Article Impact Level: HIGH Data Quality: STRONG Summary of NPJ Digital Medicine. https://doi.org/10.1038/s41746-025-01705-z Dr. Nansu Zong et al.
Points
- Electronic health record data complexities often create discrepancies between real-world evidence and randomized controlled trials.
- This study aims to enhance efficacy prediction for repurposed heart failure drugs evaluated in clinical trials.
- A new framework integrates drug-target predictions with EHR-based Emulation Trials to derive surrogate endpoints.
- The novel drug-target prediction model demonstrated superior performance against the BETA benchmark, surpassing baseline algorithms.
- The framework accurately identified 17 repurposed drugs from 266 phase 3 heart failure trials using 59,000 Mayo Clinic patient records.
Summary
This study addresses the inherent complexities of Electronic Health Record (EHR) data, which often lead to discrepancies between real-world evidence and Randomized Controlled Trials (RCTs), particularly in evaluating treatment efficacy. The objective was to improve the prediction of efficacy direction for repurposed heart failure (HF) therapies tested in RCTs. A novel efficacy prediction framework was developed, integrating drug-target predictions with EHR-based Emulation Trials (ETs) to establish surrogate endpoints utilizing HF prognostic markers.
The proposed drug-target prediction model underwent rigorous validation against the BETA benchmark, demonstrating superior performance compared to existing baseline algorithms. The framework’s predictive accuracy was further evaluated by identifying 17 repurposed drugs derived from a cohort of 266 phase 3 HF RCTs. This evaluation utilized data from 59,000 patients treated at the Mayo Clinic, showcasing the framework’s remarkable ability to predict efficacy direction.
These findings suggest that the integrated framework can significantly enhance the prediction of treatment efficacy for repurposed drugs in HF, potentially bridging the gap between real-world data and RCT outcomes. This advancement could streamline the repurposing process for HF therapies, improving patient outcomes and resource allocation.
Link to the article: https://www.nature.com/articles/s41746-025-01705-z
References
Zong, N., Chowdhury, S., Zhou, S., Rajaganapathy, S., Yu, Y., Wang, L., Dai, Q., Li, P., Liu, X., Bielinski, S. J., Chen, J., Chen, Y., & Cerhan, J. R. (2025). Advancing efficacy prediction for electronic health records based emulated trials in repurposing heart failure therapies. Npj Digital Medicine, 8(1), 1–13. https://doi.org/10.1038/s41746-025-01705-z
