Authors :
Presenting Author: Trevor Meyer, BS – Neurava Inc.
Swagat Bhattacharyya, BS – Neurava Inc.; Patrick Lehman, BS – Neurava Inc.; Vivek Ganesh, PhD – Neurava Inc.; Joseph Ta, MD – Vanderbilt University Medical Center; Kelly Lowen, CCRP – Vanderbilt University Medical Center; Deidre Dragon, BS – University of Iowa; Rup Sainju, MBBS – University of Iowa; Brian Gehlbach, MD – University of Iowa; Jay Shah, PhD – Neurava Inc.; William Nobis, MD, PhD – Vanderbilt University Medical Center; George Richerson, MD, PhD – University of Iowa
Rationale:
Due to the heightened risk of sudden unexpected death in epilepsy (SUDEP) associated with generalized tonic-clonic seizures (GTCS), patients who experience frequent GTCS are thought to be a particularly vulnerable population. There is a critical need for wearable devices capable of accurately detecting these seizures to reduce the risk of life-threatening incidents and their repercussions. Such a wearable can also keep an accurate record of GTCSs to assist physicians in optimizing treatment. We have developed a wearable system for monitoring an array of physiological signals, including a multi-modal arm wearable for GTCS detection. Here we report on the overall performance of the arm wearable in people with epilepsy.
Methods:
We conducted a multicenter, prospective study using the Neurava arm wearable in the epilepsy monitoring units (EMUs) at Vanderbilt University Medical Center and University of Iowa. The subjects were admitted to the EMUs as part of their normal clinical care. We included people older than 18 years with confirmed epilepsy and highly suspected of having GTCSs. The primary outcome of the study was the validation of the arm wearable and its detection performance of GTCSs relative to gold-standard, time-synced video-EEG data with physician annotations. The secondary outcome was usability scores detailing the subjects’ experiences with the device, obtained by asking the subjects to complete a questionnaire based on a 5-point Likert scale.
Results:
We placed the arm wearable on 30 adults (63.3% female, mean age = 33.6 years) and collected over 1300 hours of data, including nine GTCSs. A total of 21 subjects did not experience a GTCS. All subjects completed the study. Preliminary results indicate a positive percent agreement (PPA) of 87.5% and a false alarm rate (FAR) of 0.984 per day using subject-hold-out cross-validation and an adjustable threshold. These interim metrics are consistent with other seizure monitoring devices available on the market. On average across all subjects, they agreed (score > 3.0) that the arm wearable was comfortable (score = 4.0), stable and secure (score = 3.5), did not interfere with normal activities like sleeping (score = 4.3), and that they would wear the device to sleep (score = 4.2) and would recommend it to others (score = 4.2).
Conclusions:
While the algorithm development and analysis efforts are still ongoing, these preliminary results, along with the usability data, demonstrate that the Neurava arm wearable can be a viable device for reliable GTCS detection and tracking. The arm wearable, along with the rest of Neurava’s full system, will provide people with epilepsy, their physicians, and caregivers with accurate multi-modal physiological data and GTCS detection and alerting. This full system may help reduce the risk of life-threatening events and potentially mitigate the risk of SUDEP.
Funding:
This study was funded by Neurava Inc.