2023 HSR&D/QUERI National Conference

1202 — Use of VA Primary Care and Adherence to Medications During the COVID-19 Pandemic for High-Risk Veterans with Diabetes

Lead/Presenter: Jean Yoon,  Resource Center - HERC
All Authors: Yoon J (HERC), Chen, C (Center for Innovation to Implementation, Palo Alto) Wong, E (HERC, Palo Alto) Rosland, A (Center for Health Equity Research and Promotion, Pittsburgh)

The shutdown of in-person healthcare services during the COVID-19 pandemic was a potentially disruptive shock to patients’ chronic condition care although the VA transitioned to virtual services and issued national directives to prioritize the care of high-risk patients as strategies to potentially mitigate adverse impacts. In an operations-funded evaluation, we examined changes in VA primary care utilization, medication adherence, and hospitalizations for diabetes for high-risk Veterans in the years before and after the pandemic began FY2019-2021.

We conducted longitudinal analysis on a cohort of high-risk primary care patients with diabetes over a 3-year period using VA administrative data. Patients were in the top 15th percentile of risk based on the Care Assessment Need score and previously diagnosed with diabetes. We measured patients’ use of VA primary care by modality, adherence to diabetes medications via VA pharmacy fill data, and VA hospitalizations for diabetes complications in each quarter. Using time series models, we estimated utilization outcomes during the pre-pandemic (10/1/2019-3/30/2020), early-pandemic (4/1/2020-12/31/2020), and late-pandemic (1/1/2021-9/30/2021) phases after adjusting for quarter and patients’ sociodemographic characteristics and Elixhauser comorbidities. We included a patient random effect and adjusted standard errors for clustering by clinic.

Most patients (95%) were male with mean age 68 years (±9.6) and mean Elixhauser comorbidity score of 5.7 (±2.9). Over the 3-year period, 18% of patients died. Mean in-person primary care encounters/quarter dropped from 1.5 (±2.1) pre-pandemic to 0.7 (±1.8) early-pandemic and 1.3 (±2.0) late-pandemic while mean virtual primary care encounters/quarter increased from 1.3 (±2.0) to 2.0 (±2.7) and 1.6 (±2.3), respectively. Mean medication adherence was unchanged during the pandemic phases (81-82%). Hospitalizations for diabetes complications dropped from 10 to 7 per 1,000 patients from the pre- to early- and late-pandemic phases. In adjusted models, the early pandemic period was associated with -0.7 fewer in-person primary care encounters, 0.9 more virtual primary care encounters, and -0.002 fewer hospitalizations for diabetes complications per patient (all P < 0.001). The late pandemic period was associated with 0.10 more in-person primary care encounters (P < 0.001), 0.50 more virtual primary care encounters (P < 0.001), and no difference in hospitalizations for diabetes complications (P = 0.86). There were no significant changes in diabetes medication adherence between the pre- and post-pandemic phases.

Primary care utilization remained high with more virtual care provided than in-person care during the pandemic phases, and adherence to diabetes medications was maintained at high levels. There was a temporary decline in hospitalizations for diabetes complications in the early pandemic phase, and it is unclear whether this represented an improvement in outcomes or whether patients were avoiding the hospital.

The VA’s experience with using virtual modalities and providing medications through the mail prior to the pandemic may have mitigated potentially adverse impacts from shutdown of in-person services and enabled the highest-risk patients to receive continuous, high-quality care.