DESIGN AND ANALYSIS OF MULTIARM CLINICAL-TRIALS WITH SURVIVAL END-POINTS

The clinical trials literature has paid relatively little attention to the design and analysis of K-sample trials with survival endpoints wh ere K is 3 or greater. Following the least-significant-difference appr oach proposed by Makuch and Simon [1], we derive sample size formulas by working with the logrank test and. proportional hazards model direc tly. This approach ensures the type I error rate to be the nominal val ue when the global null hypothesis is true. For power considerations, planning the study based on the least favorable alternative is recomme nded. The resulting sample size requirements are presented in graphic form for K = 3 and 4. Assuming that there is a control group and consi dering only the alternative that the survival of the experimental trea tments is at least as good as that of the control group, power investi gations indicate that the proposed strategy has good power for detecti ng the difference between the control and the best treatment. The ''ov erall power,'' defined as the chance of the global test and subsequent pairwise comparisons all being correct, is good when all treatments a re similar to either the control or the best treatment. Overall power is poor when the hazards are more evenly spread out between the contro l and the best group because the sample size is inadequate to detect s uch differences.