A Comparison of Point Estimation Methods in Adaptive Enrichment Designs
10.11783/j.issn.1002-3674.2025.01.007
- VernacularTitle:适应性富集设计中的效应点估计方法比较
- Author:
Shuang WANG
1
;
Zihang ZHONG
1
;
Zihan ZHU
1
Author Information
1. 南京医科大学公共卫生学院生物统计学系(211166)
- Publication Type:Journal Article
- Keywords:
Enrichment designs;
Point estimation;
Interim analysis;
Subgroup
- From:
Chinese Journal of Health Statistics
2025;42(1):38-43,49
- CountryChina
- Language:Chinese
-
Abstract:
Objective To explore methods for point estimation in adaptive enrichment design,providing a reference for effect estimation inn this context.Methods We simulated efficacy data under adaptive enrichment design and subsequentially estimated the effects using five methods:maximum likelihood method,stage 2 method,Lindley's method,parametric Bootstrap method,and the uniformly minimum variance conditionally unbiased estimated(UMVCUE)method.The performance of these methods was further evaluated.Results The simulation study showed that the maximum likelihood method has a large bias and is very sensitive to some parameters,while Lindley's method tends to underestimate and the parametric bootstrap method sometimes overestimates.The stage 2 method and UMVCUE are almost unbiased.In terms of mean squared error(MSE),all methods perform well except stage 2 estimator.Conclusion In adaptive enrichment design,the UMVCUE performs well in terms of bias and MSE.Therefore,we recommend it as the point estimator for the treatment effect in adaptive enrichment design.
