This CSSI Seminar is jointly hosted with the Statistics and Data Science Seminar. Note unusual time.
James Robins, Harvard Chan School of Public Health
Target Trials and Structural Nested Models: Emulating RCTs using Observational Longitudinal Data
Abstract: Target trials are RCTs one would like to conduct but cannot for ethical, financial, and/or logistical reasons. Consequently, we must emulate such trials from observational data. A novel aspect of target trial methodology is that, for purposes of data analysis, each subject in the observational study is ‘enrolled’ in all target trials for which the subject is eligible, instead of a single trial. I will describe recent theoretical results connecting target trial methodology and structural nested models. I will discuss a novel inferential conundrum that arises from this connection. Finally I will discuss the question: How do we validate causal estimates from observational data. Based on joint research with Fuyu Guo, Oliver Dukes, and Mats Stensrud.
Bio: James Robins is LaFoley Dong Professor of Epidemiology at the TH Chan School of Public Health. Robins is widely recognized as a founding figure in the modern statistical field of causal inference. Beginning with his 1986 paper “A new approach to causal inference in mortality studies with a sustained exposure period—application to control of the healthy worker survivor effect,” he introduced some of the most important analytic methods and concepts in the field, including those collectively referred to as “g-methods,” used for drawing causal inferences from complex observational and randomized studies with time-varying treatments or exposure. Over a career of more than 40 years, Robins has applied his methods to research in a variety of topics in epidemiology, public health, and the social sciences, such as HIV/AIDS treatment, arsenic exposure, and bariatric surgery.
