Session: Health Equity/Social Determinants of Health 9
190 - Racial and Ethnic Differences in the Use of Stigmatizing Language in the Pediatric Emergency Department
Monday, April 28, 2025
7:00am – 9:15am HST
Publication Number: 190.6389
Meleah Boyle, Children's National Health System, Rockville, MD, United States; Muhammad M. Rahman, Children's National Hospital, Silver Spring, MD, United States; Gia Badolato, Children's National Health System, Washington, DC, United States; Monika Goyal, Children's National Health System, Washington, DC, United States
Staff Scientist Children's National Health System Rockville, Maryland, United States
Background: Racial and ethnic inequities in the provision of emergency department (ED) care have been documented across a wide range of medical conditions, including mental health. Clinician bias may be one potential source of such inequities. Electronic heath record (EHR) notes capture clinicians' observations and interactions with patients and the words they use may provide insight into potential biases. Objective: To examine racial and ethnic differences in the use of positive and negative terms within EHR notes in the pediatric ED. Design/Methods: This cross-sectional study included patients who presented with a mental health related complaint and were treated and evaluated by a psychiatry social worker in the pediatric ED from January 1, 2019 to August 31, 2023. Patient demographics, visit characteristics, and clinical notes were extracted from the EHR. Descriptive terms used in the clinical notes were identified using natural language processing techniques, such as topic modeling. Sentiment analysis was conducted on these terms, followed by contextual refinement and additional terms were added based on a literature review. This analysis included 14 positive terms and 10 negative terms. Descriptive statistics and logistic regression were used to evaluate use of terms by race and ethnicity. Results: There were a total of 203,397 clinical notes across 7,276 ED visits and 5,228 unique patients. Patients had an average age of 13.7 (SD: 2.8) years, were majority female (61.7%), and 55.8% were non-Hispanic (NH) Black, 18.7% were Hispanic, and 16.1% were NH White. Use of positive and negative terms varied by race and ethnicity. Compared to NH White patients, NH Black patients had lower odds of any positive term, and the terms believed, compliant, engaged, interacted well, pleasant, shared, and using coping skills used in their notes; while Hispanic patients had lower odds of calm, interacted well, and pleasant. NH Black patients had higher odds of having several negative terms, including aggressive, combative, destroyed, non-compliant, refused, and reportedly used in their notes, while Hispanic patients had lower odds of aggressive, combative, destroyed, and ignoring compared to NH White patients. (Table 1)
Conclusion(s): Overall, positive terms were used less often in notes for NH Black and Hispanic patients, while negative terms were used more often in notes for NH Black patients compared to NH white patients. This indicates the presence of potential for bias in EHR notes. Future research should examine whether use of stigmatizing language is associated with care outcomes.
Table 1. Use of Positive and Negative Terms by Race and Ethnicity