Cost Minimization Analysis of a Teledermatology Triage System in a Managed Care Setting.

Importance: Teledermatology (TD) enables remote triage and management of dermatology patients. Previous analyses of TD systems have demonstrated improved access to care but an inconsistent fiscal impact. Objective: To compare the organizationwide cost of managing newly referred dermatology patients within a TD triage system vs a conventional dermatology care model at the Zuckerberg San Francisco General Hospital and Trauma Center (hereafter referred to as the ZSFG) in California. Design, Setting, and Participants: A retrospective cost minimization analysis was conducted of 2098 patients referred to the dermatology department at the ZSFG between June 1 and December 31, 2017. Intervention: Implementation of the TD triage system in January 2015. Main Outcomes and Measures: The main outcome was mean cost to the health care organization to manage newly referred dermatology patients with or without TD triage. To estimate costs, decision-tree models were constructed to characterize possible care paths with TD triage and within a conventional dermatology care model. Costs associated with primary care visits, dermatology visits, and TD visits were then applied to the decision-tree models to estimate the mean cost of managing patients following each care path for 6 months. The mean cost for each visit type incorporated personnel costs, with the mean cost per TD consultation also incorporating software implementation and maintenance costs. Finally, ZSFG patient data were applied within the models to evaluate branch probabilities, enabling calculation of mean cost per patient within each model. Results: The analysis captured 2098 patients (1154 men [55.0%]; mean [SD] age, 53.4 [16.8] years), with 1099 (52.4%) having Medi-Cal insurance and 879 (41.9%) identifying as non-White. In the decision-tree model with TD triage, the mean (SD) cost per patient to the health care organization was $559.84 ($319.29). In the decision-tree model for conventional dermatology care, the mean (SD) cost per patient was $699.96 ($390.24). Therefore, the TD model demonstrated a statistically significant mean (SE) cost savings of $140.12 ($11.01) per patient. Given an annual dermatology referral volume of 3150 patients, the analysis estimates an annual savings of $441 378. Conclusions and Relevance: Implementation of a TD triage system within the dermatology department at the ZSFG was associated with cost savings, suggesting that managed health care settings may experience significant cost savings from using TD to triage and manage patients.