SBIR Phase I: Project PAIR: Optimized Managed Care Through Personalized AI for Individuals in Recovery

这项小企业创新研究（SBIR）第一阶段项目的更广泛影响/商业潜力是开发一种远程监测系统，可以提醒护理人员注意阿片类药物使用的复发。超过2300万美国人对毒品和酒精上瘾，这些上瘾每年数十亿美元。大多数帮助人们保持康复状态的工具成功率很低或好坏参半。减少复发可以挽救生命和家庭，并减少再逮捕、再监禁和再住院。在本提案中，机器学习和模式识别这两种形式的人工智能（AI）将有助于识别和应对潜在的复发。好处包括节约应急资源，但更重要的是，提高长期干预的成功率。这个小企业创新研究（SBIR）第一阶段项目将利用生理和智能手机数据确定和预测未来渴望/痴迷或复发状态的可行性，这是一个生理基础警报改变当前护理协调工作流程的用例，也是一个可以显著避免住院康复出院后复发的用例。技术目标包括：1)设计一个新的数据驱动框架，使用个性化分类模型准确客观地估计重新使用阿片类药物的概率；2)通过管理医疗服务提供者的反馈，在成瘾治疗中心部署模型风险分层系统和监测仪表板，并评估其效果；3)评估和比较即时干预与标准实践的渴望与预测模型的效果。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

The broader impact /commercial potential of this Small Business Innovation Research (SBIR) Phase I project is to develop a remote monitoring system that can alert caregivers to relapses in opioid use. Over 23 million Americans are addicted to drugs and alcohol, and these addictions billions per year. Most tools to help people stay in recovery have low or mixed success rates. Reducing relapse saves lives and families and it reduces rearrests, reincarcerations, and rehospitalizations. In this proposal machine learning and pattern recognition, both forms of artificial intelligence (AI) will be aid identification of and response to potential relapse. Benefits include conserving emergency response resources, but more importantly, improving long-term intervention success.This Small Business Innovation Research (SBIR) Phase I project will establish the feasibility of identifying and predicting a future state of craving / obsession or relapse using physiological and smartphone data, a use-case where physiologically underpinned alerts alter current care coordination workflows, and a use-case where relapse after discharge from inpatient facilities for rehabilitation can be significantly averted. Technical objectives include: 1) devise a novel data-driven framework for accurately and objectively estimating probability for relapsing into opioid use using individualized classification models; 2) Deploy and assess efficacy of model risk stratification system and monitoring dashboard at addiction treatment centers through feedback from managed care providers; 3) Assess and compare efficacy of craving vs. prediction models for just-in-time interventions vs. standard practices.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.