EHA 2025 | Synthetic cohorts and digital twin technology to address elderly underrepresentation in trials

Thank you for the question. Synthetic patient cohorts and digital twin technology offer a transformative way to address the chronic under-representation of elderly patients in clinical trials, especially relevant in geriatric hematology where older adults often make up the majority of the affected population. These technologies use advanced statistical modeling and machine learning to generate synthetic data that mirrors the complex clinical demographic characteristics of real-world patients. A key technique here is data augmentation, which allows researchers to expand existing data sets to include more elderly profiles, ensuring that control groups in virtual trials are more representative. This approach has already been implemented in the GIMEMA activities in the PROMIS Observational International Study on Myelodysplastic Syndromes. A synthetic cohort of 5,000 elderly patients was created using data augmentation based on real-world data from over 900 patients. These synthetic data sets preserve not only clinical variables, but also patients’ reported outcomes, like fatigue, which are especially relevant in older populations. Overall, synthetic cohorts make it possible to ethically include frail and elderly patients by avoiding exposure to suboptimal therapies while improving statistical power, reducing cost and enhancing the external validity of trial findings.

This transcript is AI-generated. While we strive for accuracy, please verify this copy with the video.