Early recognition and intervention are critical for patient survival. High quality CPR and defibrillation by AED prior to EMS arrival improves survival after OHCA. Therefore, the chance of surviving OHCA is highly correlated with bystander and medical dispatchers’ recognition of the condition during the emergency calls.
At EMS-Copenhagen we have been developing and testing a computer with AI, which can ‘listen in’ on the call to 1-1-2 and find patterns in the calls that are indicative for OHCA. Retrospective studies have shown, that such a computer can recognise significantly more OHCA than medically trained dispatchers and do this faster. (Blomberg, Stig Nikolaj, et al. “Machine learning as a supportive tool to recognize cardiac arrest in emergency calls.” Resuscitation (2019).)
The technology is presently being tested in an RCT-study, and preliminary results are encouraging.
Perspectives on using AI in dispatch is promising, and Copenhagen-EMS are now investigating the use of AI on other diagnoses such as stroke and sepsis. Also, a more general approach as a triage-tool is being investigated.