A computer simulation of 28,349 real cardiac arrests around Paris found that drones carrying automated external defibrillators (AEDs) could put a device on scene within five minutes in more than 95% of cases. Fetching an existing wall-mounted AED by road manages that in 30 to 40% of cases. Dr. Hillary Minka, an emergency physician at Lariboisière Hospital in Paris, presented the study at the European Emergency Medicine Congress in Paris on Saturday, September 26.
The team modeled seven departments of the Île-de-France region, leaving out central Paris, and measured how long an AED takes to reach the center of each census block under three strategies: more fixed AEDs, drones alone, or both. The hybrid version cut average AED access time by 3.76 minutes across every block, and by more than ten minutes in some outlying and congested ones.
Every one of those flights was a straight line drawn on a map. The researchers flew no drones.
Fixed AEDs Reach 30% of Paris-Region Cardiac Arrests Inside 500 Meters
Only 30% of the 28,349 out-of-hospital cardiac arrests recorded between 2011 and 2024 happened within 500 meters (1,640 feet) of one of the region’s 1,893 registered fixed AEDs, Minka told the congress. Closing that gap with cabinets alone would take 1,712 more devices, for 3,605 in total.
The shortages sit in the countryside and on the outskirts of Paris, according to the study team, and an extra 910 fixed AEDs would bring 84.5% of cases inside the 500-meter target. The 30% figure flatters the cabinets. Matthieu Heidet, a professor of emergency medicine at Henri-Mondor University Hospital in Créteil and a co-author, told The Guardian that in France only 8% of cardiac arrest patients get a public AED applied before emergency crews arrive, partly because so many devices sit inside private buildings that are not open around the clock.
Two Hundred Drone Bases Do the Work of 1,700 New AED Cabinets
A drone base covers a 3,900-meter (2.4-mile) radius against 500 meters for a fixed AED, so each base added cut the cabinets needed. With 100 bases and 26 new AEDs, coverage passed 97%. With 200 bases, 871 existing AED sites and four new ones, it reached 99.4%.
The candidate base locations were existing AED sites, mobile intensive care units and fire stations. Minka’s design puts a trained operator or an automated system in charge of each flight under emergency-service supervision, with a bystander collecting the AED on arrival and a dispatcher talking them through the shock. “Drones will never replace bystanders, resuscitation efforts or emergency medical teams,” she said. Dr. Felix Lorang, who heads the emergency department at Klinikum Lippe in Germany and sits on the congress abstract committee, put the same limit the other way: a bystander who cannot do CPR or work the device still waits for the ambulance, drone or no drone.
The Model Flew Nothing, but France Already Flies AED Drones
The study computed drone travel time as the crow flies, so weather, airspace rules and operational delays are absent from every number above. Minka said the findings need prospective trials and real-world tests before anyone acts on them. Abstract OA066 is not peer reviewed and received no funding.
Those real-world numbers exist, and some of them are French. Everdrone entered live dispatch service for SAMU 76 in Forges-les-Eaux, Normandy, on December 2, 2025, the first French deployment tied into the emergency dispatch chain, roughly an hour and a half by road from Paris. The same operator’s Swedish network beat the ambulance to a Borås cardiac arrest in July, and a bystander restarted the patient’s heart. Sweden’s 2023 Lancet Digital Health trial measured actual drone arrivals against actual ambulances over 11 months, which is the study design Minka is asking for. In the United States, Duke Health began flying AEDs to live 911 calls in July and James City County, Virginia followed on August 1.
DroneXL’s Take
The number I’d want a French health minister to remember from this study is not 99.4%, it’s 8%. Heidet’s figure says a region with 1,893 registered defibrillators gets one onto a chest before the ambulance in fewer than one case in twelve, and the Paris model explains why: the cabinets are in the wrong places, and a good share of them are behind a locked door at 2 a.m. A drone base doesn’t care which side of the door it’s on. That is the case for AED drones, and it’s the kind of delivery I’ve argued for since I first covered Everdrone’s Trollhättan save in January 2022: a flight that has to happen, not a toothpaste run.
Where I part ways with the study is the method. Straight-line simulations were the state of the art a decade ago. Sweden has flown real AED drones against real ambulances since 2020 and folded them into routine dispatch, and France has had a base flying live calls in Normandy since December. My read is that Île-de-France doesn’t need another model to find out whether drones beat the road network. It needs SAMU 76’s flight logs and one base on the outskirts of Paris.
Sources: European Society for Emergency Medicine, The Guardian, Everdrone
DroneXL uses automated tools to support research and source retrieval. All reporting and editorial perspectives are by Haye Kesteloo. Read our editorial standards.
