LO
PilotCognitive automation2021-01-04
Alerting dispatchers with a machine-learning model did not significantly improve their recognition of cardiac arrest in emergency calls, a Copenhagen randomised trial found
Paramedic / EMToccupation page →
Event date / reported
2021-01-04
Evidence stage
PilotSmall-scale trial in a real setting. Tells us the deployment conditions are being tested, not that they hold — so one pilot is never enough on its own; two independent ones are.
Tasks this bears on
Assessing callers by phone
Assessing patients over the phone from the control room, giving advice, and deciding whether an ambulance is needed.
Being augmented✓ Evidence-backed
Where this applies
Copenhagen, Denmark. A double-masked randomised trial at the city's emergency medical services from September 2018 to December 2019. A speech-recognition model screened 169,049 calls and flagged 5,242 as suspected cardiac arrest; dispatchers in the intervention group received an alert. With alerts they recognised 93.1% of confirmed cardiac arrests against 90.5% without (P = .15). The model alone was more sensitive than dispatchers without alerts (85.0% vs 77.5%) but had a much lower positive predictive value (17.8% vs 55.8%). The authors conclude that machine learning support did not significantly improve dispatchers' recognition even though the AI surpassed human recognition.
What this means
Even an AI that hears cardiac arrest better than people did not, as an alert, change what people did.
What it does not yet show
One trial with dispatchers in one city; later versions and other services may differ.
What you can check
Open the trial's PubMed abstract (JAMA Network Open, PMID 33404620) and find "did not find any significant improvement".
Does it change the assessment?
No. The impact index is never moved by a single event, and this stage does not move one on its own: a "Pilot" record counts toward a judgement but needs a second, independent record before the judgement rests on evidence. This one is counted; on its own it changed nothing.
Source
Blomberg et al. — Effect of Machine Learning on Dispatcher Recognition of Out-of-Hospital Cardiac Arrest During Calls to Emergency Medical Services: A Randomized Clinical Trial, JAMA Network Open (4 January 2021) · verified 2026-09-30 · Claude (VOLO agent) · interpreted 2026-09-30 · Claude (VOLO agent)
Primary source — published by the party that did this, or the authority of record. No co-signature needed.