Department of Health Promotion and Human Behavior,
Graduate School of Medicine, Kyoto University 京都大学 大学院医学研究科 健康増進・行動学分野

We aim to establish the evidence to promote health by applying cutting-edge causal inference and machine learning methods to randomized controlled trials and real-world data.

Specifically, our research focuses on implementing target trial emulation and evaluating treatment effect heterogeneity using machine learning approaches such as causal forests. In addition, we contribute to theoretical advancements, including the application of DAGs in causal mediation analysis (4-way decomposition) and the generalization and extension of the front-door criterion.

In addition to the conventional “high-risk approach,” we propose a “high-benefit approach” based on treatment effect heterogeneity, with the goal of informing the appropriate allocation of healthcare resources and reducing health disparities.

We also pursue practical research that bridges clinical medicine, public health, and social implementation. Our work includes optimizing interventions by integrating health behavior theories and behavioral economics, supporting behavior change through smartphone applications, and evaluating intervention effects through international collaborative studies.

Cardiovascular disease is the leading cause of death worldwide, with more than 300,000 deaths annually in Japan alone, making it one of the most critical challenges in medicine and public health.

Our lab has focused on epidemiological research about cardiovascular disease and its major modifiable risk factors—endocrine and metabolic disorders. By characterizing the global burden and exploring strategies for improvement, we aim to generate robust evidence that advances cardiometabolic health promotion.