O1018: Racial/ethnic and Language Disparities in Electronic Health Record-embedded Behavioral Health Flag Use in a Pediatric Emergency Department

Background: There are racial disparities in the management of youth presenting with mental health concerns (MHC) to the emergency department (ED), with minoritized youth experiencing higher rates of chemical and physical restraints. Behavioral health flags (BHF) in the electronic health record (EHR) alert providers of the potential risk of harmful or violent encounters. Racial and ethnic disparities in BHF use have been demonstrated in the adult population, but little is known about their use in the pediatric population. The objective of this study was to investigate the presence of racial, ethnic, and language disparities in the use of BHFs among youth presenting to the ED with MHC.

Methods: Cross-sectional study of visits from January 2020 through December 2022 with an ICD code related to MHC (F90-F99) in an urban pediatric ED. We extracted patient and visit level data from the EHR. Race and ethnicity were categorized as non-Hispanic (NH) white, NH Black, Hispanic and NH other, and preferred language was categorized as English, Spanish and Other. Multivariable logistic regression used to examine racial, ethnic and language preference differences and the presence of BHFs. Models were adjusted for age, sex, and insurance. In a subpopulation of patients with the BHF, we examined the differences in being flagged as high risk of aggression vs moderate and low risk.

Results: There were 3074 visits representing 2369 unique patients with MHC. The majority of patients were male (56.3%), NH Black (58.5%), and publicly insured (63.8%) with a mean age of 13 (+/- 3.9) years (Table 1). BHFs were present in 41.8% of the visits. NH Black patients had higher odds of having a BHF compared to NH whites (47.4% vs 33.2%; aOR:1.7; [95% CI 1.4, 2.2]). Compared to patients who preferred English (43.5%), patients who preferred Spanish (30.1%; aOR: 0.5; [95% CI 0.4, 0.7]) or another language (30.0%; aOR: 0.6 [95% CI 0.4, 0.8]) had lower odds of having a BHF. In patients with BHFs (n=1286), the majority were marked as low risk of aggression (55.1%), followed by high risk (26.2%), and moderate risk (18.7%.) NH Black patients had higher odds of being flagged as high risk than NH white patients (29.8% vs 17.2%; aOR: 1.8; [95% CI 1.2, 2.8]). Patients who preferred a language other than English had higher odds of being flagged as high risk of aggression when compared to patients who preferred English (41.2% vs 26.1%; aOR: 2.0; [95% CI 1.1, 3.5]) (Table 2).

Conclusion: These findings suggest that disparities exist in the use of BHFs, with minoritized youth more likely to be affected. Multi-center studies are needed to confirm these results and inform interventions to standardize the use of BHFs as well as create institutional changes to address structural racism in mental health care.