Challenges of analyzing synthetic tabular data generated from 115 phase 3 oncology trials
- 1Synthetic Data for Research Acceleration (SYNDARA), Ghent University Hospital
Abstract
Synthetic data have been proposed to facilitate data sharing in privacy-sensitive contexts, including clinical trials. It remains unclear, however, how well original treatment effect estimates can be replicated in synthetic data analyses. Therefore, we synthesized and reanalyzed 128 treatment comparisons from 115 phase 3 randomized oncology trials using sixteen different generative models. Our findings demonstrate that careful methodological choices are essential for drawing valid statistical conclusions from synthetic data analyses. Naive analyses frequently yield falsely significant treatment effects, occurring in up to half of the trials created by deep generative models. Correcting standard errors to reflect the uncertainty inherent in synthetic data generation reduces these false positives, but primarily suffices for trials generated by parametric models. Although this correction entails some power loss, it can be mitigated by increasing the synthetic sample size. Thus, at present, large synthetic trials generated by parametric models and analyzed with corrected standard errors are more likely to preserve inferential utility. Advancing valid statistical inference from synthetic data created by deep generative models remains an important direction for future research.
Related
- Beyond Distributional Fidelity: Causal-Penalized Diffusion for Synthetic Tabular Data
- Seq2Synth: Benchmarking Temporal Fidelity in Synthetic Sequential Tabular Data
- Classical methods for modern applications: synthetic tabular data generation using lightweight discrete response continuous latent variable model
- SynthGuard-ReleaseBench: Locked-Audit Evidence for Synthetic Tabular Data Releases
- Synthetic data for open medical research: current challenges and perspectives
BibTeX
@article{decruyenaere2026challenges,
title = {{Challenges of analyzing synthetic tabular data generated from 115 phase 3 oncology trials}},
author = {Alexander Decruyenaere and Christiaan Polet and Johan Decruyenaere and Heidelinde Dehaene and Paloma Rabaey and Thomas Demeester and Stijn Vansteelandt and Sylvie Rottey},
journal = {Scientific Reports},
year = {2026},
month = jul,
doi = {10.1038/s41598-026-62212-z},
url = {https://doi.org/10.1038/s41598-026-62212-z},
note = {CC BY},
}
RIS
TY - JOUR
TI - Challenges of analyzing synthetic tabular data generated from 115 phase 3 oncology trials
AU - Alexander Decruyenaere
AU - Christiaan Polet
AU - Johan Decruyenaere
AU - Heidelinde Dehaene
AU - Paloma Rabaey
AU - Thomas Demeester
AU - Stijn Vansteelandt
AU - Sylvie Rottey
PY - 2026
DA - 2026/07/01
JO - Scientific Reports
DO - 10.1038/s41598-026-62212-z
UR - https://doi.org/10.1038/s41598-026-62212-z
AB - Synthetic data have been proposed to facilitate data sharing in privacy-sensitive contexts, including clinical trials. It remains unclear, however, how well original treatment effect estimates can be replicated in synthetic data analyses. Therefore, we synthesized and reanalyzed 128 treatment comparisons from 115 phase 3 randomized oncology trials using sixteen different generative models. Our findings demonstrate that careful methodological choices are essential for drawing valid statistical conclusions from synthetic data analyses. Naive analyses frequently yield falsely significant treatment effects, occurring in up to half of the trials created by deep generative models. Correcting standard errors to reflect the uncertainty inherent in synthetic data generation reduces these false positives, but primarily suffices for trials generated by parametric models. Although this correction entails some power loss, it can be mitigated by increasing the synthetic sample size. Thus, at present, large synthetic trials generated by parametric models and analyzed with corrected standard errors are more likely to preserve inferential utility. Advancing valid statistical inference from synthetic data created by deep generative models remains an important direction for future research.
SN - 2045-2322
ER -