Carreiro S, Ramanand P, Taylor M, et al. Evaluation of a digital tool for detecting stress and craving in SUD recovery: An observational trial of accuracy and engagement. Drug Alcohol Depend. 2024;261:111353. doi:10.1016/j.drugalcdep.2024.111353
This observational study evaluated digital stress and craving detection of the Realize, Analyze, Engage (RAE) app when used on a commercial-grade wrist sensor. The RAE system combines a wearable device, a mobile app, and a clinician portal to provide substance use recovery support. The system is designed to detect stress and craving in the user and then automatically send interventions to the user, and notifications as necessary, to the clinician. In this study, previous validation data of the RAE system using research-grade wearable devices was compared to commercial-grade wearable devices. Here, the RAE system was modified for use with a Garmin Vivosmart 4, and a novel biomarker model was developed. Adults were recruited from outpatient SUD treatment centers (N=60) and asked to wear the provided Garmin device for 30 days and provide real-time feedback when they experienced stress or a craving in the RAE mobile app. From these participants, 4359 hours of sensor data with 8873 stress events and 186 craving events were analyzed. The most common reasons participants failed to report stress events were inconvenient timing, not having access to their phone, and forgetting to report. Attrition rates were highest for participants whose phones had simple/less expensive operating systems. It may be that these were ill-equipped to manage the RAE app, leading participants to drop out due to a dampened user experience. Despite these challenges, the model developed from the Garmin device was comparable in its ability to predict stress and craving events compared to the research-grade device data. Future research is needed to determine the efficacy of RAE in protecting users from relapse and assessing the user feedback on new features such as engagement tracking, stress and craving trend visuals, connectivity confirmations, and disconnection notifications.