Review of software to fit generalized estimating equation regression models

Researchers are often interested in analyzing data that arise from a longit udinal or clustered design. Although there are a variety of standard likeli hood-based approaches to analysis when the outcome variables are approximat ely multivariate normal, models for discrete-type outcomes generally requir e a different approach. Liang and Zeger formalized an approach to this prob lem using generalized estimating equations (GEEs) to extend generalized lin ear models (GLMs) to a regression setting with correlated observations with in subjects. Tn this article, we briefly review GLM, the GEE methodology, i ntroduce some examples, and compare the GEE implementations of several gene ral purpose statistical packages (SAS, Stata, SUDAAN, and S-Plus). We focus on the user interface, accuracy, and completeness of implementations of th is methodology.