Professor Miguel Hernán uses health data and causal inference methods to learn what works. As Director of the CAUSALab at Harvard, he and his collaborators repurpose real world data into scientific evidence for the prevention and treatment of infectious diseases, cancer, cardiovascular disease, and mental illness. His work shapes health policy and research methodology worldwide.

Professor Hernán joined the Harvard School of Public Health in 1999, becoming a professor in 2011 before being appointed Kolokotrones Professor of Biostatistics and Epidemiology, Harvard T.H. Chan School of Public Health in 2016. In 2021, Prof Hernán was named Director, CAUSALab, Harvard T.H. Chan School of Public Health, and he is also an Associate Member, Broad Institute of MIT and Harvard.

He is currently as Associate Editor of Annals of Internal Medicine, and Editor Emeritus of Epidemiology, and previously as Associate Editor of Biometrics, American Journal of Epidemiology, and Journal of the American Statistical Association.

Professor Hernán has been awarded the Rousseeuw Prize for Statistics, Rothman Epidemiology Prize, and the MERIT award from the National Institutes of Health, and has been elected Fellow of the American Association for the Advancement of Science and of the American Statistical Associationne.

Seminar Overview:

When randomized trials are not available, the causal effects of medical interventions are often estimated using large healthcare databases. Causal inference from these observational data can then be viewed as an attempt to emulate a hypothetical randomized trial—the target trial—that would quantify the causal effect of interest. The seminar will start with an outline of the two steps of the target trial framework: 1) specification of the protocol of the target trial, and 2) explicit emulation of the protocol’s components.

We will then explain why, contrary to what is generally believed, many well-known failures of observational studies for causal inference were the result of not adequately emulating a target trial rather than limitations of the observational data. We will describe those methodological failures, including immortal time and selection biases, in non-technical language and describe how a systematic application of the target trial framework prevents them. Examples from medical interventions for the prevention of cardiovascular disease and cancer will be used