%0 Journal Article %A Lin, Ray S. %A Lin, Ji %A Roychoudhury, Satrajit %A Anderson, Keaven M. %A Hu, Tianle %A Huang, Bo %A Leon, Larry F %A Liao, Jason J.Z. %A Liu, Rong %A Luo, Xiaodong %A Mukhopadhyay, Pralay %A Qin, Rui %A Tatsuoka, Kay %A Wang, Xuejing %A Wang, Yang %A Zhu, Jian %A Chen, Tai-Tsang %A Iacona, Renee %D 2019 %T Alternative Analysis Methods for Time to Event Endpoints under Non-proportional Hazards: A Comparative Analysis %U https://tandf.figshare.com/articles/journal_contribution/Alternative_Analysis_Methods_for_Time_to_Event_Endpoints_under_Non-proportional_Hazards_A_Comparative_Analysis/11310851 %R 10.6084/m9.figshare.11310851.v1 %2 https://tandf.figshare.com/ndownloader/files/20047583 %K Oncology trial %K survival analysis %K non-proportional hazards %K log-rank test %K Fleming-Harrington test %X

The log-rank test is most powerful under proportional hazards (PH). In practice, non-PH patterns are often observed in clinical trials, such as in immuno-oncology; therefore, alternative methods are needed to restore the efficiency of statistical testing. Three categories of testing methods were evaluated, including weighted log-rank tests, Kaplan-Meier curve-based tests (including weighted Kaplan-Meier and Restricted Mean Survival Time, RMST), and combination tests (including Breslow test, Lee’s combo test, and MaxCombo test). Nine scenarios representing the PH and various non-PH patterns were simulated. The power, type-I error, and effect estimate of each method was compared. In general, all tests control type I error well. There is not a single most powerful test across all scenarios. In the absence of prior knowledge regarding the PH or non-PH patterns, the MaxCombo test is relatively robust across patterns. Since the treatment effect changes overtime under non-PH, the overall profile of the treatment effect may not be represented comprehensively based on a single measure. Thus, multiple measures of the treatment effect should be pre-specified as sensitivity analyses to evaluate the totality of the data.

%I Taylor & Francis