EUVIMEDEUVIMED entdecken
Crossref

A hybrid prior Bayesian method for combining domestic real-world data and overseas data in global drug development

Keer Chen iD, Zengyue Zheng iD, Pengfei Zhu, Shuping Jiang, Nan Li, Jumin Deng, Pingyan Chen, Zhenyu Wu, Ying Wu iD

Statistical Methods in Medical Research · 2026

DOI: 10.1177/09622802251414586 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Statistical Methods in Medical Research, 2026. Verfasst von Keer Chen und 8 weiteren Autor:innen. Beteiligte Einrichtung: Department of Biostatistics, School of Public Health, Southern Medical University, Guangzhou, China. Hauptthemen laut MEDLINE: Drug Development. Die Arbeit nennt 44 Literaturangaben. Gefördert von: Guangzhou Basic and Applied Foundation Project, Real World Research Project Grant Fund from the Hainan Institute of Real World data, National Natural Science Foundation of China.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background

Hybrid clinical trial design integrates traditional randomized controlled trials (RCTs) with real-world data (RWD), aiming to enhance trial efficiency through dynamic incorporation of external data (External trial data and RWD). However, existing methods, such as the Meta-Analytic Predictive (MAP) Prior, exhibit serious limitations in controlling data heterogeneity, adjusting baseline discrepancies, and optimizing dynamic borrowing proportions. These limitations often introduce external bias or compromise evidence reliability, hindering their application in complex analyses like bridging trials and multi-regional clinical trials (MRCTs).

Objective

This study proposes a novel hybrid Bayesian framework, EQPS Robust MAP (rMAP), to address heterogeneity and bias in multi-source data integration. Its feasibility and robustness are validated through systematic simulations and retrospective case analyses, using two independent datasets to evaluate the effect of Risankizumab in patients with moderate-to-severe plaque psoriasis. Design and Methods The EQPS-rMAP method operates in three stages: (1) Eliminating baseline covariate discrepancies through propensity score stratification; (2) constructing stratum-specific MAP priors to dynamically adjust weights for external data; and (3) introducing equivalence probability weights to quantify data conflict risks. The study evaluates the method's performance across six simulated analyses (heterogeneity differences, baseline shifts, etc.), comparing it with traditional methods (MAP, PSMAP, Empirical Bayes MAP) in terms of estimation bias, type I error control, and sample size requirements. Real-world case analyses further validate its applicability.

Results

Simulations demonstrate that EQPS-rMAP maintains estimation robustness under considerable heterogeneity while reducing sample size demands and enhancing trial efficiency. Case analyses confirm its ability to control external bias while preserving high estimation accuracy compared to conventional approaches.

Conclusion

The EQPS-rMAP method provides empirical evidence for the feasibility of hybrid clinical designs. Its methodological advancements—resolving baseline and heterogeneity conflicts through adaptive mechanisms—offer broader applicability for integrating external and RWD across diverse analyses, including bridging trials, MRCTs, and post-marketing studies.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Drug DevelopmentBayes TheoremHumansRandomized Controlled Trials as TopicComputer SimulationResearch DesignMeta-Analysis as TopicPsoriasisRetrospective Studies

Stichwörter der Autor:innen

BayesianHybrid clinical trial designexternal trial datameta-analytic approachespropensity scorereal world data

Autor:innen und Institutionen

  • Department of Biostatistics, School of Public Health, Southern Medical University, Guangzhou, China.
  • BARDS, MSD China, Shanghai, China.
  • BARDS, MSD R&D (China) Co., Ltd, Beijing, China.
  • Hainan Institute of Real World Data, The Administration of Boao Lecheng International Medical Tourism Pilot Zone, Hainan, China.
  • Department of Biostatistics, School of Public Health, Fudan University, Shanghai, China.

ORCID-Profile: Keer Chen, Zengyue Zheng, Ying Wu

Förderung

  • Guangzhou Basic and Applied Foundation Project
  • Real World Research Project Grant Fund from the Hainan Institute of Real World data
  • National Natural Science Foundation of China

Zitieren

APA

Chen, K., Zheng, Z., Zhu, P., Jiang, S., Li, N., Deng, J., Chen, P., Wu, Z., & Wu, Y. (2026). A hybrid prior Bayesian method for combining domestic real-world data and overseas data in global drug development. Statistical Methods in Medical Research. https://doi.org/10.1177/09622802251414586

Vancouver

Chen K, Zheng Z, Zhu P, Jiang S, Li N, Deng J, et al. A hybrid prior Bayesian method for combining domestic real-world data and overseas data in global drug development. Statistical Methods in Medical Research. 2026. doi:10.1177/09622802251414586

BibTeX
@article{Chen2026,
  author = {Chen, Keer and Zheng, Zengyue and Zhu, Pengfei and Jiang, Shuping and Li, Nan and Deng, Jumin and Chen, Pingyan and Wu, Zhenyu and Wu, Ying},
  title = {A hybrid prior Bayesian method for combining domestic real-world data and overseas data in global drug development},
  journal = {Statistical Methods in Medical Research},
  year = {2026},
  doi = {10.1177/09622802251414586}
}
Kostenlos starten →