Stigmatizing and Positive Language in Birth Clinical Notes Associated With Race and Ethnicity.

Importance: Language used in clinical documentation can reflect biases, potentially contributing to health disparities. Understanding associations between patient race and ethnicity and documentation of stigmatizing and positive language in clinical notes is crucial for addressing health disparities and improving patient care. Objective: To examine associations of race and ethnicity with stigmatizing and positive language documentation in clinical notes from hospital birth admission. Design, Setting, and Participants: This cross-sectional study included birthing patients at 2 metropolitan hospitals in the Northeastern US between 2017 and 2019. Eligible participants were admitted for labor and birth and had at least 1 free-text clinical note. Analysis was conducted using natural language processing. Data were analyzed between March and December 2024. Exposures: Patient race and ethnicity, categorized into mutually exclusive groups of Asian or Pacific Islander, Black, Hispanic, and White. Main Outcome and Measures: Presence of 4 stigmatizing language categories (marginalized language or identities, difficult patient, unilateral or authoritarian decisions, and questioning patient credibility) and 2 positive language categories (preferred and/or autonomy, power and/or privilege). Results: Among the 18 646 patients included in the study (mean [SD] age, 30.5 [6.2] years), 2121 were Black (11.4%), 11 078 were Hispanic (59.4%), and 4270 were White (22.9%). The majority (10 559 patients [56.6%]) were insured by Medicaid. Compared with White patients, Black patients had higher odds of having any stigmatizing language (model 2: odds ratio [OR], 1.25; 95% CI, 1.05-1.49; P < .001), after adjustment for demographic characteristics. Black patients also had higher odds of any positive language documented (model 2: OR, 1.18; 95% CI, 1.05-1.32; P = .006). Hispanic patients had lower odds of documented positive language (model 2: OR, 0.90; 95% CI, 0.82-0.99; P = .03). Asian or Pacific Islander patients had lower odds of language documented in the power and/or privilege category (model 2: OR, 0.71; 95% CI, 0.57-0.88; P = .002). Conclusions and Relevance: In this cross-sectional study examining clinical notes of 18 646 patients admitted for labor and birth, there were notable disparities in how stigmatizing and positive language was documented across racial and ethnic groups. This underscores the necessity for improving documentation and communication practices to reduce the use of stigmatizing language.