Epidemiologic Inference with Mechanistic Models: Merging the ‘Why’ with the ‘How’
With the exception of infectious disease epidemiology, associative statistical models have dominated epidemiologic research. However, mechanistic models, both computational and mathematical, have a rich tradition in other sciences, ranging from ecology to physics. Computational power and methodological innovations now allow mechanistic and semi-mechanistic models to be used for proper statistical inference. However, these approaches have not been widely adopted in epidemiology. This symposium aims to stimulate epidemiologic inference with mechanistic models. We will start by contrasting the epidemiologic approach to inference vis-à-vis other disciplines and examine the historical origins of our siloed approach. Next, applications of mechanistic models to infectious disease and social epidemiology will be presented, along with ecological approaches. Finally, the session will conclude with a panel discussion on possible new frontiers, as well as pitfalls and limitations.
Justin Lessler, Johns Hopkins Bloomberg School of Public Health
Arijit Nandi, McGill University
Thomas Glass, Johns Hopkins Bloomberg School of Public Health
Eleanor Murray, Harvard TH Chan School of Public Health
Pejman Rohani, University of Georgia
Daniel Westreich, University of North Carolina, Chapel Hill