Changes in Healthcare Utilization Among Medicare Beneficiaries With Diabetes 2 Years Into the COVID-19 Pandemic

INTRODUCTION: The COVID-19 pandemic abruptly impacted healthcare service delivery and utilization. However, the impact on older adults with diabetes in the U.S. is unclear. This study aimed to estimate changes in healthcare utilization among older adults with diabetes during the initial 2 years of the COVID-19 pandemic compared with the changes in the 2 years before and to examine the variation in utilization changes by demographic and socioeconomic characteristics. METHODS: In this study, we analyzed changes in utilization, measured by the average use of healthcare services per 1,000 persons with diabetes, using medical claims for Medicare fee-for-service beneficiaries aged ≥67 years. Utilization changes by setting (acute inpatient, emergency room, hospital outpatient, physician office, and ambulatory surgery center) and by media (telehealth and in person) were examined for 22 months of the pandemic (March 2020–December 2021) compared with those during prepandemic period (March 2018–December 2019). We also estimated utilization changes by beneficiaries’ age group, sex, race/ethnicity, and residential urbanicity. RESULTS: The study sample consisted of approximately 6 million beneficiaries with diabetes each month. In the first 2 years of the pandemic, the average use of healthcare services by setting was 5%–17% lower than the prepandemic level for all types of services. Phase 1 (March 2020−May 2020) had the largest decrease in utilization: physician office visits changed by −51.2% (95% CI= −55.0%, −47.5%), ambulatory surgery center procedures changed by −45.1% (95% CI= −49.8%, −40.4%), emergency room visits changed by −36.9% (95% CI= −39.0%, −34.7%), acute inpatient stays changed by −31.5% (95% CI= −33.6%, −29.3%), and hospital outpatient visits changed by −27% (95% CI= −29.3%, −24.8%). The reduction in utilization varied by sociodemographic subgroup. During the pandemic, the use of telehealth visits increased by 511.1% (95% CI=502.2%, 520.0%) compared with that in the prepandemic period. The increase was smaller among rural residents. CONCLUSIONS: Medicare beneficiaries with diabetes experienced a reduction in the use of healthcare services during the COVID-19 pandemic, some of which persisted through 2 years into the pandemic. Telehealth visits increased but not enough to overcome decreases in in-person visits. Understanding these patterns may help to optimize the use of healthcare resources for diabetes management in the postpandemic era and during future emergencies.