Background: An estimated 350,000 individuals experience out-of-hospital cardiac arrest (OHCA) in the United States (US) annually; only approximately 10% survive. The probability of survival after OHCA doubles when a bystander uses an automated external defibrillator (AED) before emergency management services arrives; however, bystander AED use is less than 2% in the US. We conducted a series of randomized trials in a community setting comparing the delivery of an AED by aerial drone to that of a bystander searching for and retrieving an AED from fixed locations in the area. Methods: We conducted 35 simulation trials of an OHCA using a life-sized manikin accompanied by two participants, paired by gender and age (18-34, 35-49, 50-65 years), in a community setting, with seven trials per zone in five distinct zones. Zones represented different environmental challenges to drone navigation and pedestrian acquisition of an AED and varied in number of fixed AEDs in the area (range 1-8 AEDs in a 600 ft radius from the OHCA site). We used a DJI Matrice 600 Pro drone, modified to carry a standard AED, flying autonomously from a preprogrammed flight path to the OHCA site. Drone launch sites varied by zone and ranged from 800 to 1300 ft from the OHCA site. We randomized participants either to call a mock 9-1-1 telecommunicator who would initiate the drone’s flight sequence, or to conduct a ground search to locate and retrieve an AED from a fixed location. We compared the delivery time of an AED by drone to that of an AED acquired by ground search. We conducted pre- and post-trial interviews with each participant to query perceptions. Results: We conducted 18 trials with women and 17 with men. These included 15, 11, and 9 trials across the three age strata, respectively. The average time (minutes: seconds) from onset of the simulated OHCA to AED delivery was 1:21 faster by drone (4:45, standard deviation (SD) = 0.34) compared to ground search (6:06, SD = 3:21). In 71% of trials (n=25 of 35), the drone delivered the AED to the event site within 5 minutes, compared to 51% (n=18 of 35) of ground searches (p=0.09). Median AED delivery time was faster by drone than by ground search in four of the five zones (range 0:17 to 2:56). Ground search was faster than the drone in one zone with 7 AEDs within 600 ft of the simulated OHCA event site and the shortest average distance to an AED (254 ft) of all the five zones. Among participants randomized to call for the drone, 89% reported that they felt comfortable as the drone approached, and 72% reported having no safety concerns. Nearly half of participants randomized to conduct a ground search reported difficulty finding an AED. Conclusion: Our study suggests that drone delivery of an AED to the site of an OHCA is feasible and acceptable in a community setting. Drones may provide more timely access to early defibrillation compared to bystander search and acquisition of an AED from the surrounding area.