9 NOV 2026 — Two randomised trials of smartphone apps that summon CPR-trained volunteers to cardiac arrests landed on opposite sides when they were presented on 7 November in Chicago, at a resuscitation science symposium run by the American Heart Association. In British Columbia and Columbus, Ohio, switching on PulsePoint Respond alerts did not raise the share of patients who got bystander CPR or a defibrillator before paramedics arrived. In Freiburg, Germany, a volunteer reached the patient before the ambulance almost twice as often when the system picked who to alert by estimated travel time instead of a fixed 600-metre circle.
Both are early results. The North American trial is a conference abstract, which the association says is not peer-reviewed and should be treated as preliminary. The Freiburg trial, called ALARM, was published the same day as a full paper in the Journal of the American Heart Association.
The five arms side by side
The trials measured different things, so read down each trial, not across them.
| Trial arm | Who gets the alert | Outcome measured | Result |
|---|---|---|---|
| PulsePoint, alerts on | App users near a public-place arrest, alerted alongside the ambulance | Bystander CPR or AED use before professionals arrived | 123 of 178 (69%) |
| PulsePoint, alerts suppressed | No one; standard dispatch only | Same | 127 of 177 (72%) |
| Freiburg, 600 m circle | Volunteers within about 600 metres | A volunteer arrived before the ambulance | 31% |
| Freiburg, 2,500 m circle | Volunteers within about 2,500 metres | Same | 57% |
| Freiburg, smart alerting | Volunteers expected to beat the ambulance, by estimated travel time | Same | 58% |
A high baseline left little room
The PulsePoint trial randomly assigned 17,547 emergency calls for possible cardiac arrest in public places, from 8 June 2021 to 27 March 2026. Only 355 made the primary analysis, because a case counted only if paramedics treated a confirmed arrest with at least one app user within 400 metres at dispatch. By our arithmetic that is about one call in 49.
Without alerts, 73% of patients already received bystander CPR. "This unexpectedly high rate of bystander CPR may have limited the opportunity for alerted responders to provide additional benefit," said Steven Brooks, the senior lead author, of Queen's University in Kingston, Ontario. About 21% of patients in each group survived to hospital discharge. The volunteer pool was also thin. When alerts went out, 57% of incidents had a single app user inside 400 metres.
The release says the result does not settle what happens in private homes, where most out-of-hospital arrests occur, or in places where ambulances take longer.
Freiburg tested the radius itself
In the Freiburg study, the region's alerting system switched at random each day between the three strategies from 1 January to 31 July 2024. Neither dispatchers nor volunteers knew which was running. Arrival times came from the volunteers' phone GPS, and the system could assign at most four volunteers to one emergency.
A 2,500-metre circle covers about 19.6 square kilometres against about 1.1 for a 600-metre one, roughly 17 times the area by our calculation. Widening the circle raised arrivals before the ambulance from 31% to 57%.
Smart alerting matched the wide circle, with a 1.2 percentage point difference that the paper reports as not statistically significant, while sending fewer people. Under smart alerting, 69% of volunteers who accepted were given a task, against 94% under the wide circle. It held back 31% of volunteers who had accepted an alert because they would probably arrive after the ambulance. "A circle on a map simply cannot capture this," said Michael Mueller, the senior author, of St. Josef's Hospital in Freiburg.
The paper reports 823 activations and 1,482 volunteers dispatched over the 213-day trial. That works out to about 1.8 volunteers per activation and just under four activations a day. It did not measure survival, and Mueller said that question needs further research.
Singapore's app is also built on circles
Neither trial included Singapore, where the Singapore Civil Defence Force runs myResponder. HTX, which co-developed the app, says the first version from 2015 alerted volunteers to suspected cardiac arrests within 400 metres. The version launched on 4 July 2024 lets volunteers who choose to cycle or drive take alerts up to 800 metres or 1.5 kilometres away, and shows the expected time to reach the patient by each mode. A 1.5-kilometre circle covers about 14 times the area of a 400-metre one. HTX said 76 cardiac arrest patients had been saved since 2015 by volunteers activated through the app.
The SCDF's App Store listing describes alerts to "nearby fire and medical cases" and opens registration to people aged 15 and above. Neither source says whether myResponder chooses responders by estimated travel time, as Freiburg's smart system did.
View author profile → · Editorial policy
