Telehealth for Endocrinology: Effect on Utilization in Non-Pandemic Care

Telehealth (TH) usage increased exponentially with the COVID-19 pandemic, aided by more permissive policies. Prior to the pandemic, policy makers and payers expressed concern that telehealth might lead to over-utilization of services. This contributed to a limited reimbursement environment. To assess evidence for this concern, we studied real-world TH at an academic medical center and hypothesized that TH would not increase utilization. We compared pre-pandemic utilization in a retrospective cohort of adult endocrinology patients with either TH (synchronous audio-visual visits) or conventional visits (CV) between 11/1/2017 - 3/16/2020. We focused on a pre-pandemic cohort to reflect usual practice. TH patients (i.e., 1+ TH visit with endocrinology in the study period) were identified using billing modifiers in EHR-derived data. CV patients were matched for provider and primary diagnosis, then selected from EHR-derived data in a 1:3 ratio. A minority of providers were performing TH visits pre-pandemic. Matching was performed to eliminate confounding from individual practice variation and to assure ability to compare patients with similar diagnoses. The majority of TH encounters in our practice during the study were clinic-to-clinic, with patients at a TH-enabled remote clinic. 75 TH and 225 CV patients were identified. 77% (55) of TH and 72% (163) of CV patients were female. The mean age was 45.2 (SD 12.4) and 49.6 (SD 16.7) years, respectively. Among TH patients, 68% (51) were white and 12% (9) were black with no statistically significant demographic differences between groups. The most common TH primary diagnoses were DM (33 patients) and thyroid disease (33 patients). Other TH diagnoses included pituitary/adrenal conditions, hyperparathyroidism, PCOS, and hypogonadism. TH patients had median follow-up of 21 months (IQR 13-25) and CV patients 18 months (IQR 11-25), p=0.613. TH patients had a range of 1 to 9 TH encounters during this time (median 2, IQR 1-3). Median TH visits per month were 0.20 (IQR 0.11-0.3). Looking at all endocrine encounters, TH patients had a median of 0.24 (IQR 0.015-0.36) encounters per month. CV patients had median of 0.20 encounters per month (IQR 0.11-0.37). Total encounters per month did not vary significantly between groups (p=0.0512, Wilcoxon rank-sum test). Patients with primary thyroid diagnoses had similar frequencies of TSH testing with TH median 4 tests (IQR 2-7); CV median 3 tests (IQR 1-5), p=0.057. A1c testing frequency was similar in patients with primary diagnosis of DM. TH patients had median 2 tests (IQR 1-3.5); CV median 3 (IQR 1-5), p=0.8012. In conclusion, TH was not associated with higher utilization of visits or lab testing in a small study of a pre-pandemic telehealth program. Coupled with larger scale and replicative studies, these findings might inform policy making and support continuation of telehealth in the post-pandemic environment.