SOFTWARE FOR HIERARCHICAL MODELING OF EPIDEMIOLOGIC DATA

Hierarchical models can provide more reasonable and stable parameter e stimates than conventional analytical approaches. This technique also deals with problems of multiple comparisons and allows one to model mu ltilevel data within a hierarchical framework. Hence, one would antici pate a surge in applying hierarchical models to epidemiologic data. Di fficulties in fitting hierarchical models, however, seem to have limit ed their use. To help address this problem, we describe the existing s oftware packages that one can use to fit hierarchical models. Since th ese packages have limited familiarity and applicability in epidemiolog y, we also present SAS code for analyzing epidemiologic data with hier archical models. These results allow epidemiologists to fit hierarchic al models with readily available software.