Nonparametric estimation of waiting time distribution functions and other characteristics of multi-state models

Abstract:  As a type of multivariate survival data, multi-state models have a wide range of applications, notably in cancer and infectious disease progression studies. Neither empirical nor rate-based (Kaplan-Meier) calculations are possible using current status data when each subject is inspected only once during the entire study period. In this talk, I will present two nonparametric approaches, one using a product limit formula and another using a fractional at-risk set, for calculation of state occupation probabilities as well as distribution functions of state entry and exit times using isotonic regression and kernel smoothing. Based on the fractional at-risk set method, we also construct nonparametric estimators of state waiting time distribution functions in a Markov multi-state model using current status data. This is a particularly difficult problem since neither the entry nor the exit times of a given state are directly observed. Detailed simulation studies will be presented in support of our estimators. We will also illustrate our methods using a pubertal development data set obtained from the NHANES III.

Refreshments and socializing: 2:30 - 3:00 PM