Fat and Calories in the Epidemiology of Colon Cancer in Western New York

In this article, we contrast three models of macronutrients in the epidemiology of colon cancer. We then evaluate a dataset collected in Buffalo showing that the macronutrient model we use to guide our data analysis has a decisive bearing on the answers we obtain. We attempt to demonstrate that the method of statistical analysis we employ depends on the biologic–behavioral model that we assume and that, if we assume the wrong model, we may perform the wrong analysis and generate an answer that is wrong. We focus on dietary fat, because the intake of fat has received so much attention in the nutritional epidemiology of cancer. In addition, the intake of fat brings the issue of collinearity between a macronutrient and total caloric intake into focus and helps accentuate the contours of our debate.