The Analysis of Current Status Data on Point Processes

Abstract

This article considers the analysis of event count data in which each subject is observed at only one time point and no information is available on subjects between their entry time and observation points. This type of data, often referred to as current status data, arises frequently—for example, in demography, epidemiology, and reliability studies. Statistical methods for the analysis of current status data from point processes are proposed. Specifically, a statistic for testing the equality of the mean functions of point processes is presented and its asymptotic distribution obtained. For illustration, the proposed method is used to analyze multiple tumor data from a tumorgenicity experiment, with focus on the comparison of tumor growth rates in male and female rats. The adequacy of the asymptotic distribution of the test statistic is evaluated in a small simulation study. Power comparisons with the usual parametric model are also obtained. Finally, some possible directions for further research are discussed.

Key Words:

• Comparison test
• Isotonic regression
• Tumorgenicity study