# An episode-based cost analysis of virtual-first versus in-person-first care to treat common acute conditions among members of a large national payor

## TL;DR

This study compares the costs of virtual-first and in-person-first care for treating common acute conditions and finds that virtual care can be more cost-effective.

## Contribution

The paper introduces an episode-based cost analysis to evaluate virtual-first care's economic impact compared to in-person care for acute conditions.

## Key findings

- Virtual-first care was associated with 10–24% lower episode costs for six conditions in Medicare Advantage members.
- In commercial fully-insured members, virtual-first care showed 9–33% lower costs for 12 of 16 conditions.
- No significant cost differences were observed for certain conditions like skin inflammation and low back pain.

## Abstract

The potential of virtual care as an alternative to in-person visits is promising, yet its economic impact is insufficiently understood. This evaluation represents an episode-based, cost impact analysis of virtual-first (versus in-person first) care to treat the most prevalent primary care acute conditions among Medicare Advantage (MA) and commercial fully-insured (C-FI) members of a large national health plan in the United States.

Retrospective episodes-of-care and medical claims analyses identified members (N = 366,195; MA: 126,363, C-FI: 239,832) with resolved, pre-specified, acute primary care episodes (N = 455,231; MA: 141,034, C-FI: 314,197) between 1/1/2022–6/30/2022. Propensity score weighting estimated % difference in healthcare expenditures between virtual-first episodes and an adjusted cohort of in-person-first episodes.

Within the MA cohort, 7.6% (range: 0.7–24.8%) of episodes utilized virtual-first care with observed cost-of-episode 10–24% lower than in-person-first care for 6 of 11 included conditions (all P < 0.05), including: otolaryngology disease (-24 ± 2%), rhinitis (-20 ± 4%), gastroenterology disease (-20 ± 7%), minor bacterial skin infections (-17 ± 7%), sinusitis (-14 ± 4%), and bronchitis (-11 ± 4%). Within the C-FI cohort, 12.6% (range: 2.8–40.4%) of episodes utilized virtual-first care with observed cost-of-episode 9–33% lower than in-person-first care (all P < 0.001) for 12 of 16 included conditions (all P < 0.001), including: urinary tract infection (-33 ± 5%), viral skin infection (-29 ± 6%), gastroenterology disease (-27 ± 5%), rhinitis (-28 ± 5%), otolaryngology disease (-25 ± 2%), sinusitis (-25 ± 2%), urological disease (-23 ± 9%), contact dermatitis (-19 ± 5%), viral pneumonia (-17 ± 12%), bronchitis (-15 ± 4%), fungal skin infection (-11 ± 6%), and minor bacterial skin infection (-9 ± 7%), and 4 ± 2% higher to treat exposure to infectious disease (P = 0.001). There were no between-group differences in cost-of-episode to treat: skin inflammation (MA & C-FI), urinary tract infection (MA), exposure to infectious disease (MA), fungal skin infection (MA), low back pain (C-FI), or migraine headache (C-FI) (all P > 0.081).

This real-world study of a large national sample of geographically diverse members demonstrates the potential of virtual-first care to resolve acute conditions at lower cost compared to in-person-first care. The use of episode-based analytical tools enhances the significance of these findings by enabling a proxy for clinical outcomes.