DESIGN AND ANALYSIS OF MULTILEVEL ANALYTIC STUDIES WITH APPLICATIONS TO A STUDY OF AIR-POLLUTION

We discuss a hybrid epidemiologic design that aims to combine two appr oaches to studying exposure-disease associations. The analytic approac h is based on comparisons between individuals, e.g., case-control and cohort studies, and the ecologic approach is based on comparisons betw een groups. The analytic approach generally provides a stronger basis for inference, in part because of freedom from between-group confoundi ng and better quality data, but the ecologic approach is less suscepti ble to attenuation bias from measurement error and may provide greater variability in exposure. The design we propose entails selection of a number of groups and enrollment of individuals within each group. Exp osures, outcomes, confounders, and modifiers would be assessed on each individual; but additional exposure data might be available on the gr oups. The analysis would then combine the individual-level and the gro up-level comparisons, with appropriate adjustments for exposure measur ement errors, and would test for compatibility between the two levels of analysis, e.g., to determine whether the associations at the indivi dual level can account for the differences in disease rates between gr oups. Trade-offs between numbers of groups, numbers of individuals, an d the extent of the individual and group measurement protocols are dis cussed in terms of design efficiency. These issues are illustrated in the context of an on-going study of the health effects of air pollutio n in southern California, in which 12 communities with different level s and types of pollution have been selected and 3500 school children a re being enrolled in a ten-year cohort study. Exposure is being assess ed through a combination of ambient monitoring, microenvironmental sam pling, personal monitoring, and questionnaire data on time-activity an d household characteristics. These data will be used to develop a mode l for personal exposures for use in the individual-level analyses, as well as for the group mean exposures for the group-level analyses.