1112 Performance Evaluation of a Multisensor Ring During a Clinical Multiple Sleep Latency Test

Abstract Introduction Commercially-available wearable devices are increasingly used to detect sleep-wake patterns in free-living conditions. These devices, however, are infrequently assessed relative to their ability to detect naps. The present study evaluated whether a commercially-available wearable device was able to detect nap onset during a clinical Multiple Sleep Latency Test (MSLT). Methods Participants referred for a clinical MSLT wore a commercially-available multisensor ring device (Happy Ring; Happy Health) during their in-lab assessment at one of two AASM-accredited sleep centers. Each sleep lab followed AASM guidelines for administering the MSLT and scoring the polysomnogram for sleep onset. Due to potential differences in scoring across sites, a third reviewer (INSERT CREDENTIALS HERE) blindly scored all records, also according to AASM guidelines. Results Internal sleep lab scoring somewhat differed from the independent scorer, with 90% agreement regarding the proportion of naps that were clinically significant; there were 2 cases where only the sleep lab reported a clinically significant MSLT and 3 cases where only the independent rater reported a clinically significant MSLT. The ring achieved 86% diagnostic accuracy relative to the lab scorer and 84% relative to the independent scorer. The mean difference between the lab and independent scorer for sleep latency was 142 seconds, with N=30 records that achieved agreement of < 150 seconds. Of these 30 records, the Happy Ring achieved 96.7% accuracy (only 1 case of disagreement). The mean difference between the lab scorer and the ring was 152 seconds. Conclusion The Happy Ring shows promise in its ability to detect sleep onset during an MSLT. Future work could be used to optimize wearables for being able to predict who will test positive in an MSLT, thus reducing costs, wait times, and patient burden. Support (if any) Happy Health