Artificial intelligence-enabled, real-time risk monitoring for substance use recovery in Veterans: An initial evaluation of the Behaivior Recovery Platform

Abstract: INTRODUCTION: Substance use disorders (SUDs) are a growing concern in U.S. Military populations, including Veterans, reflecting both national trends and unique service-related risk factors. Behaivior's cloud-based predictive analytics system integrates artificial intelligence (AI) with smartwatches to predict and intervene in SUD-related health crises. Early success in civilian populations underscores its potential to support SUD recovery in Veterans. By integrating continuous monitoring with timely, context-specific support, the platform can foster sustainable behavior change, mitigate risks, and improve treatment adherence. Accessible via the World Wide Web and smartphone applications, the platform uses AI-driven algorithms to analyze biometric data, informing the delivery of tailored interventions and provider alerts. MATERIALS AND METHODS: A user experience study including 10 Veterans with cocaine use disorder and a mobile health intervention study involving 27 Veterans with opioid use disorder were conducted. Both were IRB-approved through the VA Pittsburgh Healthcare System. The user experience study assessed the acceptability and desirability of mobile health interventions, leveraging methods including surveys and interviews. The intervention study evaluated preliminary engagement and efficacy signals over a 4-week trial of the Behaivior platform, capturing self-report data and wearable-monitored outcomes. RESULTS: In the user experience study, all participants favored mobile health interventions for SUD treatment, with an average of 82.9% of proposed wearable device-enabled functions receiving positive endorsements. In the intervention study, engagement and adherence with wearable data collection were sufficient, with participants completing over 8 hours of remote physiological monitoring per day on average and completing most surveys. Substance use outcomes revealed a 29.6% return-to-opioid-use rate, with only 3.3% of specimens testing positive for opioids. CONCLUSION: Behaivior's technology demonstrates strong potential to support SUD recovery by synthesizing real-time metrics, proactively identifying risk states, and informing provider decision-making. This platform holds promise for enhancing the health of Veterans and may also have potential to benefit active duty service members.