Causal Association

One of the primary purposes of epidemiological research is to find out an association between risk factors and a disease. Some of these associations could be statistically significant, and some are not. When the association is statistically significant, we also tend to infer that we found a causal association. However, epidemiologists should be cautious in describing an association between variables to be causal because not all associations, even statistically significant, are necessarily causal associations. One of the factors we must rule out is the association due to the confounding effect of some other factors. In addition, Bradford Hill proposed a few criteria which can be applied to identify if there is a causal relationship between the exposure and a disease or an event. We will describe the concepts of a cause-and-effect relationship between variables with practical examples. The chapter will also describe a multifactorial disease model, known as the sufficient-component cause model, popularly called Rothman’s causal pie. In this connection, the readers will differentiate between several terms suggested by Rothman such as component cause, necessary cause, and sufficient cause of a disease which occurs due to multiple factors.