Behavioral Health Flag Use by Race and Ethnicity in a Pediatric Emergency Department.

Importance: Behavioral health flags are intended to serve as safety alerts for staff by identifying patients at risk of aggression but may also be prone to racial bias, leading to inequities in care.

Objective: To investigate racial and ethnic differences in the use of an aggression risk evaluation tool and being labeled as high aggression risk in the electronic health record among youths presenting to a pediatric emergency department.

Design, setting, and participants: This cross-sectional study of electronic health record data was conducted at a single-center pediatric emergency department. Visits from patients aged 21 years or younger presenting with mental health-related concerns between January 2020 and December 2022 were included.

Exposure: Patient race and ethnicity (Hispanic, non-Hispanic Black [hereafter, Black], non-Hispanic White [hereafter, White], or other, which includes individuals who identify as American Indian or Alaska Native, Asian, multiple races, or any other race).

Main outcomes and measures: The primary outcomes were (1) the use of an aggression risk evaluation tool by a psychiatric social worker, (2) being labeled with a high aggression risk behavioral health flag, or (3) being labeled with a high aggression risk behavioral health flag with a history of violent behavior documented with no other aggression risk behaviors documented. Separate logistic regression models were used to examine association of race and ethnicity with each outcome.

Results: There were 5121 visits for mental health concerns. The majority of patients were female (3198 patients [62.5%]) and publicly insured (3161 patients [61.8%), with a mean (SD) age of 13.8 (2.7) years. With regard to race and ethnicity, 3061 patients (59.8%) were Black, 893 patients (17.4%) were Hispanic, 778 (15.2%) were White, and 389 (7.6%) belonged to other racial and ethnic groups. Most visits (4119 visits [80.4%]) had an aggression risk evaluation tool completed, and 627 (15.2%) were labeled with a high aggression risk behavioral health flag. There were no racial and ethnic differences in use of the aggression risk evaluation tool. Compared with White youths, Black youths were more likely to be labeled as high aggression risk (60 White youths [9.8%] vs 486 Black youths [19.8%]; adjusted odds ratio, 1.71; 95% CI 1.24-2.35). Among patients who had a history of violent behavior and no other aggression risk behaviors documented on the evaluation tool, Black patients had higher odds of being labeled as high aggression risk compared with White patients (72 Black patients [32.4%] vs 3 White patients [10.7%]; odds ratio, 4.00; 95% CI, 1.16-13.69).

Conclusions and relevance: In this cross-sectional study, Black youths were more likely to be labeled with a high aggression risk behavioral health flag despite similar responses to an aggression risk evaluation tool. Processes to ensure equitable care delivery are critically needed.