본문으로 건너뛰기
← 뒤로

Control arm augmentation and hierarchical modeling in time-to-event trials: advantages and pitfalls.

1/5 보강
Biostatistics (Oxford, England) 2025 Vol.26(1)
Retraction 확인
출처

Alt EM, Chang X, Liu Q, Jiang X, Mo M, Xia HA, Ibrahim JG

📝 환자 설명용 한 줄

In clinical trials, it is often valuable to borrow information from external data sources.

이 논문을 인용하기

↓ .bib ↓ .ris
APA Alt EM, Chang X, et al. (2025). Control arm augmentation and hierarchical modeling in time-to-event trials: advantages and pitfalls.. Biostatistics (Oxford, England), 26(1). https://doi.org/10.1093/biostatistics/kxaf021
MLA Alt EM, et al.. "Control arm augmentation and hierarchical modeling in time-to-event trials: advantages and pitfalls.." Biostatistics (Oxford, England), vol. 26, no. 1, 2025.
PMID 40795860 ↗

Abstract

In clinical trials, it is often valuable to borrow information from external data sources. Unfortunately, when the external data are fully or partially incompatible with the current trial data, type I error rates can be highly inflated under traditional blanket discounting schemes such as power priors, commensurate priors, and meta-analytic predictive priors. However, such inflation of the probability of a false positive can be necessary, as the alternative is to have an underpowered study. For clinical trials with time-to-event (TTE) outcomes, this problem is exacerbated since many observations are censored. In this paper, we develop the latent exchangeability prior for TTE data. We also present a novel framework to borrow information about the treatment effect between groups as well as incorporate information from external controls. Simulation results suggest that, although efficiency gains can be achieved by borrowing information among external controls, operating characteristics in general can be quite poor under a lack of exchangeability. We apply our approach to a real clinical trial in second-line metastatic colorectal cancer.

🏷️ 키워드 / MeSH 📖 같은 키워드 OA만

🏷️ 같은 키워드 · 무료전문 — 이 논문 MeSH/keyword 기반