0178 2B-Alert App: Personalized Caffeine Recommendations for Optimal Alertness

Abstract

Abstract Introduction Sleep loss causes millions of individuals around the world to carry out daily activities with impaired alertness, affecting injury risk and productivity. If properly consumed, caffeine can safely and effectively mitigate the effects of sleep loss on alertness. However, to be effective, caffeine should be consumed at the right time and in the correct amount, depending on sleep history, work schedule and, importantly, sensitivity to sleep loss. Here, we present the 2B-Alert app, a unique smartphone application with capabilities to learn an individual’s trait-like response to sleep loss and to provide personalized caffeine recommendations to optimize alertness. Methods We conducted a prospective 62-h total sleep deprivation study to validate the capabilities of 2B-Alert. Throughout the challenge, 21 participants used the app to measure their alertness via the psychomotor vigilance test (PVT). The app used the PVT data collected during the first 36 h of wakefulness to learn each participant’s sleep-loss response and provided on-the-fly personalized caffeine recommendations (from 0 to 800 mg), so that each participant would sustain alertness at a pre-specified target level (mean PVT response time of 270 ms) during a 6-h period starting at 44 h into the challenge. To assess the effectiveness of the personalized recommendations, we computed the amount of PVT data during the 6-h period that fell within the target alertness level +/-2 times the within-subject variability of alertness impairment (i.e., +/-60 ms). Results We observed a wide range of responses to sleep loss, with some participants displaying high levels of resilience and others high levels of vulnerability. Accordingly, 2B-Alert recommended no caffeine to five participants, 100-400 mg to 11 participants, and 500-800 mg to five participants. Regardless of their sleep-loss response and the consumed amount of caffeine, participants sustained the target alertness level ~80% of the time during the 6-h period. Conclusion 2B-Alert automatically learns an individual’s trait-like response to sleep loss and provides personalized caffeine recommendations in real time so that individuals achieve a desired alertness level regardless of their susceptibility to sleep loss. Support (if any)