What is RFAN?

• RFAN is a novel two-stage clinical trial design that meets regulatory requirements while enhancing real-world outcomes and fairness.


• It begins with a randomized stage, followed by an adaptive phase driven by active learning and Bayesian uncertainty estimation.


• The design introduces two key post-trial objectives: Post-Trial Mean Benefit (PTMB) and Post-Trial Fairness (PTF).

Where can I find the paper and code?

• The paper is available at https://proceedings.mlr.press/v258/klein25a.html


• Code and implementation are available on GitHub.

How to cite RFAN?

• Klein, O. N., Hüyük, A., Shamir, R., Shalit, U., & van der Schaar, M. (2025, April). Towards Regulatory-Confirmed Adaptive Clinical Trials: Machine Learning Opportunities and Solutions. In International Conference on Artificial Intelligence and Statistics (pp. 4969–4977). PMLR.

Need help?

• For questions or feedback, please contact Omer Noy Klein: omernoy@mail.tau.ac.il