Neural networks speed up adaptive data borrowing in drug studies
Amortized Data Borrowing with Exchangeability-Aware Neural Posterior Estimation
Machine Learning
Summary
Combining small new studies with older similar data can help drug research, but traditional methods are slow and need careful expert setup. The authors explore training a neural network to quickly estimate how much to trust external data based on many simulated examples. Their method is much faster than usual approaches and works better when study results differ between groups. They also tested it on Alzheimer's data, where it correctly adjusted risk estimates without overstating confidence.
What this means in practice
- •For biostatisticians in drug development: Obtain rapid and adaptive Bayesian borrowing of external clinical data for small drug trials with variable outcomes.
- •For healthcare data analysts: Analyze combined clinical cohorts with differing outcome distributions to better estimate risk factors while controlling error rates.
Authors
Chin-Hung Huang, JooChul Lee, Huan He
Abstract
Augmenting small concurrent studies with external or historical cohorts is attractive in drug development, where enrollment is slow, follow-up is expensive, and closely related trial or real-world data are often already available. Bayesian dynamic borrowing (BDB) provides a principled framework for adaptively controlling the influence of external data, but classical implementations often depend on hand-specified priors and MCMC-based inference, which can be computationally expensive and not generalizable. In this work, we study amortized neural posterior estimation (NPE) as a flexible alternative. A single network is pretrained on simulated current/external dataset pairs spanning covariate shift, outcome drift, and joint non-exchangeability, and then returns an approximate posterior for a scalar current-study target in a single forward pass. Through simulation studies, we find that NPE is most useful under outcome drift and joint mismatch: in the harder outcome-drift regimes, it gives up to about five-fold lower absolute bias than the best classical baseline and keeps Type I error close to nominal. After pretraining, posterior summaries are obtained in about 8 ms per dataset, roughly $10^3\times$ faster than MCMC-based borrowing baselines in our timing experiment. We further analyze Alzheimer's Disease Neuroimaging Initiative (ADNI) data and show that, when mild cognitive impairment outcomes differ across cohorts, the NPE formulation recovers the later-cohort risk level in this example without claiming greater precision. Code is available at https://github.com/ChinHungScott/NPE-for-Bayesian-Dynamic-Borrowing-MLHC-.