CMStatistics 2015: Start Registration
View Submission - CMStatistics
B0679
Title: On the use of Fleming and Harrington's test to detect late effects in clinical trials Authors:  Nicolas Savy - Toulouse Institute of Mathematics (France) [presenting]
Valerie Gares - University of Sydney NHMRC Clinical Trials Centre (Australia)
Jean-Francois Dupuy - INSA de Rennes (France)
Sandrine Andrieu - INSERM U 1027 (France)
Abstract: The question of detection of late effects in the setting of clinical trials has been previously investigated. The most natural test for detecting this kind of effects depends on a parameter that, in the context of clinical trials, must be chosen a priori. We examine the reasons why this test is adapted to the detection of late effects by studying its optimality in terms of Pitman Asymptotic Relative Efficiency. We give an explicit form of the function describing alternatives for which the test is optimal. Moreover, we observe by means of a simulation study that this test is not very sensitive to the value of the parameter, which is very reassuring for its use in clinical trials. The main conclusion is that the logrank test is optimal to test the proportional risks assumption, while Fleming Harrington test is optimal to test the late effect assumption. However, it is not always easy to choose between these alternatives. A test, based on a so-called expert a priori, has been introduced to override this difficulty.