改善以存活時間為主要目標的三臂等效性試驗所需樣本數的計算

Project Details

Description

In this project, we will investigate three directions of improvement to reduce the minimum required sample size for three-arm equivalence trials with survival time as the primary endpoint. First, delayed treatment effects challenge the proportionality assumption of the Cox model. Thus, for this situation, the nonproportionality survival models including time-varying regression coefficient Cox model, AFT model, Additive survival models will be investigated and compared the performance with conventional remedies such as weighted log-rank test and piecewise constant hazard model. Second, in the literature, the score testing procedure or replacement of MLE by RMLE in a Wald testing procedure significantly reduce the minimum required sample size in many two-arm trial cases. Consequently, we will develop either a score test procedure or RMLE replacement Wald test procedure for the three-arm trials under survival outcomes. Third, model misspecification may overestimate the minimum required sample size in clinical trials. Thus, we will develop a general model selection method for semi-parametric hazard regression models to avoid severe model misspecification. The model selection criterion is based on the concordance index C, which is a function of hazard function conditional on covariates under suitable assumptions. We illustrate the performance of all improvements by considering a randomized, active- and placebo-controlled trial in subjects with travelers’ diarrhea. The validity and usefulness of the proposed method will be investigated by simulation studies.
StatusFinished
Effective start/end date1/08/2231/10/23

Keywords

  • Concordance index
  • Delayed treatment effects
  • RMLE
  • Semi-parametric hazards regression model

Fingerprint

Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.