NEW HANOVER COUNTY — Over the last decade, as New Hanover County added 30,000 new people, the region’s emergency medical services reduced its response time by nearly a minute.
With so many new developments and increasing traffic congestion, how did that happen?
“We started to leverage technology more and more,” Rick O’Donnell, New Hanover Regional Medical Center (NHRMC) Emergency Medical Services (EMS) director, said.
In 2012, EMS started to see a downward trend (in a good way), shedding response time down to six minutes and four seconds this year.
O’Donnell credits the reduction in time to entering a “modern era” of EMS technology. This shift ultimately landed his team in the top one percent of EMS systems in the country last year, after receiving accreditation from the Commission on Accreditation of Ambulance Services.
And in 2019, NHRMC’s EMS department plans to get its artificial intelligence software fully online.
Road to AI
EMS plays a game of chess all day long. At peak volume, the system has 17 units in service, according to NHRMC’s spokesperson, Julian March. That’s about one truck for every 13,000 residents in the county.
But depending on available units, the location of an incident, road impedance, road connectivity, and traffic congestion, it’s not as simple as getting truck A to point B. Now, these decisions are made manually — by a human, that is. But soon, the moves could be made by artificial intelligence (AI), with a human standing by. In other words, an automated system would decide which units respond to which incidents, based on predictive analytics of traffic situations.
“Hopefully, a lot of it will be automated, but you’ll have the human experience and the trained professionals working in concert with the technology,” O’Donnell said.
NHRMC’s EMS already incorporates AI technology in its decision-making, informed by the software program, Optima. One function of this program is predicting the future (or, at least, the future of traffic conditions). Now, O’Donnell said EMS brings on additional units based on anticipated call volume.
“Where we could anticipate call volume on a July Monday at 5 o’ clock could be completely different from where we’re anticipating call volume on a December Monday at 5 o’ clock,” he said.
Two steps ahead
Each truck has automatic vehicle location (AVL) units already installed. These units relay GPS location and navigation data to EMS’ AI software. “The system learns,” O’Donnell said. “And constantly looks over three years worth of data. It learns your road networks and road impedance.”
When asked if — aside from a media inquiry — other agencies or groups have asked to see any of this data, O’Donnell said no.
Instead of analyzing conditions as they are, O’Donnell said EMS’ new technology is allowing his team to consider what resources will be like in 10, 15, 30 minutes. “We’re constantly moving the pieces around to make sure we’re two steps ahead of what’s going to come our way,” he said.
Accounting for the season, time of day, week, and recorded patterns, the system shows EMS “hot spots.”
“So you’re not looking at, ‘Oh, this is what I have available now.'” he said. “That’s the human tendency, right?”
Basic insight and pattern recognition might help EMS responders stay alert, say, around 2 a.m. on a Saturday in July near Wrightsville or Carolina Beach. But AI is helping the system pick up on what humans could miss.
For instance, O’Donnell said he’s noticed spikes around bedtime in some neighborhoods, that he wouldn’t otherwise consider a retirement community.
“You’ll see little hot spots on certain neighborhoods at certain times that maybe you otherwise wouldn’t pick up on,” he said.
One EMS AI step already underway is installing a monitor at the county’s 911 dispatch center. EMS is not funded by taxpayer dollars, but O’Donnell said he wants dispatchers to see what he’s seeing when they receive a call. Through working with New Hanover County’s Information Technology department, EMS is getting a monitor installed to help prompt more efficient responses.
“Any second, or seconds you can save, is important,” O’Donnell said.
Send tips and comments to Johanna Ferebee at firstname.lastname@example.org