Analysis of smoking trends with incomplete longitudinal binary responses

The generalized estimating equations procedure (GEE) widely applied in the analysis of correlated binary data requires that missing data depend only o n remote covariates or that they be missing completely at random (MCAR); ot herwise GEE regression parameter estimates are biased. A weighted generaliz ed estimating equations (WGEE) approach that accounts for dropouts under th e less stringent assumption of missing at random (MAR) through dependence o n observed responses gives unbiased estimation of parameters in the model f or the marginal means if the dropout mechanism is specified correctly. WGEE s are applied in the estimation of 7-year trends in cigarette smoking in th e United States from a cohort of 5,078 black and white young adults. Analys is using WGEE suggests that there was a general decline in cigarette smokin g only among white females, whereas the only other subgroup for which smoki ng declined was white males of the older birth cohort (1955-1962) with coll ege degrees. The results of WC;EE are compared to a likelihood-based method valid under MAR that does not require specification of a missing data mode l.