Background: Stigmatizing language (SL) is common in the medical record and can perpetuate bias, which has downstream effects on clinical decision-making and patient care and promotes healthcare inequity1-3. Internal medicine residents are often front-line clinicians for composing documentation and residency faculty are tasked with providing feedback and training on best documentation practices. However, the extent to which residents and faculty are prepared to address SL in documentation is unknown. We sought to understand resident, advanced practice providers (APPs), and faculty experiences with, recognition of, and attitudes around SL in clinical documentation (CD) at two large academic medical centers (AMCs).
Methods: We developed a survey using best practices of survey design including pretesting with cognitive interviewing. We distributed the survey electronically via REDCap to residents, APPs and faculty at these centers. Participants were recruited via email invitations, with reminders sent every one to two weeks. At one of the two AMCs, respondents were provided with a $15 gift card after completion of the survey. The survey included five knowledge questions with statements containing a total of 11 evidence-based SL terms and questions around respondent experiences and attitudes connected to SL.
Results: Sixty three total residents responded (21% response rate) and eighty-one Attendings or APPs responded (52% response rate). Analysis revealed variability in recognition of SL across the eleven terms. 3% of residents and 4% of faculty/APPs correctly identified all 11 terms, whereas 68% of residents v 72% of attendings selected at least 6 or more correct choices. The majority of residents and faculty/APPs correctly identified “non-compliant” (95 vs 81), “claims” (79 % v 90%) and “poorly controlled” (69.% v 75%). However, residents and faculty/APPs were less likely to label as SL the use of the patient’s own words contained in quotation marks. For example, ’horrible’ and ‘only medication that works’ (34% vs 45% and 46% vs 45%). The majority of residents (92%) and faculty/APPs (89%) indicated that they notice the use of SL when reading clinical documentation. Few reported feeling very or extremely confident (24% v 10%) that their CD consistently avoids SL. However, faculty infrequently provide feedback to residents (17%) citing time (66%) and lack of prior training (29%) as the most common reasons. The majority of faculty (70%) felt that is important for trainees to avoid the use of SL in CD.
Conclusions: Residents are in the formative stages of their professional identity formation and the documentation habits they currently employ will be carried into the future. Our study showed variability in resident and Faculty/APP ability to recognize SL in patient descriptors. The majority of Faculty/APPs have not received previous training on SL (57.3%). To provide a more equitable healthcare landscape for all patients, targeted educational interventions must be created to help faculty, APPs, and trainees identify and replace SL within their own documentation and when coaching their learners.