Apple's Heart Study, announced with Stanford in Nov. 2017, finds ~0.5% of 400K+ participants using Apple Watch up to Series 3 received irregular pulse alerts
Chris Welch / The Verge :
Context & Ripple Effects
The results close the loop on the heart research app Apple launched with Stanford in November 2017: of the 400K-plus participants running it on Apple Watches up to Series 3, about half a percent received an irregular pulse notification. That yield sits on top of earlier evidence that the hardware could see arrhythmia at all — a [[a:918922|UCSF and Cardiogram study had already shown the Watch detecting atrial fibrillation with 97% accuracy]] via deep neural network analysis.
First-order effects
- Roughly 2,000 of the 400K+ participants now hold an irregular pulse alert that points them toward follow-up care, making this one of the largest single screenings for atrial fibrillation risk ever run through a consumer device.
Second-order effects
- The low alert rate cuts both ways for Apple: it supports the case for the Watch as a screening tool, but subsequent reporting that the device largely misses AFib above 120 beats per minute gives clinicians grounds to question whether a clean result means a clean bill of health.
Third-order effects
- Academic partners are already recalibrating — a Harvard biostatistician has been [[a:969012|reconsidering plans to use the Apple Watch in a study after finding inconsistencies in its heart rate variability data]] — so the structural question is whether consumer wearables become standard research instruments or require clinical-grade validation first.
The trend: Consumer wearables are scaling into population-scale medical screening platforms, with each large study exposing accuracy limits that determine how far medicine will trust them.