Estimating efficacy in trials with selective crossover

Abstract

When one arm in a trial has a worse early endpoint such as recurrence, a data‐monitoring committee might recommend that all participants are offered the apparently superior treatment. The resultant crossover makes it difficult to measure differences between arms thereafter, including for longer‐term endpoints such as mortality and disease‐specific mortality. In this paper, we consider estimators of the efficacy of treatment on those who would not cross over if randomised to the apparently inferior arm. Binomial and proportional hazards maximum likelihood estimators are developed. The binomial estimator is applied to analysis of a breast cancer treatment trial and compared with intention‐to‐treat and inverse probability weighting estimators. Full and partial likelihood proportional‐hazard model estimators are assessed through computer simulations, where they had similar bias and variance. The new efficacy estimators extend those for all‐or‐none compliance to this important problem. © 2017 The Authors. Statistics in Medicine Published by John Wiley & Sons Ltd

Comment on “ALPPS Improves Survival Compared With TSH in Patients Affected of CRLM. Survival Analysis From the Randomized Controlled Trial LIGRO” by K. Hasselgren, et al., Annals of Surgery 2020 The Jury is Still Out

“…When 1 arm has a worse early endpoint, the resultant crossover confounds differences in late endpoints, such as survival, between the 2 arms. 8 The main oncological outcome of the study reported by Sandstro ¨m et al is presented in Figure 2, which displays the overall survival based on an intention-to-treat design. Although these findings are impressive, one could argue that the final conclusions should also be drawn from Figure 3, which represents a more direct comparison of the 44 patients who completed ALPPS with the 27 patients who completed TSH, excluding those with rescue ALPPS.…”

“…In this situation, sequential HVE could be performed instead of the first stage of ALPPS; HVE could be performed under local anesthesia and light sedation, on an outpatient basis, and, most importantly, with no risk of severe morbidity or mortality. When 1 arm has a worse early endpoint, the resultant crossover confounds differences in late endpoints, such as survival, between the 2 arms 8 …”

Two-stage estimation to adjust for treatment switching in randomised trials: a simulation study investigating the use of inverse probability weighting instead of re-censoring

“…Other types of switching occur – switching is sometimes triggered by an interim analysis [28] or patients may switch on to treatments other than those under investigation in the RCT [1, 5]. The type of switching may influence whether adjustment is appropriate, and what the target of estimation should be [29] but these issues are not the focus of this study. We use the simulation study previously reported by Latimer et al [18] but extend it to include the TSE method in combination with IPCW in addition to the previously investigated TSE with and without re-censoring methods.…”