Sankhya: The Indian Journal of Statistics

2000, Volume 62, Series B, Pt. 1, 134--148

LONGITUDINAL DATA ANALYSIS OF CONTINUOUS AND DISCRETE RESPONSES FOR PRE-POST DESIGNS

By

KUNG--YEE LIANG,

and

SCOTT L. ZEGER, Johns Hopkins University

SUMMARY. The availability of biomarkers has led to the increasing adoption of the pre- and post-randomization designs in clinical trials. In this paper we discuss the use of random effects models when the primary objective of the trial is to assess the efficacy of newly developed treatments in terms of progression of biomarkers over time. One issue of particular interest is how to best utilize the pre-randomization responses. We discuss the pros and cons of several inferential procedures including the conditional and full likelihood approaches. Throughout, these issues are illustrated by an analysis of data from a schizophrenic trial by the Janssen Research Foundation.

AMS (1991) subject classification. Primary 62J10, 62J12.

Key words and phrases. Clinical trials; longitudinal data; pre-post designs; random effects model.

Full paper (PDF)

This article in Mathematical Reviews