BRINGING CONTEXT BACK INTO EPIDEMIOLOGY - VARIABLES AND FALLACIES IN MULTILEVEL ANALYSIS

A large portion of current epidemiologic research is based on methodol ogic individualism: the notion that the distribution of health and dis ease in populations can be explained exclusively in terms of the chara cteristics of individuals. The present paper discusses the need to inc lude group-or macro-level variables in epidemiologic studies, thus inc orporating multiple levels of determination in the study of health out comes. These types of analyses, which have been called contextual or m ultilevel analyses, challenge epidemiologists to develop theoretical m odels of disease causation that extend across levels and explain how g roup-level and individual-level variables interact in shaping health a nd disease. They also raise a series of methodological issues, includi ng the need to select the appropriate contextual unit and contextual v ariables, to correctly specify the individual-level model, and, in som e cases, to account for residual correlation between individuals withi n contexts. Despite its complexities, multilevel analysis holds potent ial for reemphasizing the role of macro-level variables in shaping hea lth and disease in populations.