Demographic and socioeconomic disparities in the hybrid ophthalmology telemedicine model.

IMPORTANCE: As telemedicine use expands, it is important to evaluate demographic and socioeconomic disparities among patients receiving ophthalmic care through new hybrid telemedicine models.

OBJECTIVE: To evaluate whether there are demographic and socioeconomic disparities in the delivery of the hybrid telemedicine model.

DESIGN: Retrospective, cross-sectional, case-control analysis of patient encounters from April to December 2020.

SETTING: A single, academic, hospital-based eye clinic in Boston, Massachusetts.

METHODS: Electronic medical records of all patient encounters from April to December 2020 were reviewed and categorized into hybrid, virtual-only, and standard in-person visits. Patient-level data for all visits were extracted including age, sex, race/ethnicity, primary language, Area Deprivation Index (ADI), insurance type, and marital status. Visit-level data for all hybrid visits were also extracted from the medical record including the visit dates and patient adherence. Demographics for the cohort of patients with at least one no-show visit were compared with demographics for the cohort of patients who only had completed visits. The primary study outcomes were the differences in demographic characteristics between the hybrid visit show and no-show groups. The secondary outcomes included demographic characteristics of patients who did not complete their hybrid visit versus a time-matched cohort of patients who did not complete their standard in-person visit. Continuous variables for patient characteristics were compared with independent samples t -tests and categorical variables were compared using Pearson chi-square tests. Multivariate logistic regression was used to examine the differences between the cohorts. Variables with missing values other than suppressed ADI values were imputed using multiple imputations by chained equations.

RESULTS: Of a total of 1025 patients who were scheduled for a hybrid visit, 145 (14.1%) patients failed to complete their visit. Primary language and insurance were found to be statistically different between patients who completed and did not complete their hybrid visits. More English speakers and fewer Haitian Creole speakers completed their hybrid visits ( p  = 0.007) while more patients with private insurance and fewer patients with Medicaid completed their hybrid telemedicine visits ( p  = 0.026). No associations were found between hybrid telemedicine visit adherence and age, sex, race/ethnicity, marital status, or ADI. When the 145 patients who failed to complete their hybrid visits were compared to a time-matched cohort of patients who failed to complete their standard in-person visit, we found that patients who missed hybrid visits were similar to those who missed standard in-person visits except for patients insured by Medicare. These patients were more likely to miss a hybrid visit than a standard in-person visit (Odds Ratio 2.199, 95% confidence interval 1.136-4.259, p  = 0.019). No associations were found between patient nonadherence with hybrid telemedicine versus with standard in-person visits based on age, sex, primary language, race/ethnicity, marital status, or ADI.

CONCLUSION: The hybrid telemedicine model was associated with insurance and language-based disparities. Patients with non-English primary language and Medicaid recipients were more likely to miss a hybrid visit than their counterparts. Our findings support developing deliberate interventions to ensure hybrid telemedicine care is delivered equitably to all patients.