AI&DHI Innovation Spotlight: Building Effective and Reliable Digital Health Intervention Systems to Improve Mental Health and Well-being
This month’s AI&DHI Innovation Spotlight focuses on work led by Dr. Zhenke Wu, Associate Professor of Biostatistics in the School of Public Health.
With AI&DHI’s support, Dr. Wu’s team is building reliable and effective digital intervention systems aimed at supporting the mental health of medical interns—a group particularly vulnerable to stress and burnout.
Regular physical activity and quality sleep are key to reducing the risk of conditions such as cardiovascular disease, obesity and depression. Yet, statistics show that only a quarter of adults in the United States are getting the recommended level of physical activity, and many are struggling to achieve healthy sleep. This is especially true for physicians in medical internships. These demanding training programs often leave physicians exhausted and with little time for traditional mental health support. Complementarily, among patients seeking mental health care, the problem is worsened by long wait times for mental health services.
To tackle these challenges, Dr. Wu’s team is leveraging recent breakthroughs in wearable devices and smartphones. These digital tools can deliver low-burden, personalized mental health interventions right at users’ fingertips, enabling them to provide support that’s both immediate and tailored to a person’s individual needs.
“By combining clinical expertise with digital technology and AI, we are setting the stage for smarter, more personalized care.”
Associate Professor, Biostatistics
School of Public Health
Over the course of a seven-year collaboration with the Intern Health Study—an NIH-funded longitudinal cohort study that assesses stress and mood in medical interns at institutions around the US and China—Dr. Wu’s team has engaged more than 10,000 new physicians in micro-randomized trials since July 2018, testing the effectiveness and timing of smartphone-delivered interventions such as physical activity prompts and mood tracking. Notably, they introduced gamified elements—like team-based competitions—to improve App and intervention engagement. They are also working within a larger cross-disciplinary team to deploy advanced AI methods, such as reinforcement learning (RL), to personalize support based on real-time behavioral and physiological data. The team envisions their next steps will be to deploy, monitor, evaluate, and improve these algorithms in practice by leveraging the critical resources and talents in the digital health space at Michigan, notably at AI&DHI and the Depression Center.
“The development of RL methods to use real-time behavioral and physiological data for optimizing mental and behavioral health treatments represents an initial step towards a broader application of machine learning in healthcare settings, aiming to inspire wider adoption, monitoring, and evaluation of such innovative approaches,” said Dr. Wu. "By combining clinical expertise with digital technology and AI, we are setting the stage for smarter, more personalized care.”
How AI&DHI Supported This Work
AI&DHI facilitated the long-term collaboration between Dr. Wu’s lab and the Intern Health Study team, enabling them to conduct methodological research in causal inference and reinforcement learning aimed at improving mental health and well-being. The findings from this work are now actively being used to inform new intervention designs and implementations for future cohorts as well as for other ongoing digital health intervention studies. Additionally, this work facilitated the training and professional development of two Biostatistics PhD students and has also laid the groundwork for a larger grant proposal focused on designing and implementing digital mental health intervention systems using reinforcement learning.
Relevant Publications
NeCamp T, Sen S, Frank E, Walton M, Ionides E, Fang Y, Tewari A, Wu Z (2020). Assessing real-time moderation for developing adaptive mobile health interventions for medical interns: micro-randomized trial. Journal of Medical Internet Research (JMIR) 22(3): e15033. doi: 10.2196/15033. PMID: 32229469
Wang J, Fang Y, Frank E, Walton MA, Burmester M, Tewari A, Dempsey W, NeCamp T, Sen S, Wu Z (2023). Effectiveness of gamified competition in the context of mHealth intervention for medical interns: a micro-randomized trial. npj Digital Medicine. doi:10.1038/s41746-022-00715-5
Wang J, Shi C, Wu Z (2023). A Robust Test for the Stationarity Assumption in Sequential Decision Making. 40th International Conference on Machine Learning (ICML).
Li M, Shi C, Wu Z, Fryzlewicz P (2025). Testing stationarity and change point detection in reinforcement learning. Annals of Statistics. 53(3): 1230-1256.