Optimizing Just-in-Time Adaptive Intervention to Improve Dietary Adherence in Behavioral Obesity Treatment: A Micro-randomized Trial

项目总结/摘要 行为肥胖治疗产生临床上显著的体重减轻和降低的疾病风险/严重程度， 许多人超重/肥胖和心血管疾病。然而，大约一半的患者没有达到 预期的结果，这在很大程度上可以归因于从推荐的饮食失误。我们的工作 饮食失误（不遵守饮食目标的具体情况）在减肥过程中很常见， 尝试（每周约3 - 4次），与减肥效果较差相关，并由瞬间改变触发 状态（例如，情绪的变化或可口食物的可用性）。因此，显然需要创新。 解决方案，可以提供动态的即时干预，以提高遵守规定的饮食， 肥胖治疗我们的研究团队是第一个开发基于智能手机的即时自适应 干预措施（JITAI）包括：1）每日生态瞬时评估（EMA;通过 移动终端）的相关行为，心理和环境触发失效; 2）机器 学习算法，使用通过EMA收集的信息来确定实时失效风险;以及3）交付 在高风险时刻进行短暂干预。我们的试点工作表明，吉泰是可行的，可以接受的， 并降低了平均延时频率。然而，我们还没有显示出直接的影响， JITAI对失智风险升高时的饮食行为的认识不足，对干预措施的类型知之甚少 最能有效地减少失误因此，我们建议通过一个微- 随机试验（MRT），一种方法，涉及随机分配干预（或对照）， 具体的决策点，即，当我们的算法预测失误的风险增加时。捷运局将决定 在特定时刻的特定干预是否达到了预期效果。因此，我们将把我们的吉泰港， 一个更具可扩展性的在线平台，并进行MRT，以评估通用失效风险警报的影响 信息和理论驱动的饮食失误及时干预。在n = 15的细化测试后， 确保我们更新的JITAI技术功能正常，超重/肥胖成人（n = 159）将参与 在一个完善的12周在线肥胖治疗计划+JITAI中，有12周的JITAI随访。 当一个人处于失智风险时，他/她将被随机分配到不干预、一般风险警报或一个 4理论驱动的干预与互动技能培训。关注的结果将是发生 (or缺乏）的饮食失误，如主观测量的（即，通过EMA）和客观地（即，通过手腕- 基于摄入量监测）。MRT的结果将为优化的 干预交付的算法，将推动最终的JITAI。未来的随机对照试验将比较 使用和不使用优化的JITAI治疗肥胖。这种高度创新的方法将推动 通过支持发展复杂的粘附理论模型， 行为和告知JITAI目标是坚持其他健康行为（例如，药物，活动目标）。

PROJECT SUMMARY/ABSTRACT Behavioral obesity treatment produces clinically significant weight loss and reduced disease risk/severity for many individuals with overweight/obesity and cardiovascular disease. Yet, about half of patients fall short of expected outcomes, which can be largely attributed to lapses from the recommended diet. Our work has shown that dietary lapses (specific instances of nonadherence to dietary goals) are frequent during weight loss attempts (~3-4 times per week), associated with poorer weight losses, and triggered by momentary changing states (e.g., changes in mood or availability of palatable food). Thus, there is a clear need for innovative solutions that can provide dynamic in-the-moment interventions to improve adherence to the prescribed diet in obesity treatment. Our research team was the first to develop a smartphone-based just-in-time adaptive intervention (JITAI) that includes: 1) daily ecological momentary assessment (EMA; repeated sampling via mobile device) of relevant behavioral, psychological, and environmental triggers for lapse; 2) a machine learning algorithm that uses information gathered via EMA to determine real-time lapse risk; & 3) delivery of brief intervention during high-risk moments. Our pilot work revealed that the JITAI was feasible, acceptable, and produced reductions in average lapse frequency. However, we have not yet shown a direct effect of the JITAI on eating behavior in the moment of heightened lapse risk and know little about the types of interventions that are most effective for reducing lapse. We therefore propose to extend our research via a micro- randomized trial (MRT), a methodology that involves random assignment to intervention (or control) at a specific decision point, i.e., when our algorithm predicts heightened risk for a lapse. The MRT will determine whether a specific intervention in a specific moment had its intended effect. We will therefore port our JITAI to a more scalable online platform and conduct a MRT to evaluate the effects of a generic lapse risk alert message and theory-driven just-in-time interventions on dietary lapses. After refinement testing with n=15 to ensure proper technical functioning of our updated JITAI, adults with overweight/obesity (n=159) will participate in a well-established 12-week online obesity treatment program + JITAI, with 12 weeks of JITAI-only follow-up. When an individual is at risk for lapsing s/he will be randomized to no intervention, a generic risk alert, or one of 4 theory-driven interventions with interactive skills training. The outcome of interest will be the occurrence (or lack thereof) of dietary lapse, as measured both subjectively (i.e., via EMA) and objectively (i.e., via wrist- based intake monitoring), in the hours following randomization. Results of the MRT will inform an optimized algorithm for intervention delivery that will drive the finalized JITAI. A future RCT will compare weight loss in obesity treatment with and without the optimized JITAI. This highly innovative approach will advance the science of adherence by supporting the development of sophisticated theoretical models of adherence behavior and informing JITAIs that target adherence to other health behaviors (e.g., medication, activity goals).