About gestimation/Shiro Tanaka

I’m a Professor of Clinical Biostatistics at Kyoto University Graduate School of Medicine. This site offers story-driven, reproducible notes on clinical research, statistics, and causal thinking—written for researchers and clinicians who want the ideas to feel usable in practice.

I have served as a trial statistician in nationwide clinical trial groups (JCOG, JCCG, and A-TOP) and as an advisory expert for the PMDA. I’ve been fortunate to learn through collaborations with researchers such as Marc A. Brookhart, Jason P. Fine, and Thomas H. Scheike. This site is my small way of giving back to the community that shaped my training—through books, open-source software, and freely shared resources.

Why “gestimation”?

The name comes from g-estimation in causal inference—a class of methods introduced by James Robins. In my own work, I have found this approach especially useful in applied clinical research:

• Tanaka S, Matsuyama Y, Shiraki M, Ohashi Y. Estimating the effects of time-varying osteoporosis treatments on incidence of fractures among Japanese postmenopausal women. Epidemiology 2007;18(5):529–36.

• Tanaka S, Brookhart MA, Fine JP. G-estimation of structural nested mean models for competing risks data using pseudo-observations. Biostatistics 2020;21(4):860–75.

• Tanaka S, Brookhart MA, Fine JP. G-estimation of structural nested mean models for interval-censored data using pseudo-observations. Statistics in Medicine 2023;42(21):3877–91.

These are not required reading for the stories here. But if you ever feel like following the path from coffee chats → DAGs → g-estimation, this is one of the places it eventually leads.