Telemedicine use increased sharply during the COVID-19 pandemic. Although its utilization declined afterwards. It remained higher than before the pandemic. At Penn Medicine, more than 2.1 million virtual visits occurred from 2020 to 2022. In 2024, about 4% of over 3 million outpatient visits were conducted virtually. However, the effect of telemedicine on overall health care charges is still unclear. Most prior studies have focused on specific payer groups or clinical outcomes and have not provided a systemwide financial perspective. David A et al. conducted a target trial emulation using billing data and electronic health records (EHRs). They compared telemedicine with in-person visits across 10 common diagnoses and assessed 30-day care charges as well as follow-up visits.
In this study, billing data and EHRs from five Penn Medicine hospitals were analyzed between January 1 and April 30, 2024. A total of 163,308 visits (mean age = 49.2±19.1 years, female = 66.4%) were included, comprising outpatient telemedicine visits (n = 29,446, female = 67%, male = 33%) and in-person visits (n = 133,862, female = 66.2%, male = 33.8%). Excluded inpatient and emergency encounters. Index visits initiated a new episode of ambulatory care. Refill, imaging-only, telephone encounters, and laboratory observations were excluded as index visits but were counted as downstream services.
Analyses were restricted to the 10 most common telemedicine diagnostic categories based on ICD-10 codes. The study compared 30-day subsequent visits and episode charges within a window spanning 7 days before to 30 days following the index visit. Charges reflected billed amounts excluding facility and professional fees. Propensity score matching adjusted for socio-economic, demographic, and clinical factors. Regression models estimated mean charge differences and relative risks with robust standard errors, and multiple sensitivity analyses were performed.
Lower use of telemedicine was observed for abnormal findings without diagnosis at 4.3%, skin inflammation at 9.3%, obesity at 23%, skin inflammation at 9.3%, neurodevelopmental disorders at 23%, and respiratory symptoms at 10.3%. After propensity score matching, 27,541 telemedicine visits were matched to 72,481 in-person visits. Mean 30-day episode charges were found to be $96.60 (95% confidence interval [CI]: $92.24-$100.96) for telemedicine and $509.21 (95% CI: $500.65-$517.77) for in-person care, with a difference of $412.62 and p<0.001. Telemedicine was also associated with fewer follow-up visits (3.44 vs 4.44) with a relative reduction of 23%.
Additionally, charge reductions were consistent across five hospitals, largest at Penn Presbyterian Medical Center (PPMC) with a decrease of $550.70 (95% CI: -$575 to -$526.40) and at Pennsylvania Hospital (PAH) with a decrease of $473.25 (95% CI: -$495.25 to -$451.25). By conditions, the greatest savings occurred for respiratory signs and symptoms with a reduction of $828.78 (95% CI Penn Presbyterian Medical Center: -$853.23 to -$804.34), neurodevelopmental disorders with a reduction of $28.88 (95% CI: -$54.72 to -$3.04), and depressive disorders with a reduction of $69.47 (95% CI: -$100.90 to -$38.40), as well as abnormal findings with diagnosis with reduction of $991.55 (95% CI: -$1021.11 to -$961.99).
COVID-19 diagnoses were associated with substantially fewer revisits with a risk ratio (RR) of 0.35 (95% CI: 0.30-0.41). Conversely, obesity showed no meaningful difference in revisit risk with RR of 1.02 (95% CI: 0.92-1.12). While sleep-awake disorders were linked to a slightly higher risk of revisits, with an RR of 1.27 (95% CI: 1.18-1.37).
This study’s limitations include its single-health system design, potential selection of milder cases for telemedicine, focus on short-term charges, possible residual confounding, and reliance on the HER billing and coding data that may vary in accuracy.
In conclusion, this study highlights that telemedicine was associated with lower charges and fewer short-term follow-up visits across conditions. These findings support the hybrid care models and inform value-based strategies and reimbursement policies to ensure sustainable outpatient telemedicine delivery.
Reference: Zhang B, Li L, Lu Y, et al. Episode Charges and Subsequent Visits After Telemedicine vs In-Person Care. JAMA Netw Open. 2026;9(2):e2556127. doi:10.1001/jamanetworkopen.2025.56127
