Abstract:Generating synthetic tabular health data is challenging, and evaluating their quality is equally, if not more, complex. This systematic review highlights the critical importance of rigorous evaluation of synthetic health data to ensure reliability, clinical relevance, and appropriate use. From an initial identification of 2067 relevant papers published in the last ten years, 134 studies were selected for detailed analysis. Our review identifies key challenges, including lack of consensus on evaluation methods, inconsistent application of evaluation metrics, limited involvement of domain experts, inadequate reporting of dataset characteristics, and limited reproducibility of results. In response, we provide a structured consolidation of synthetic data generation and evaluation methods into taxonomies, alongside practical guidelines to support more robust and standardised evaluation practices. These findings aim to support the responsible development and use of synthetic health data, aligned with emerging expectations around transparency, reproducibility, and governance, ultimately enabling the community to fully harness its transformative potential and accelerate innovation.
| Comments: | 32 pages |
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computers and Society (cs.CY) |
| Cite as: | arXiv:2504.18544 [cs.LG] |
| (or arXiv:2504.18544v3 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2504.18544 arXiv-issued DOI via DataCite |
Submission history
From: Nazia Nafis Ms [view email]
[v1]
Thu, 10 Apr 2025 02:48:20 UTC (21,276 KB)
[v2]
Thu, 11 Sep 2025 16:27:30 UTC (8,713 KB)
[v3]
Thu, 14 May 2026 13:41:04 UTC (3,987 KB)