InfluenceBorrowing: Adaptive Influence-Based Borrowing for Hybrid Control Trials

Implements the adaptive influence-based borrowing framework proposed by Qinwei Yang, Jingyi Li, Peng Wu, and Shu Yang (2026+) in the paper “Improving Treatment Effect Estimation in Trials through Adaptive Borrowing of External Controls" <doi:10.48550/arXiv.2604.13973> for augmenting Randomized Controlled Trials (RCTs) with External Control (EC) data. This package provides a comprehensive workflow to: (1) quantify the comparability of external control samples using influence scores approximated via the influence function of the M-estimator; (2) construct candidate borrowing subsets and select the optimal subset that minimizes the Mean Squared Error (MSE); and (3) calibrate systematic differences in external outcomes using R-learner methods implemented via Ordinary Least Squares or Kernel Ridge Regression.

Documentation:

Downloads:

Linking:

Please use the canonical form https://CRAN.R-project.org/package=InfluenceBorrowing to link to this page.