Objective To compare the commonly used methods for analyzing treatment switching in clinical trials to facilitate selection of optimal methods in different scenarios. Methods Based on the data characteristics of patient conversion in oncology clinical trials, we simulated the survival time of patients across different scenarios and compared the bias, mean square error and coverages of the treatment effects derived from different methods. Results The sample size had an almost negligible impact on the outcomes of the various methods. Compared to conventional methods, more complex methods (RPSFTM, IPCW, TSE, and IPE) resulted in lower errors across different scenarios. The IPCW method could cause a significant increase in errors in cases where the probability of conversion was high. The TSE method had the lowest error and mean squared error when the risk was low and the probability of conversion was high. The IPE method had an obvious advantage in the scenario with a low probability of conversion, but it may slightly underestimate the treatment effect when the inflation factor was small. Conclusion The choice of a specific method for analyzing cohort transition should be made based on considerations of both the probability of conversion and inflation factor in different scenarios.
HarmonisationICF. E9(R1) Statistical Principles for Clinical Trials: Addendum: Estimands and Sensitivity Analysis in Clinical Trials; International Council for Harmonisation; Draft Guidance for Industry; Availability[J]. Fed Regist, 2017, 82(209): 7-14.
[2]
ShihA, HsuCY, ShyrY. Power and sample size calculation for non-inferiority trials with treatment switching in intention-to-treat analysis comparing RMSTs.[J]. Research square, 2024.
[3]
LejunD, ChihYuanH, YuS. Power and sample sizes estimation in clinical trials with treatment switching in intention-to-treat analysis: a simulation study.[J]. BMC Med Res Methodol, 2023, 23(1): 49.
LatimerNR, DewdneyA, CampioniM. A cautionary tale: an evaluation of the performance of treatment switching adjustment methods in a real world case study[J]. BMC Med Res Methodol, 2024, 24(1): 17.
[6]
MordenJP, LambertPC, LatimerN, et al. Assessing methods for dealing with treatment switching in randomised controlled trials: a simulation study[J]. BMC Med Res Methodol, 2011, 11: 4.
[7]
LatimerNR, AbramsKR, LambertPC, et al. Assessing methods for dealing with treatment switching in clinical trials: a follow-up simulation study[J]. Stat Methods Med Res, 2018, 27(3): 765-84.
[8]
XuanJ, IsaSM, LatimerN, et al. Is inverse probability of censoring weighting a safer choice than per-protocol analysis in clinical trials?[J]. Stat Methods Med Res, 2024, 34(2): 1029045591.
EunJinA, HyunK. Intention-to-treat versus as-treated versus per-protocol approaches to analysis.[J]. Korean J Anesthesiol, 2023, 76(6): 531-9.
[11]
LawMG, KaldorJM. Author's Reply: Survival analysis of randomized clinical trials adjusted for patients who switch treatments by M. G. Law and J. M. Kaldor, Statistics in Medicine, 15, 2069-2076 (1996)[J]. Statist Med, 1997, 16(22): 2620-1.
[12]
XuanJ, IsaSM, LatimerN, et al. Is inverse probability of censoring weighting a safer choice than per-protocol analysis in clinical trials[J]? Stat Methods Med Res, 2024, 34(2): 1029045591.
[13]
ReckM, DeTL, Paz-AresL, et al. Treatment-switching adjustment of overall survival in CheckMate 227 part 1 evaluating first-line nivolumab plus ipilimumab versus chemotherapy for metastatic nonsmall cell lung cancer[J]. Clin Lung Cancer, 2024, 25(7): e362-8.
[14]
RobinsJM, HernánMA, BrumbackB. Marginal structural models and causal inference in epidemiology[J]. Epidemiology, 2000, 11(5): 550-60.
GorrodHB, IsaSM, XuanJ, et al. Adjusting for switches to multiple treatments: Should switches be handled separately or combined?[J]. Stat Methods Med Res, 2025, 34(2): 1029056081.
LatimerNR, AbramsKR, LambertPC, et al. Adjusting for treatment switching in randomised controlled trials - A simulation study and a simplified two-stage method[J]. Stat Methods Med Res, 2017, 26(2): 724-51.
[19]
LatimerNR, WhiteIR, AbramsKR, et al. Causal inference for long-term survival in randomised trials with treatment switching: Should re-censoring be applied when estimating counterfactual survival times?[J]. Stat Methods Med Res, 2019, 28(8): 2475-93.
[20]
School Of Health And Related Research UOSS, MRC Clinical Trials Unit UCLL, Population Health Sciences BMSU, et al. Improved two-stage estimation to adjust for treatment switching in randomised trials: g-estimation to address time-dependent confounding[J]. Stat Methods Med Res, 2020, 29(10): 2900-18.
[21]
JacksonD, RanD, ZhangF, et al. New methods for two-stage treatment switching estimation[J]. Pharm Stat, 2025, 24(1): e2462.
[22]
RobinsJM, TsiatisAA. Correcting for non-compliance in randomized trials using rank preserving structural failure time models[J]. Commun Stat Theory Meth, 1991, 20(8): 2609-31.
[23]
SullivanTR, LatimerNR, GrayJ, et al. Adjusting for Treatment Switching in Oncology Trials: A Systematic Review and Recommendations for Reporting[J]. Value Health, 2020, 23(3): 388-96.
[24]
Bell GorrodH, LatimerNR, DamianD, et al. Assessing the long-term effectiveness of cladribine vs. placebo in the relapsing-remitting multiple sclerosis CLARITY randomized controlled trial and CLARITY extension using treatment switching adjustment methods[J]. Adv Ther, 2020, 37(1): 225-39.
[25]
LatimerNR, AbramsKR, AmonkarMM, et al. Adjusting for the confounding effects of treatment switching-the BREAK-3 trial: dabrafenib versus dacarbazine[J]. Oncologist, 2015, 20(7): 798-805.
[26]
BenderR, AugustinT, BlettnerM. Generating survival times to simulate Cox proportional hazards models[J]. Stat Med, 2005, 24(11): 1713-23.
[27]
WalkerAS, WhiteIR, BabikerAG. Parametric randomization-based methods for correcting for treatment changes in the assessment of the causal effect of treatment[J]. Stat Med, 2004, 23(4): 571-90.
[28]
CollettD. Modelling survival data in medical research[M]. 3rd ed. Taylor and Francis, 2015: 204-219
[29]
BurtonA, AltmanDG, RoystonP, et al. The design of simulation studies in medical statistics[J]. Stat Med, 2006, 25(24): 4279-92.
[30]
LatimerNR. Treatment switching in oncology trials and the acceptability of adjustment methods[J]. Expert Rev Pharmacoecon Outcomes Res, 2015, 15(4): 561-4.