The Drones Are Coming

In December an autonomous drone was used to deliver an automated external defibrillator (AED) to help save a 71-year-old man who suffered a cardiac arrest while shoveling snow outside his home in Trollhättan, Sweden.

It took just more than three minutes from the time the alarm was first raised to the Everdrone unit’s delivery, which enabled a doctor on scene to begin defibrillation before an ambulance arrived. 

The idea of using drones to deliver AEDs to the scenes of out-of-hospital cardiac arrests (OHCAs) is gaining traction following two academic studies—one in Sweden and one in the US—of its feasibility. 

There is a strong case to be made. Despite advances in emergency cardiac care, OHCA remains a significant health problem in the United States. More than 350,000 sudden cardiac arrests occur each year in the US, with fewer than 10% of victims surviving.

The chance of OHCA survival decreases by 10% for every minute without defibrillation, notes Wayne Rosamond, PhD, MS, a professor in the University of North Carolina at Chapel Hill Gillings School of Global Public Health’s Department of Epidemiology.

The probability of survival doubles when a bystander administers CPR and uses an AED before EMS arrives, but bystander AED use occurs in less than 2% of cardiac arrests in the US. AEDs can be difficult to locate and are rarely available in the home, where some 80% of OHCAs occur.

Survival is most likely when CPR and defibrillation are delivered within 5 minutes after the start of a cardiac arrest. However, many EMS response times are longer than that.

Studies show that the time saved using drones to deliver AEDs can be between 2–3 minutes.

“Recent data suggests 2 or 3 minutes of action makes the difference between life and death,” notes Rosamond.

Studies Show Success

In Sweden, “After GIS analysis of optimal placement, we ran simulations with beyond-visual-line-of-sight flights in 2016 and published in JAMA in 2017,” says lead researcher Andreas Claesson, PhD, associate professor at the Center for Resuscitation Science in the Department of Clinical Science and Education, Södersjukhuset, Karolinska Institutet.

Researchers there contacted Everdrone and launched a four-month clinical study in west Sweden on June 1, 2020. Drones were dispatched to 12 out of 14 cases of suspected cardiac arrest and successfully delivered AEDs in all but one case. In seven cases the drone arrived before an ambulance.

The University of North Carolina at Chapel Hill and North Carolina State University’s Institute for Transportation Research and Education conducted a study in 2019 examining how medical drones could reduce the time it takes to get lifesaving assistance to victims and improve the OHCA survival rate. 

“You might see AEDs in shopping malls, at the airport, in stores. We wanted to see if drones do any better than people trying to find one,” says Rosamond, who served as the US study’s principal investigator. “We did a test to see if someone could find one faster than a drone can fly autonomously and locate the victim. The vast majority of drones found the victim faster.”

The North Carolina universities conducted a randomized trial consisting of 35 tests in a community setting in which an AED was delivered by an autonomously flying drone while a bystander searched for and retrieved a fixed-location AED from the surrounding area. The study was published in the September 17, 2020 issue of the New England Journal of Medicine. 

In each test an out-of-hospital cardiac arrest was simulated with the use of a life-size manikin, with two participants matched by sex and age present at the scene.

Seven tests were conducted in each of five different geographic zones. Zones were selected to present different environmental challenges to acquisition of the AED by bystanders and to drone navigation. Sites varied by the number of AEDs available within 600 feet of the simulated cardiac arrest. 

Drone launch sites were not visible from the cardiac arrest sites, and the distance from the launch site to the site of the arrest differed by zone, ranging from 780–1290 feet.

In each test one participant was randomly assigned to call a mock 9-1-1 telecommunicator, who initiated the drone’s autonomous flight using an automated flight path and GPS coordinates to deliver the AED to the simulated victim. 

The other participant simultaneously conducted a ground search to locate an AED and return to the site with it. 

In many trials the drone reached the simulated victim first. The difference in the median AED delivery time between drone delivery and the ground search method differed by zone and ranged from less than 2 minutes, 56 seconds to 1 minute, 42 seconds.

While one zone had the highest density of AEDs in one area, access to them was limited, while another zone demonstrated the shortest mean distance to an AED in the area, suggesting the relative timeliness of drone delivery and ground search depends on the physical setting.

Among participants randomly assigned to call for the drone, 89% reported in post-trial interviews that they felt comfortable as the drone approached. Nearly half the participants randomly assigned to conduct ground searches reported difficulty finding an AED. 

Welcome Help

Rosamond says researchers were interested in how people would react to drones coming to the scenes of cardiac arrests: Would they be afraid of them? Would they not want to interact? All participants reported they would be willing to access an AED drone-delivery system in a true out-of-hospital cardiac arrest.

“It actually felt good that they knew that something was coming directly to help them, so they didn’t have to go out and find an AED on their own,” says Rosamond. 

Study participants found the drone-delivered AEDs to be user-friendly as the devices “talked” to them and guided them through the experience. “That suggests it’s acceptable by the population and would be seen as a good thing,” Rosamond adds.

Future Directions

The purpose of the North Carolina study was to inform future investigations for designing and testing medical drones beyond the campus environment, integrate the concept with existing emergency medical systems, and move it outside to other parts of the state. 

Rosamond says feedback from EMS and aerospace engineering partners regarding the use of drones in delivering AEDs has been positive.

Researchers have begun working with EMS leaders in Orange County, California—which has a significant rural area that also could benefit from the technology—to “see how we can take what is a theoretical solution to actual processes,” says Rosamond.

Rosamond emphasizes drones aren’t being used to replace ambulances.

“There are a lot of great 9-1-1 emergency medicine systems in this country,” he says. “We’re trying to further our partnering with local EMS systems and working on ways to best integrate this technology.

“For example, someone calls 9-1-1 because they see someone having a cardiac arrest. The person who answers that 9-1-1 call in most places would very quickly know the longitude and latitude of the caller with cell phone and 9-1-1 technology. We’re looking at how we can get that information to the brain of a drone that can get launched and get there quickly.”

He envisions several avenues for fitting the approach into the 9-1-1 dispatch process.

“There might be multiple drone stations in a county, and you can envision which station would be the best for getting a flight plan together using GPS coordinates and the safest and most effective routes,” Rosamond says. “There are a lot of important logistical issues we’re working with.”

Drone bases could be independently located or placed in fire or EMS stations, Rosamond says.

“An essential drone technologist could be in charge at two or three of these stations,” he says. “When the call comes in, an operator could call it up on his screen and launch and view the flight in real time.”

Busy 9-1-1 operators could also transfer victim location information to a parallel system that automatically launches the drone. 

Rosamond notes that with the drone, “you can put eyes on the ground. If a drone brought the AED, there will probably be an operator watching through video capability who can tell you exactly what to do.”

There can be other medical uses for drones, such as using them to deliver EpiPens and other emergency equipment to supplement getting an ambulance to the scene as soon as possible, Rosamond says, echoing the sentiment of the Swedish researchers.

Building a System

There are certain metrics to determine the economic feasibility of such interventions. For instance, a drone capable of carrying an AED and its drone docking station are estimated to cost $15,000, with annual maintenance being an estimated 20% of the purchase price.

“By most standards it is economically feasible to do this when you think about other alternatives, such as the cost of an equipped ambulance,” Rosamond says.

In Sweden the drone deployment was the result of a collaborative effort involving the Karolinska Institutet, national emergency operator SOS Alarm, Region Västra Götaland, and Everdrone.

There a command pilot oversees the operation, including obtaining air traffic control clearance for the drone. 

“The dispatch center is notified on arrival and delivery of the AED on site so they can instruct the caller to retrieve the AED in line with T-CPR instructions,” says Claesson.

In terms of time saved by using drones, Claesson notes in the Karolinska Institutet’s published study, 1 minute, 52 seconds was saved in the seven cases where the drone arrived first.

Meanwhile, Everdrone is exploring bringing the technology to other countries.

Of its feasibility in the US, Claesson points out while OHCA cases are the same in the US as in Sweden, regulations are not.

“GIS studies on optimal placement are needed, as well as integration with the US EMS and dispatch organizations,” he adds.

Claesson says AED-equipped drones are a complement to standard care that should be evaluated for feasibility, safety, and clinical outcomes.

“In my opinion, feasibility is shown,” he says. “With more flights more survivors will appear.” 

Fast Facts: Unmanned Aircraft Systems

• 859,998 drones registered 
  • 328,171 commercial drones registered
  • 529,239 recreational drones registered
  • 2588 paper registrations
• 262,964 remote pilots certified
• 204,565 completion certificates for The Recreational UAS Safety Test (TRUST) issued by test administrators

—Source: Federal Aviation Administration, February 2022 statistics

Carol Brzozowski is a freelance journalist and former daily newspaper reporter in South Florida. Her work has been published in more than 200 media outlets.