One limitation of the study is that researchers used data only on confirmed cardiac arrest cases, not every suspected cardiac arrest called in to 911 dispatchers, the authors note. This might mean they underestimated the number of drones needed, because they would still need to be deployed on calls that turned out to be false alarms.

Another drawback of the study is it doesn’t calculate costs, which the authors note might be considerable.

An effective drone delivery network would also require a massive public education effort to make sure enough bystanders know how to use the defibrillators, said Dr. Peter Pons, a professor emeritus of emergency medicine at the University of Colorado School of Medicine who wasn’t involved in the study.

“This concept is in its infancy but there have been some initial efforts in this regard,” Dr. Pons said by email.

Drones were used to deliver small aid packages in Haiti after the earthquake, Dr. Pons said. They have also been used in a trial situation to fly laboratory test specimens from a remote location to an urban testing lab. “There is certainly the potential for a drone to deliver a defibrillator when needed,” Dr. Pons added.