Causal Inference in Epidemiology: What Was It, What Is It, and What Will It Become?
Summary: Epidemiology is centrally concerned with identifying causes of health and disease, so as to inform the search for effective interventions, either in public health policy or in the clinic. The epidemiology of the second half of the twentieth century saw the connection between a cause of disease and an effective intervention as a very loose one, with the intervention to be uncovered later after further biomedical research. By contrast, the first part of this century has witnessed a strong push to connect the notions of cause and intervention. This movement, often going under the misleadingly broad label 鈥渃ausal inference鈥, sees a very tight connection between interventions and causes, such that a causal question is not even well defined for the purposes of epidemiological research unless there is a well-specified intervention on that cause, against which the causal effect is measured. This movement is inspired in part by a pragmatic concern with achieving effective interventions and in part by the appeal of the powerful mathematical tools that can be used if causal questions are restricted in this way. It is the development and deployment of these tools that various recent workshops books on 鈥渃ausality鈥 (Pearl 2009) and 鈥渃ausal inference鈥 (Hern谩n and Robins 2015) focus upon. This technical focus hides the revolutionary nature of this new way of thinking about causality and causal inference. This workshop seeks to understand the conceptual framework of this movement, to place it in context against traditional epidemiological thinking, and to establish both the advantages and the risks of accompanying this 鈥渕ethodological revolution鈥.
This event has been organised as part of the Design and Analysis theme of the Centre for Evaluation.
Admission
