Historical Data Borrowing with Meta Analytic Predictive Methodology in Adaptive Design Clinical Trials
10.11783/j.issn.1002-3674.2025.04.004
- VernacularTitle:基于MAP先验的历史数据信息借用在适应性设计临床试验中的应用研究
- Author:
Yaru HUANG
1
;
Binbin KANG
;
Siyu XIONG
Author Information
1. 空军军医大学军事预防医学系军队卫生统计学教研室,教育部特殊作业环境危害评估与防治重点实验室,陕西省环境健康危害评估与防护重点实验室,陕西省自由基生物学与医学重点实验室 710032
- Publication Type:Journal Article
- Keywords:
Clinical trials;
Historical data;
Bayesian method;
MAP prior;
Adaptive design
- From:
Chinese Journal of Health Statistics
2025;42(4):496-501,509
- CountryChina
- Language:Chinese
-
Abstract:
Objective Exploring the MAP(meta analytic predictive)methods for historical information borrowing for adaptive trials.Methods Taking the clinical trial of recurrent or metastatic head and neck squamous cell carcinoma as an example,the applicability of the MAP method in the adaptive trial was evaluated with different heterogeneity of historical control data and different degrees of data conflict.Results For the five applicable extreme protocol historical studies with a total of 253 subjects,the MAP prior converted to an effective sample size(ESS)of 19 cases,and the RMAP(robust MAP)prior converted to ESS of 17 cases.The RMAP method balanced the potential conflicts between historical data and current trial data better than the MAP method.With the increase of the heterogeneity among historical controls and the conflict between the prior and the current data,the type-I error of both the MAP method and the RMAP method were slightly inflated.When the current control data corresponded with the historical controls,the power of both the MAP prior and the RMAP prior decreased with the increase of heterogeneity.For the two-stage adaptive trial,borrowing information saved 30%-35% of ESS in control group.Conclusion The MAP method can be a valid and robust methodoiogical tool for the effective application of historical information borrowing for the adaptive design trials,which can help to optimize experimental design,cinserve resources,and assist for trial decision.
