Comparison of state occupation, entry, exit and waiting times in two or more groups based on current status data in a multistate model.

摘要

We propose mean absolute distance-based nonparametric bootstrap tests comparing the occupation probabilities, entry, exit and waiting times in a given state in two or more groups in a multistate model. The state occupied is determined at a single inspection time for each individual. Nonparametric monotonic regression and smoothing are used to uniquely define the various marginal estimators for each group which are then compared via a mean absolute distance criterion. A detailed simulation study shows that the proposed tests all have close to nominal size and reasonable power. An illustrative application to a pubertal development data set obtained from the NHANES III (Analytic and Reporting Guidelines: The Third National Health and Nutrition Examination Survey (1988-94). Hyattsville, MD: National Center for Health Statistics, Centers for Disease Control and Prevention, 1997) is also presented.