B1052
Title: A general class of nonparametric test statistics for interval-censored data
Authors: Ramon Oller - Fundacio Universitaria Balmes (Spain) [presenting]
Abstract: An extension of the Jones and Crowley's class of nonparametric test statistics is proposed for interval-censored data. The test statistics are formulated in terms of the estimated distribution function $\hat{F}_i(t)$ of each individual in the sample. They measure the relationship between survival times and a quantitative vector label $\mathbf{z}(t)$ attached to each individual. Under the null hypothesis of no relationship and the assumption that the underlying censoring process is identical for all individuals, the permutation distribution of the test statistics is derived. In this general class, we obtain and investigate several particular statistics. Some of them are well known, for instance a set of weighted log-rank test statistics. Others are new and give new insights for analyzing interval-censored data, for instance a set of Kendall-type test statistics.
Journal