A Multi-Payer Evaluation of Telemedicine on Spending and Utilization

Research objective: Telemedicine, the use of telecommunications technology to remotely diagnose and treat patients, has the potential to improve access to care and provide lower-cost alternatives for health care services, particularly in underserved areas. However, little is known about the impact of telemedicine encounters on follow-up care and spending, and how improving access to telemedicine may affect the overall use of health care services. In this study, I first investigate how episodes of care initiated by a telemedicine visit differ than those initiated by a face-to-face encounter. Next, I exploit the change in telemedicine coverage under the Minnesota Telemedicine Act (MTA), which was passed in 2015 and mandated reimbursement parity between telemedicine and in-person services for Medicaid enrollees, to examine whether improved reimbursement for telemedicine services increased its use, and led to a displacement of in-person visits.

Study design: This analysis draws upon the Minnesota All Payer Claims Database (MN APCD), which contains integrated medical and pharmacy claims and provider files from private and public payers in MN over the period 2009-2016. Telemedicine visits are identified using telemedicine-specific CPT codes or CPT code modifiers. In order to investigate whether episodes of care initiated by a telemedicine visit are less costly than those initiated by a face-to-face encounter, I aggregated enrollee claims into 30 and 60-day episodes of care, initiated after a telemedicine or non-telemedicine evaluation and management encounter. Because the use of telemedicine requires the buy-in of both patients and providers, I first estimated the likelihood of a telemedicine encounter as a function of patient, provider, and encounter characteristics, demographic controls, and controls for broadband coverage. Additionally, I controlled for whether the provider is affiliated with an ACO, and clinic and hospital experience with telemedicine and IT systems, using survey data from the state. Next, I examined the impact of this initial telemedicine visit relative to a non-telemedicine visit, on follow-up utilization, including emergency department use, office and inpatient visits, prescriptions, and on total allowed charges and paid amounts for that episode of care. I then examined the causal effects of the MTA on the total number of telemedicine and non-telemedicine services. Using a differences-in-differences model, where Medicare beneficiaries in non-rural areas, who are not affected by state telemedicine policies served as a control group, I evaluated the change in telemedicine and non-telemedicine services following the MTA, across all payer types.

Population studied: The state repository of health care claims data, which contains claims for approximately 89% of Minnesotans with health care coverage. I identified 32,142 unique telemedicine users.

Principal Findings: The findings from this study are forthcoming.

Implications for policy: This work makes a novel contribution to the literature by leveraging a unique all-payer claims dataset to characterize telemedicine utilization and spending. This study is one of the first to look across multiple payers at how telemedicine encounters affect subsequent utilization and spending relative to traditional care, as well as to examine the effect of an exogenous change to telemedicine reimbursement on the overall use of health services.