For every minute that someone has a heart attack, their chances of survival decreases approximately 10 percent with each minute that passes. Most heart attacks aren't properly diagnosed until first responders are on the scene, which leaves emergency operators to estimate the severity of the situation while on the phone. However, one impressive AI could be a key in diagnosing heart attacks by using clues that most human operators wouldn't pick up.
The Corti Signal is an AI developed by the company of the same name that uses real-time speech analysis with advanced machine learning to help read the context clues of critical conversations. It's been so successful that Copenhagen EMS uses it to detect cardiac arrest faster and with better accuracy.
"Conversations are noisy, implicit, and hard to understand, but they contain a goldmine of information," the company's website noted. "We have developed a multitude of deep neural networks that listen directly to a sound stream and extract the most important features. The better the quality of these features, the better our prediction and reasoning frameworks are."
The partnership between Corti and the Danish city started in 2016. For over a year, dispatchers have been able to have an assistant in triaging a patient. The AI not only listens to the caller's conversation, but it also picks up on nonverbal cues like breathing patterns. That data is synthesized by Corti against millions of earlier emergency calls for patterns.
Over 1.5 million heart attacks and strokes happen each year in the United States alone. Cardiovascular emergencies account for one in every three deaths in the U.S. as well. Those numbers get just as difficult overseas and elsewhere around the world. However, thanks to Corti, dispatchers in Copenhagen have help in lowering those numbers.
"This is an innovation with the potential to change the way Emergency Medical Services handle emergency calls," said Freddy Lippert, the CEO of EMS Copenhagen.
With each dispatch, Corti's understanding of the emergency situations increase. Corti chief executive Andreas Cleve detailed one such story in Fast Company. A woman called for an emergency dispatch after she believed a man broke his back. However, Corti heard the patient's breathing and analyzed the rattling noises to determine that his heart had stopped and that the man had fallen because he'd gone into cardiac arrest. However, Corti was still in the learning phase of development and was unable to assist that particular dispatcher in the situation. The man did not survive, according to reports.
However, Cleve said now Corti is learning and growing, and the team is closer to getting the AI to partner with more dispatch teams around the world.
"I would always, especially when it comes to my health, prefer human contact. But augmented by a supportive system that might be using AI - that, to me, is sort of an end-game scenario."