Prevalent Cohort Models in Clinical Treatment Studies: An Analysis of Zidovudine Treatment Data

Mei-Cheng Wang, Department of Biostatistics, Johns Hopkins University
Ron Brookmeyer, Department of Biostatistics, Johns Hopkins University
Nicholas P. Jewell, Division of Biostatistics, School of Public Health, University of California, Berkeley

Abstract

In prospective cohort studies individuals are sometimes recruited according to a certain cross-sectional sampling criterion. A prevalent cohort is defined as a group of individuals who satisfy a certain criterion when they enter the study. In this paper the situation that the time of enrollment coincides with the starting time of a treatment is considered and emphasized. Statistical models in one-sample and regression problems are discussed. The incident proportional hazards model in which the times of disease diagnoses are retrospectively ascertained as the origins of failure times is studied. The relationship between the incident proportional hazards model and the prevalent proportional hazards model, where the time origin is the time of treatment commencement, is explored. The analysis of data from an observational study of zidovudine (ZVD) in patients with the acquired immunodeficiency syndrome (AIDS) is presented.