ASSOCIATION MODELS FOR PERIODONTAL-DISEASE PROGRESSION - A COMPARISONOF METHODS FOR CLUSTERED BINARY DATA

We investigate population-averaged (PA) and cluster-specific (CS) asso ciations for clustered binary logistic regression in the context of a longitudinal clinical trial that investigated the association between tooth-specific visual elastase kit results and periodontal disease pro gression within 26 weeks of follow-up. We address estimation of popula tion-averaged logistic regression models with generalized estimating e quations (GEE), and conditional likelihood (CL) and mixed effects (ME) estimation of CS logistic regression models. Of particular interest i s the impact of clusters that do not provide information for condition al likelihood methods (non-informative clusters) on inferences based u pon the various methodologies. The empirical and analytical results in dicate that CL methods yield smaller test statistics than ME methods w hen noninformative clusters exist, and that CL estimates are less effi cient than ME estimates under certain conditions. Moreover, previously reported relationships between population-averaged and cluster-specif ic parameters appear to hold for the corresponding estimates in the pr esence of these clusters.