VIRTUALLY DECIDED: A MIXED-METHODS EXPLORATION OF PATIENT, PROVIDER, AND SYSTEM-LEVEL FACTORS INFLUENCING VA CARDIOLOGISTS' PROVISION OF TELEMEDICINE

ABSTRACT

BACKGROUND: Virtual cardiology care, defined as care delivered by phone or video, expanded rapidly in the Veterans Health Administration (VA) at the onset of the COVID-19 pandemic and remains a significant proportion of all VA cardiology care. However, factors influencing whether a visit is conducted virtually are poorly understood. METHODS: In this mixed-methods study, we first analyzed a nationwide cohort of Veterans who had a cardiology visit before COVID-19 (1/1/2019-3/ 10/2020), some of whom had follow-up visits before COVID and others afterwards (3/10-2020-3/10/2021). We assessed the hazard of receiving cardiology-related video care and any virtual care with a survival model adjusted for baseline patient sociodemographic and clinical characteristics;we performed analyses with and without adjustment for geographic region via Veterans Integrated Service Network location (VISN). Then, we conducted qualitative interviews with VA cardiologists to further characterize the variation identified in the hazard of video and virtual care utilization. RESULTS: We analyzed 416,621 Veterans;average patient age was 69.1 years and 5.0% were female. Older, low-income, and rural-dwelling Veterans had a lower hazard (i.e. time to event) of using video care (adjusted hazard ratio for ages 75 and older 0.80, 95% CI 0.75-0.86;for low-income status 0.94, 95% CI 0.89-0.98;for highly rural residents 0.77, 95% CI 0.68-0.87). The hazard ratios for a video-based encounter varied across geographic regions, with adjusted hazard ratios for use of video care as low as 0.06 (95% CI 0.04-0.07) compared to the reference region with highest use of video care. In our qualitative assessment, cardiologists (N=7) suggested patient, provider, and system-level factors influencing visit modality. At the patient level, clinicians perceived that older, lower-income, and rural-dwelling Veterans had more difficulty accessing video technology, but also benefited disproportionately from virtual care from the convenience of avoiding travel to a VA facility. At the provider level, clinicians preferred virtual care for routine follow-up visits and visits for conditions when most pertinent information could be collected from history (e.g. stable coronary artery disease). At the system level, clinicians noted explicit and implicit nudges toward certain modalities, such as differential productivity accounting (e.g. video visits counting as more productivity units than phone visits) and praise for high video care users, and differed in their perception of whether the system or clinician primarily drove choice of visit modality. CONCLUSIONS: Likelihood and timing of virtual cardiology care varies across VA patients and sites due to patient, clinician, and system factors. VA cardiologists perceive variability in the degree to which autonomy over visit modality choice lies with providers versus the system. Policies intended to alter visit modality mix should consider these types of influences as well as varying autonomy in modality choice.