Benchmark RadarCoMedBench
CoMedBench evaluates synthetic medical data fidelity and downstream utility across 37 dataset-task pairs from seven public sources, using a common clinical-validity framework and shared training and evaluation engine.
- Released
- 2026-08-13
- Readiness
- Paper only
- Primary field
- Health & Life Sciences
Why it matters
Provides a reproducible benchmark for comparing synthetic data generators across multiple datasets and tasks, addressing the lack of comprehensive evaluation in prior studies and aiding in deciding when synthetic data is viable for model development.
Motivation
Access to clinical data is essential for developing reliable healthcare machine learning systems, but direct use of electronic health records is constrained by privacy regulation, institutional review, data-use agreements, and the risk of re-identification.
Primary resources
Benchmark Radar records only publicly supported details and links back to primary sources for verification.