Racial/Ethnic and Language Disparities in Telepsychiatry during the COVID‐19 Pandemic: Results from an Urban Safety‐Net Hospital System

RESEARCH OBJECTIVE: To estimate racial/ethnic and linguistic disparities in access to and utilization of outpatient mental health care before and after the rapid implementation of telepsychiatry during the COVID‐19 pandemic. STUDY DESIGN: We used electronic health record data from an urban safety‐net healthcare system on appointments, patient's diagnoses, demographics, location of care, and provider type. Data included outpatient mental health service use for the periods before (pre‐period = March–June 2019) and after (post‐period = March–June 2020) the introduction of telepsychiatry. Outcomes consisted of outpatient mental health service use and no‐show visits, which were operationalized as both dichotomous and count variables. We estimated multivariable linear probability regression models (dichotomous outcome) and negative binomial regression models (count outcome). The primary term of interest was an interaction between race/ethnicity (White [referent], Black, Latino, and Asian) and time indicator (1 = post‐period, March–June 2019; and 0 = pre‐period, March–June 2020) that yielded the difference in service use between White patients and racial/ethnic minority patients. In secondary analysis, the term of interest was an interaction between language of service preference (English [referent], Spanish, Portuguese, and Haitian Creole) and time indicator. We used predictive margins methods to estimate predicted probabilities, predicted means, interpret interactions in nonlinear models, and estimate within group differences. All regression models adjusted for a rich set of covariates. POPULATION STUDIED: Analytic sample consisted of adults between the ages of 18–64 that received mental health services from the safety‐net healthcare system under study between March 2019 and June 2020 (n = 254,995). We further classified our population by race/ethnicity (White, Black, Latino, and Asian) and language of service preference (English, Spanish, Portuguese, and Haitian Creole). PRINCIPAL FINDINGS: In adjusted results, between pre‐ and post‐periods, all racial/ethnic and linguistic groups had significant within‐group increases in access to and use of outpatient mental health care (all p < 0.05). Relative to White patients, Asian patients had a significant decrease in having any outpatient mental health visit (−1.1%, p < 0.05) and Latinos had a significant increase in the number of outpatient mental health visits per 1000 patients (+469.7, p < 0.05). Relative to English‐speaking patients, Portuguese speaking patients had a significant increase in any outpatient mental health visits (+0.5%, p < 0.05). Within‐group no‐show rates increased significantly for White and Asian patients, but mean number of no‐show visits per 1000 patients decreased significantly for all racial/ethnic groups (all p < 0.05). CONCLUSIONS: Telepsychiatry provides patients a means to initiate or continue their treatment regimen while socially distancing and quarantining. This mode of mental health care also removes many of the traditional barriers to care (e.g. child care and transportation) experienced by racial/ethnic minority groups. However, disparities still remain and may have been exacerbated by the introduction of new technological barriers to care. IMPLICATIONS FOR POLICY OR PRACTICE: Clinicians and hospital administrators should consider the impact of digital literacy and the digital divide in the continued use of telepsychiatry. The transition to telepsychiatry highlights the need for a robust medical interpreter pool to meet the increasing patient demand.