Systematic comparison of bayesian basket trial designs with unequal sample sizes and proposal of a new method based on power priors
Basket trials examine the efficacy of a single intervention in multiple patient groups simultaneously. The assignment of patients to the subgroups, called baskets, is based on matching medical characteristics, which can result in small sample sizes within baskets that are also likely to differ. Spar...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article (Journal) |
| Language: | English |
| Published: |
03 Jun 2025
|
| In: |
Statistics in biopharmaceutical research
Year: 2025, Pages: 1-13 |
| ISSN: | 1946-6315 |
| DOI: | 10.1080/19466315.2025.2486231 |
| Online Access: | Verlag, lizenzpflichtig, Volltext: https://doi.org/10.1080/19466315.2025.2486231
|
| Author Notes: | Sabrina Schmitt, Lukas Baumann |
| Summary: | Basket trials examine the efficacy of a single intervention in multiple patient groups simultaneously. The assignment of patients to the subgroups, called baskets, is based on matching medical characteristics, which can result in small sample sizes within baskets that are also likely to differ. Sparse data complicate statistical inference. A number of Bayesian methods have been proposed in the literature that share information across baskets to increase statistical power. The power prior design incorporates data from all baskets using a weighted likelihood that shares information according to the similarity of the individual baskets. However, with unequal basket sizes, there is a risk of information loss, as the information of small baskets may be superimposed by that of larger baskets. We propose a new weighting method that accounts for differing sample sizes by limiting the amount of information shared between baskets. Considering unequally sized baskets, we systematically compare the power prior design with our new weighting method to existing methods as well as to other competing designs. The results of our simulation study show a superior performance of the power prior design with our new quantity limiting weights in reliably detecting an effect of the intervention under consideration. Supplementary materials for this article are available online. |
|---|---|
| Item Description: | Gesehen am 11.09.2025 |
| Physical Description: | Online Resource |
| ISSN: | 1946-6315 |
| DOI: | 10.1080/19466315.2025.2486231 |