A Decision-Analytic Model to Evaluate Cost-Effectiveness of Regional Implementation of a Mobile Stroke Unit

Background and Objectives Mobile stroke units (MSUs) have the potential to improve functional outcome of ischemic stroke patients, through shortening onset-to-treatment times. Previous cost-effectiveness studies have limited generalizability to nonmetropolitan settings and did not evaluate cost-effectiveness over a lifetime horizon. We aimed to develop a regionally adaptable decision-analytic model, to evaluate cost-effectiveness of MSU implementation and to identify the optimal dispatch scenario. Methods We developed a generalizable state-transition microsimulation model with modifiable region-specific parameters and dispatch characteristics to evaluate the lifetime cost-effectiveness from a health care perspective of 1-year MSU implementation. We used the southwest of the Netherlands (1,770,000 inhabitants, 1,592 km ², 7 primary stroke centers, 2 thrombectomy-capable stroke centers) as an example. Region-specific input parameters for the model, such as population density, age distribution, and driving times, were obtained at the level of postal codes. We developed a virtual cohort of suspected stroke patients based on age-dependent stroke risks and the number of inhabitants per postal code. We compared the combined dispatch of an MSU and emergency medical services (EMS) with dispatch of EMS alone for patients with onset-to-alarm time <6 hours, living within the catchment area of the MSU. In the base case analysis, the MSU could be dispatched to all postal codes in the study region between 7.00 am and 11.00 pm from a central dispatch site. We assessed the long-term cost-effectiveness through incremental net monetary benefits (iNMBs). Discount rates were 1.5% for effects and 4.0% for costs. Results In the base case scenario, the MSU was dispatched to 2,080 of 3,628 patients (57.3%) with a suspected stroke and onset-to-alarm time <6 hours, resulting in a lifetime gain of 399 (95% CI 384-414) additional quality-adjusted life years, 3.9 million (95% CI 3.5 million-4.3 million) cost savings, and an iNMB of 23.9 million (95% CI 22.8 million-24.9 million). A smaller catchment area for MSU dispatch was associated with increased cost-effectiveness. Discussion Adding an MSU to the dispatch strategy for suspected stroke patients is expected to be cost-effective in our region. Our model facilitates evaluation of the cost-effectiveness of MSU implementation in different regions, settings, and scenarios with varying characteristics.