Microrandomized Trial to Optimize Use of Burden-reducing Self-monitoring Approaches in Behavioral Obesity Treatment

This clinical trial is a 24-week micro-randomized trial (MRT) designed to optimize the use of dietary self-monitoring (SM) strategies during behavioral obesity treatment (BOT). Self-monitoring is considered the cornerstone of BOT because it enables individuals to regulate behaviors that influence energy balance and weight loss. However, adherence to traditional full dietary SM (i.e., recording all food and beverages consumed each day) declines rapidly over time due to its burden, leading to poorer weight loss outcomes. To address this problem, this study evaluates five SM approaches that vary in intensity and burden: 1. Full dietary SM: Recording all food and drink consumed, with calorie estimates, every day. 2. Reduced-frequency dietary SM: Recording all food and drink consumed, with calorie estimates, on 3 days per week. 3. Self-monitoring of dietary lapses only: Recording only episodes of eating or drinking that likely cause weight gain or put weight loss at risk. 4. Smartwatch-based SM of energy intake: Using a smartwatch that detects eating gestures and estimates energy intake (i.e., "calories") via built-in sensors. 5. Body weight SM only: Weighing daily using a smart scale that automatically syncs data to the online treatment platform. All participants (N=275 adults with overweight or obesity, aged 18-70) will receive a 24-week evidence-based, online BOT. This BOT includes interactive multimedia lessons on nutrition, physical activity, and behavioral skills for weight management, with individualized calorie and activity goals, self-regulation training, and automated feedback. The only experimental manipulation is the self-monitoring approach assigned every two weeks. At the start of treatment and every two weeks thereafter, participants will be randomly assigned to one of the five SM approaches (12 total randomizations per participant). Randomization probabilities favor assignment to full dietary SM at treatment initiation, then gradually equalize across approaches. At each randomization, participants will be notified of their new SM assignment and receive a brief online training for that approach. A member of the research team will conduct a short online video call to confirm comprehension and assist with troubleshooting. Participants will complete study assessments remotely at baseline and 24 weeks via online video call and online surveys. They will be provided with study equipment, including a smart scale and a smartwatch (participants may use their own smartwatch if it is compatible with the study platform). All SM data (e.g., SM logs, smartwatch data, smart scale readings) will be automatically transferred to the secure BOT system. Participants will also complete brief questionnaires at each randomization point assessing perceived burden, motivation, satisfaction, and engagement with the assigned SM approach. The co-primary outcomes are (1) SM adherence, defined as the number of days per two-week period that the participant completes ≥2 eating recordings (or weighs daily in the body-weight-only condition), and (2) objectively measured weight change (kg) via smart scale. Secondary outcomes include perceived burden, motivation, satisfaction with SM strategy, and platform engagement metrics. This MRT design enables the research team to compare short-term (2-week) effects of each SM approach on adherence and weight loss, identify how these effects vary over time and across individual characteristics, and develop a data-driven algorithm that adaptively selects optimal SM approaches for future treatment. Reinforcement learning will be applied to the MRT data to construct an adaptive SM-selection algorithm capable of personalizing SM recommendations based on individual differences, social determinants of health, and ongoing treatment response. This project will yield both broadly applicable recommendations (e.g., which SM approaches are generally most effective at different stages of treatment) and an adaptive algorithm to guide individualized SM strategies in future clinical and research settings. The findings are expected to enhance long-term adherence to self-monitoring and improve the effectiveness and scalability of behavioral obesity treatments delivered online.