Cost-Effectiveness of Drone-Delivered Automated External Defibrillators for Cardiac Arrest.

Abstract

Background: Out-of-hospital cardiac arrest (OHCA) is a significant cause of mortality and morbidity in North America, for which timely defibrillation of shockable rhythms is essential. Drones have been proposed as an intervention to improve response time and are being implemented in practice.

Aim: To determine the cost-effectiveness of drone-delivered automated external defibrillators (AEDs) for OHCAs.

Methods: Using data from 22,017 OHCAs in Ontario, Canada over 10 years, we developed a comprehensive computational framework combining machine learning, optimization and a Markov microsimulation model to provide an economic evaluation of 964 different drone networks across a wide range of sizes and configurations. We simulated response times, survival outcomes, lifetime quality-adjusted life-years (QALYs), lifetime healthcare costs, and 10-year operational costs for each network.

Results: All 964 drone networks were cost-effective. We identified 20 networks on the cost-QALY efficient frontier, each with shorter response times, more survivors across all categories, and higher costs per survivor. Historical ambulance response (i.e., standard care) had mean response time of 6 min 21 s. On the efficient frontier, average drone response times were 32% to 71% shorter than standard care. There were 1,855 (8.4%) survivors to hospital discharge in standard care, which increased by 21% to 46% across the 20 drone networks. The smallest non-dominated drone network, with 20 drones, cost $20,912 per QALY gained. All drone networks had higher net monetary benefit than standard care. Cost-effectiveness was even greater for shockable and witnessed populations. Extensive sensitivity analyses showed that our results were robust to changes in modelling assumptions.

Conclusions: Drone-delivered AEDs were associated with reductions in response time, mortality and morbidity, and were found to be highly cost-effective relative to standard ambulance response with no drones.