Little is known on differences in healthcare utilization between telehealth users (USER) and non-users (NON-USER) in persons with diabetes mellitus (DM) during the COVID-19 pandemic. We compared healthcare utilization of USER to that of NON-USER in Medicare beneficiaries with DM who enrolled in fee-for-service plans, using 100% Medicare claims database. The USER included persons who used any telehealth services during 03/2020-06/2020. We used nearest neighbor propensity score matching to construct a NON-USER group, who did not use any telehealth services during the pandemic (03/2020-06/2021), based on demographic, healthcare utilization, and comorbidity variables measured at baseline (03/2019-02/2020). We used linear regression to compare healthcare utilization during the later pandemic period (07/2020-06/2021) between USER and NON-USER groups, controlling for baseline characteristics. Utilization was measured as per person use of each type of healthcare services in 12 months. We also performed a subgroup analysis among patients who used fewer than 15 clinician office visits (low utilizers). The USER group used more services in all service types, ranging from 0.02 to 2.08 visits in 12 months (table). These differences were smaller for the low utilizers. We suggest future studies identify reasons behind these differences.
Y.Shao: Employee; Biogen. Y.Wang: None. X.Zhou: None. D.B.Rolka: None. P.Zhang: None.
