Abstract
BACKGROUND: This study examined disparities in telehealth utilization across various health care services among adults in California, with particular focus on racial, ethnic, and linguistic differences. While telehealth has emerged as a potential tool for addressing health care disparities, evidence suggests that utilization patterns may exacerbate existing inequities. METHODS: Using the 2021-2022 California Health Interview Survey (n = 45,893) conducted in diverse languages, we employed descriptive statistics and nonlinear Blinder-Oaxaca decomposition to analyze disparities in telehealth utilization. The Blinder-Oaxaca method systematically partitions outcome differences between groups into explained components and unexplained components. We stratified analyses by visit purpose: mental health, primary care, acute care, chronic care, orthopedics, dermatology, and ophthalmology. RESULTS: Compared with non-Hispanic Whites, African Americans used telehealth 4.1% more (p < 0.01), whereas Hispanics/Latinos and Asians used it 5.9% (p < 0.01) and 11.3% (p < 0.01) less, respectively. For Hispanics/Latinos, approximately 14% of the disparity remained unexplained by measured covariates, while for Asians, 92% was unexplained (p < 0.01). Individuals with limited English proficiency (LEP) were 9.6% (p < 0.01) less likely to use telehealth overall, with 38% of this gap unexplained by measured factors (p < 0.01). The disparity was most pronounced in mental health services, where LEP individuals utilized telehealth 4.9% points less than non-LEP individuals (p < 0.01), with this gap entirely unexplained by sociodemographic factors. Subgroup analysis revealed telehealth utilization disparities between LEP and non-LEP individuals across all language groups, with LEP Spanish speakers using telehealth 4.3% points less than non-LEP Spanish speakers (p < 0.01), LEP Asian language speakers using it 7.3% points less than non-LEP Asian language speakers (p < 0.01), and LEP speakers of other languages using it 7.3% points less than their non-LEP counterparts (p < 0.01). DISCUSSION: Our findings reveal significant disparities in telehealth utilization associated with race/ethnicity and English proficiency levels, with linguistic barriers most evident in mental health services. The substantial unexplained components in our decomposition analyses suggest that cultural, structural, and linguistic factors beyond socioeconomic factors significantly influence telehealth utilization. These results underscore the need for efforts to develop linguistically appropriate telehealth services, particularly in mental health care. Addressing these barriers is crucial for harnessing telehealth's potential to promote health equity rather than exacerbating existing health disparities in an increasingly digital health care environment.