Apple Watch App Achieves 98 Percent Accuracy in Detecting Seizures
Baltimore, Thursday, 28 May 2026.
A landmark study shows the EpiWatch Apple Watch app detects severe seizures with 98% accuracy, slashing false alarms by 90% and transforming everyday wearables into reliable medical monitors.
Bridging Consumer Tech and Clinical Utility
On May 27, 2026, EpiWatch, Inc. announced the publication of its Phase III clinical study in Neurology Open Access, formally validating its seizure detection application for use on Apple Inc. (AAPL) [GPT] devices. The software is designed to detect tonic-clonic seizures (TCS) in pediatric and adult populations aged five and older [1]. The prospective, multicenter trial monitored 242 participants, who had an average age of 23, across six epilepsy monitoring units [1][2][3]. Over an average observation period of 2.5 days per patient, the study accumulated more than 16,000 hours of video-EEG clinical monitoring [2][3].
Redefining Reliability with Reduced False Alarms
A persistent barrier to the adoption of wearable seizure monitors has been the prevalence of false alarms, which can lead to device abandonment by frustrated patients and caregivers. Over the 16,000-hour monitoring period, the EpiWatch app logged only 56 false alarms [2][3]. This equates to a False Alarm Rate (FAR) of 0.08 per day, or approximately one false alert every 12.4 days [1][2]. This rate is 90% lower than that of competing monitors, which typically trigger between 0.67 and 2.52 false alarms daily [2]. During the trial, 87% of participants experienced zero false alarms, and of the 56 total false alerts recorded, 35 were triggered by repetitive arm movements such as playing video games [2][3].
Future Commercial and Clinical Horizons
While the Phase III study reported zero adverse effects, researchers noted that a primary limitation of the trial was its execution within the controlled environment of epilepsy monitoring units [2]. To address this, ongoing real-world analyses are currently evaluating the application’s performance in everyday settings [1]. Looking ahead, EpiWatch intends to submit its clinical data to the US Food and Drug Administration (FDA) for 510(k) clearance by December 2026 [5] [alert! ‘Source 5 promotional text cites 1,500 participants and 93.7% sensitivity, which conflicts with the peer-reviewed data of 242 participants and 98% sensitivity; the peer-reviewed figures are prioritized here’]. If cleared, this technology could substantially expand the commercial footprint of medical-grade applications within the broader consumer electronics market [GPT].