Identifying barriers to being offered and accepting a telehealth visit for cancer care: Unpacking the multi-levels of documented racial disparities in telehealth use.

382 Background: The use of telehealth visits in cancer care can improve accessibility, timeliness, and convenience for patients. However, disparities in its use persist with few studies having evaluated both patient- and community-level factors in relation to telehealth visit use in cancer care. Methods: We assembled a cohort of adults with an upcoming healthcare visit related to their cancer treatment at two academic medical centers (one in central NC and one in southeast MI) and their community affiliates. Black adults and those with a scheduled telehealth visit were oversampled during recruitment. Pre-visit surveys administered by telephone (N=773, RR=15%) collected information on sociodemographic characteristics, use of telehealth visits, healthcare access concerns, and patient symptoms/visit goals. We linked participant residential address to community-level measures, including broadband access. The two patient-reported outcomes of interest were 1) whether a choice in visit type (virtual or in-person) was offered and 2) scheduled visit type among those offered a choice. Associations of each outcome with sociodemographic characteristics and barriers (individual- and census tract-level) were assessed through estimation of unadjusted and adjusted prevalence differences (PDs). Results: 773 adults (RR=15%) were included in the cohort, mostly from the central NC site (N=621, RR=23%). After excluding non-recall for being offered a choice, the analytic sample (N=725) was 46% aged <65 years, 42% Black, and 67% women. Black participants were less likely than non-Black participants to be offered a choice (15% vs 23%, PD= -8.7 (-14.4, -3.0), and if offered a choice, less likely to accept a telehealth visit: 20% vs 67%, PD= -47.0 (-62.0, -32.0). Similarly, women were less likely to be offered a choice (16% vs 27%, PD= -10.9 (-17.4, -4.4)) and accept a telehealth visit 42% vs 63%, PD= -20.8 (-36.8, -4.7). Those who expressed technology-related worries were less likely to schedule a telehealth visit, such as those who were worried about internet access: 33% vs 57%, PD adj = -11.9 (-32.9, 9.2). Participants who indicated their insurance did not cover telehealth visits were less likely to be offered a choice (10% vs 24%, PD ad = -9.4 (-23.1, 4.4). Lower community technology access (e.g. broadband ownership) and higher area-level poverty were associated with both outcomes, although these estimates were attenuated after adjustment. Conclusions: Not being offered a choice in visit type and/or not accepting a telehealth visit were associated with individual- and community-level barriers as well as demographic characteristics which may reflect additional unmeasured structural inequities (namely racism and sexism). Interventions to improve access to telehealth visits in cancer care should consider patient- and community-level barriers to care.