Abstract
OBJECTIVE: This study aimed to evaluate the real-world performance of a wrist-worn seizure detection device for timely clinical interventions within an epilepsy monitoring unit (EMU). METHODS: We conducted a prospective observational study involving patients admitted to the EMU at a tertiary care center. Participants wore the Embrace2 device, which has US Food and Drug Administration approval for purchase and continuously monitored for convulsive seizures. The device was connected to the hospital's public Wi-Fi and was paired with a smartphone programmed to alert nursing staff through automated messages and phone calls upon seizure detection. Nursing staff were trained on the device's use, and implementation outcomes were documented, including technical performance, alert responses, and staff feedback. RESULTS: A total of 72 patients (57% female, mean age = 39 years, range = 19-81) were prospectively recruited between April 2024 and June 2025 at the Centre hospitalier de l'Université de Montréal EMU. The analysis included 373 monitoring days (mean = 5.1 days/patient). Eighteen focal to bilateral tonic-clonic seizures (FBTCSs), one generalized tonic-clonic seizure (GTCS), and 510 focal seizures (FSs) occurred. The Embrace2 wrist-worn device detected all 16 FBTCSs/GTCSs that occurred while the device was functioning and worn, but none of 510 FSs. Alerts successfully reached caregivers in 15 of 16 cases, with one failure due to a Wi-Fi disconnection. Seven of the successful alerts occurred during overnight shifts when electroencephalographically trained technicians are typically not on duty. Twenty-nine false alarms occurred (.077/24 h), mostly triggered by routine movements (e.g., toothbrushing). Practical challenges included the potential for missed alerts when nurses were not carrying their phones; however, no seizure events were missed for this reason during the study period. SIGNIFICANCE: Wearable devices show considerable value for seizure detection and triggering timely interventions in EMU settings, particularly during off-hours. However, successful integration requires robust coordination, reliable infrastructure, and staff engagement to maximize clinical benefit and maintain trust in the system.