A BAYESIAN HIERARCHICAL SURVIVAL MODEL FOR THE INSTITUTIONAL EFFECTS IN A MULTICENTER CANCER CLINICAL-TRIAL

In randomized clinical trials comparing treatment effects on diseases such as cancer, a multi-centre trial is usually conducted to accrue th e required number of patients in a reasonable period of time. While we interpret the average treatment effect, it is necessary to examine th e homogeneity of the observed treatment effects across institutions, t hat is, treatment-by-institution interaction. If the homogeneity is co nfirmed, the conclusions concerning treatment effects can be generaliz ed to a broader patient population. In this paper, a Bayesian hierarch ical survival model is used to investigate the institutional effects o n the efficacy of treatment as well as on the baseline risk. The margi nal posterior distributions are estimated by a Markov chain Monte Carl o method, that is, Gibbs sampling, to overcome current computational l imitations. The robustness of the inferences to the distributional ass umption for the random effects is also examined. We illustrate the met hods with analyses of data from a multi-centre cancer clinical trial, which investigated the efficacy of immunochemotherapy as an adjuvant t reatment after curative resection of gastric cancer. In this trial the re is little difference in the treatment effects across institutions a nd the treatment is shown to be effective, while there appears to be s ubstantial variation in the baseline risk across institutions. This re sult indicates that the observed treatment effects might be generalize d to a broader patient population. (C) 1998 John Wiley & Sons, Ltd.