Abstract:
|
Generalized estimating equations (GEE) are commonly utilized for the marginal analysis of longitudinal data. In order to obtain consistent regression parameter estimates, these estimating equations must be unbiased. However, when utilizing certain types of time-dependent covariates, these equations can be biased unless the independence working structure is used. Moreover, regression parameter estimation can be very inefficient because not all valid moment conditions are incorporated within the corresponding equations. Therefore, approaches utilizing modified versions of GEE or quadratic inference functions have been proposed in order to utilize all valid moment conditions. However, these approaches assumed the analyst knows the type of time-dependent covariate. In practice this may not be the case. Accordingly, we propose selection approaches, relative to existing methods, to assess a specific type of time-dependent covariate. Resulting estimates are consistent and estimated with the greatest possible efficiency. Existing and proposed methods are compared in a simulation study and application example.
|