Bringing context back into epidemiology: variables and fallacies in multilevel analysis

摘要: A large portion of current epidemiologic research is based on methodologic individualism: the notion that distribution health and disease in populations can be explained exclusively terms characteristics individuals. The present paper discusses need to include group- or macro-level variables studies, thus incorporating multiple levels determination study outcomes. These types analyses, which have been called contextual multi-level challenge epidemiologists develop theoretical models causation extend across explain how group-level individual-level interact shaping disease. They also raise a series methodological issues, including select appropriate unit variables, correctly specify model, and, some cases, account for residual correlation between individuals within contexts. Despite its complexities, ...