Artificial intelligence software identifies cardiac arrest instances

Corti has developed a new software to help recognise out-of-hospital instances of cardiac arrest “faster and more accurately” than a human.

The software that uses artificial intelligence has, so far, been tested in Copenhagen and will continue to be tested throughout Europe this summer as part of a partnership with the European Emergency Number Association (EENA).

To identify non-verbal and verbal patterns of communication, Corti’s software listens to emergency calls and uses algorithms.

The software studies the tone of a caller’s voice when they are less able to breathe, and prompts the dispatcher to ask certain questions, before guiding them to a decision about whether to send an ambulance or instruct someone to begin CPR.

Check out the demonstration of Corti’s Cardiac (Interesting Engineering):

Corti also revealed that the software has been tested on a database of 161,650 historical emergency calls and was able to accurately identify 93.1% of out-of-hospital cardiac arrests compared to 72.9% of human dispatchers.

Written by Leah Alger