Abstract:
|
In many clinical studies, the event of interest may occur more than once for each subject. Extended accelerated failure time (AFT) model is an important approach for the analysis of recurrent events. However, in clinical practice, many covariates could not be measured accurately due to natural biological ?uctuation or instrument error. For regression analysis in general, naively using mismeasured covariates in conventional inference procedures may incur substantial estimation bias. This presentation proposes an estimation procedure for recurrent events data under the AFT model in which some covariates are not accurately ascertainable. With replicated mismeasured covariates, the proposed procedure requires no distributional assumptions on either the true covariates or the error. The proposed estimating function is developed on the basis of a novel identity and distinct from conventional correction strategy. The resulting coefficient estimators are shown to be consistent and asymptotically normal. Performance of the proposed procedure was investigated by numerical studies with practical sample size. In addition, an application to clinical trial data is provided for illustration.
|
ASA Meetings Department
732 North Washington Street, Alexandria, VA 22314
(703) 684-1221 • meetings@amstat.org
Copyright © American Statistical Association.