Title: Nonparametric estimation of the joint distribution of two gap times under various types of censoring and truncation
Authors: Jacobo de Una-Alvarez - University of Vigo (Spain)
Ana Cristina Santos - Institute of Public Health - University of Porto (Portugal)
Carla Moreira - University of Minho (Portugal) [presenting]
Abstract: The statistical analysis of consecutive gap times is an issue of much importance in a number of fields, including engineering, economy, epidemiology, and survival analysis. Particular difficulties appear, for example, when information on a cohort is obtained through intermittent visits or successive cross-sections; in such cases, special combinations of left-truncated, right-censored and interval censored data will appear. We describe one of such complicated settings in a three-state progressive model, and we introduce an inverse-probability-weighted type estimator for the joint distribution of two gap times which takes the aforementioned censoring and truncation issues into account. The performance of the proposed estimator is investigated through simulations, considering that the possible dependence structure between the two gap times is ruled by a parametric copula. For illustration purposes, the estimator is applied to data from the EPIPorto adults cohort study.