted odds ratio [AOR], 2.08; 95% CI, 1.57-2.75) (Medicaid: AOR, 2.66; 95% CI, 2.08-3.40), as were unmarried patients (AOR, 2.12; 95% CI, 1.70-2.65) compared with patients who were married.

Evaluable patients were patients with at least 1 charted history or physical note in their EHR when they were admitted to an inpatient setting or emergency department or visited an outpatient setting. These notes are designed to document the patient’s reason for seeking medical care; provide a summary of medical, family, and social history; and lay out the prescribed care plan to address the medical problem. The investigators chose to center their research on the history and physical note because it is designed to create a comprehensive narrative of the patient’s experience, which will be used by other health care professionals. If a patient had multiple history and physical notes, all available notes were extracted and included for analysis.

The findings are especially troubling, given that other studies have found that less than 20% of text in inpatient progress notes are original,

 with most of the rest being imported from prior documentation. As a result, “subsequent providers may read, be affected by, and perpetuate the negative descriptors, reinforcing stigma to other health care teams,” the authors wrote. “Negative descriptors written in the admission history and physical may be likely to
be copied into subsequent notes, recommunicating and amplifying potential biases. This practice underscores the responsibility of providers who document the initial patient encounter to do so in an aware and sensitive manner.”

Furthermore, the investigators found that notes written for outpatient visits were less likely to have negative descriptors included in the patient’s EHR than inpatient encoun- ters (31% vs 68%). They theorize that this may result from inpatient settings being inherently more stressful, thereby increasing the risk of doctors using stereotypes as “cognitive shortcuts.” They also found that race-based differences in the use of negative descriptors narrowed after March 1, 2020.

Nevertheless, the authors agreed that fixing the problem will require medical 

institutions to better address the issue of implicit racial bias among providers. They cite the example of 1 physician’s use of the term “aggressive” as reflecting the physi- cian’s personal bias regarding Black men. Once such a negative label becomes part of the patient’s record, “it potentially affects the perceptions and decisions of future providers, regardless of whether future providers hold a bias about Black men being aggressive,” they wrote.

Regulatory bodies, such as the Accreditation Council for Graduate Medical Education, might also develop specific recommendations for the use of nonstigmatizing, patient-centered language to prevent the transmission of bias, they suggest. The study concludes that the need is especially important in light of the OpenNotes policies that medical institutions are adopting, which allow patients full access to their EHRs, including chart notes. Without paying attention to the language used to describe patients, providers risk “harming the patient-provider relationship, with downstream effects on patient satisfaction, trust, and even potential litigation.”

Sun M, Oliwa T, Peek ME, Tung EL. Negative patient descriptors: documenting racial bias in the electronic health record.

. 2022;41(2):203-211. doi:10.1377/hlthaff.2021.01423

Wang MD, Khanna R, Najafi N. Characterizing the source of text in electronic health record progress notes. 

 2017;177(8):1212-1213. doi:10.1001/jamainternmed.2017.1548

Articles

Fixing the problem will involve holding medical institutions accountable for addressing implicit bias among providers.

Jeffrey Bendix

Articles

Electronic Health Record Notes May Be Perpetuating Biases