Pitfalls and Solutions in Clone-Censor-Weight for Target Trial Emulation: Insights from Review, Simulation, and Real-World Analyses
- 1The University of Tokyo
Abstract
BackgroundThe clone-censor-weight (CCW) method is increasingly finding application in target trial emulation to compare treatment strategies involving grace periods. However, its performance has not been systematically evaluated against the ground truth. Many studies utilising CCW ignore time-varying covariates and informative pre-existing censoring. Heavy-tailed inverse probability of censoring weights (IPCW) may yield biased estimates and under-coverage.
MethodsSimulation Part 1 evaluated the bias, root mean squared error, and coverage probability of CCW analysis under correctly specified and misspecified IPCW models across scenarios with or without time-varying covariates and informative pre-existing censoring. Part 2 varied the confounding strength, assessing whether inference failures could be detected using an IPCW tail-heaviness index (a weighted Hill estimator-derived Pareto-type tail index) and the estimators standard deviation, the target of bootstrap standard error.
ResultsUnder correct specification, all bias estimates were below 0.005; coverage approached the nominal level of 0.95 with increasing sample size. Omitting the time-varying covariate or mishandling pre-existing censoring yielded bias up to 0.069 and coverage substantially below 0.95. Under strong confounding, confidence intervals failed to attain nominal coverage, even with correct specification and large sample sizes, when the oracle tail-heaviness index was [≤]1. When the index was >1, the coverage approached 0.95 as the estimators standard deviation decreased.
ConclusionsCCW analysis yields accurate estimates and confidence intervals with nominal coverage when implemented correctly, and confounding is not excessively strong. Investigators should account for time-varying covariates and pre-existing censoring and evaluate the tail behaviour of IPCW distribution and estimators bootstrap standard error.
Key messages- Correctly specified and implemented clone-censor-weight analyses can provide accurate strategy-specific risk estimates and confidence intervals with nominal coverage, whereas misspecified or incorrectly implemented analyses, such as those failing to account for time-varying confounders or informative pre-existing censoring, can yield biased estimates and under-coverage. - Even with correct model specifications and large samples, strong confounding can lead to heavy-tailed distributions of the inverse probability of censoring weights and invalidate standard confidence interval construction. - The tail heaviness index of the inverse probability of censoring weight distribution, in conjunction with the estimators bootstrap standard error, can help identify unreliable analyses and guide the specification of eligibility criteria and treatment strategies.
Related
- Ensuring Safe and Responsible Use of Artificial Intelligence in Medical Imaging: Pitfalls and Practical Solutions through the Lens of Obstetric Ultrasound.
- Selective serotonin reuptake inhibitors and immune checkpoint blockade in solid tumors: A target trial emulation with pan-cancer transcriptomic analyses
- Decoding adenomyosis: integrative insights from genetics, multiomics, and advanced computational analyses
- The association between inflammation and pubertal development : Insights from NHANES and Mendelian randomization analyses
- XuanBi Mitigates Rheumatoid Arthritis via Regulation of ECM‐Receptor Interactions and IL‐17 Signaling: Insights From Integrated Proteomic and Pharmacological Analyses
BibTeX
@misc{kimura2026pitfalls,
title = {{Pitfalls and Solutions in Clone-Censor-Weight for Target Trial Emulation: Insights from Review, Simulation, and Real-World Analyses}},
author = {Y. Kimura and Y. Takazawa and H. Yasunaga},
howpublished = {medRxiv},
year = {2026},
month = sep,
doi = {10.64898/2026.09.02.26362102},
url = {https://www.medrxiv.org/content/10.64898/2026.09.02.26362102v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - Pitfalls and Solutions in Clone-Censor-Weight for Target Trial Emulation: Insights from Review, Simulation, and Real-World Analyses
AU - Y. Kimura
AU - Y. Takazawa
AU - H. Yasunaga
PY - 2026
DA - 2026/09/02
PB - medRxiv
DO - 10.64898/2026.09.02.26362102
UR - https://www.medrxiv.org/content/10.64898/2026.09.02.26362102v1
AB - BackgroundThe clone-censor-weight (CCW) method is increasingly finding application in target trial emulation to compare treatment strategies involving grace periods. However, its performance has not been systematically evaluated against the ground truth. Many studies utilising CCW ignore time-varying covariates and informative pre-existing censoring. Heavy-tailed inverse probability of censoring weights (IPCW) may yield biased estimates and under-coverage. MethodsSimulation Part 1 evaluated the bias, root mean squared error, and coverage probability of CCW analysis under correctly specified and misspecified IPCW models across scenarios with or without time-varying covariates and informative pre-existing censoring. Part 2 varied the confounding strength, assessing whether inference failures could be detected using an IPCW tail-heaviness index (a weighted Hill estimator-derived Pareto-type tail index) and the estimators standard deviation, the target of bootstrap standard error. ResultsUnder correct specification, all bias estimates were below 0.005; coverage approached the nominal level of 0.95 with increasing sample size. Omitting the time-varying covariate or mishandling pre-existing censoring yielded bias up to 0.069 and coverage substantially below 0.95. Under strong confounding, confidence intervals failed to attain nominal coverage, even with correct specification and large sample sizes, when the oracle tail-heaviness index was [≤]1. When the index was >1, the coverage approached 0.95 as the estimators standard deviation decreased. ConclusionsCCW analysis yields accurate estimates and confidence intervals with nominal coverage when implemented correctly, and confounding is not excessively strong. Investigators should account for time-varying covariates and pre-existing censoring and evaluate the tail behaviour of IPCW distribution and estimators bootstrap standard error. Key messages- Correctly specified and implemented clone-censor-weight analyses can provide accurate strategy-specific risk estimates and confidence intervals with nominal coverage, whereas misspecified or incorrectly implemented analyses, such as those failing to account for time-varying confounders or informative pre-existing censoring, can yield biased estimates and under-coverage. - Even with correct model specifications and large samples, strong confounding can lead to heavy-tailed distributions of the inverse probability of censoring weights and invalidate standard confidence interval construction. - The tail heaviness index of the inverse probability of censoring weight distribution, in conjunction with the estimators bootstrap standard error, can help identify unreliable analyses and guide the specification of eligibility criteria and treatment strategies.
ER -