NHANES Synthetic Health and Demographics Dataset
Submitted September 9, 2026
This synthetic dataset comprises 4,899 records with 18 health and demographic variables including age, BMI, cholesterol levels, and HbA1c, derived from the National Health and Nutrition Examination Survey using the SPHERE synthesis algorithm. The dataset achieves high fidelity (95.4/100) with excellent preservation of means, variances, and correlations (KS 94.6), while maintaining strong privacy guarantees (98.4/100) across singling-out, linkability, and inference risks. It is suitable for epidemiological research, health outcome modeling, and algorithm development without exposing individual-level sensitive information.
Quality scores
Fidelity — composite 95.4
- Δ mean
- 100.0
- Δ variance
- 100.0
- Δ correlation
- 87.0
- KS statistic
- 94.6
Privacy — composite 98.4
- Singling-out
- 100.0
- Linkability
- 97.4
- Inference
- 97.8
Hosted by the dataset author. SPHERE World does not store the data itself.
Evaluation report generated by SPHERE, stamping SHA-256 fingerprints of both files. Tamper-evident, not cryptographically signed.
Data Cooperative member. This author has authorized inclusion of their synthetic data in curated pooled datasets.
Cite SPHERE
This synthetic dataset was generated with SPHERE. If you use it, please cite the SPHERE preprint.
Zihuai He, Junyoung Park, Rafael Catoia Pulgrossi, Justin Lee, Robert R Butler III, Audrey Weber, Lu Tian, Xiang Zhang, Julie Fangran Wang, Sharon Sha, Elizabeth C. Mormino, Tony Wyss-Coray, Victor W. Henderson, Frank M. Longo, James Zou, Manisha Desai, Russ Altman. Unlocking Sensitive Data with SPHERE in the Age of AI. bioRxiv, 2026. doi:10.64898/2026.09.01.748580
BibTeX
@article{he2026sphere,
title = {Unlocking Sensitive Data with {SPHERE} in the Age of {AI}},
author = {He, Zihuai and Park, Junyoung and Pulgrossi, Rafael Catoia and Lee, Justin and
Butler, III, Robert R and Weber, Audrey and Tian, Lu and Zhang, Xiang and
Wang, Julie Fangran and Sha, Sharon and Mormino, Elizabeth C. and
Wyss-Coray, Tony and Henderson, Victor W. and Longo, Frank M. and Zou, James and
Desai, Manisha and Altman, Russ},
journal = {bioRxiv},
year = {2026},
doi = {10.64898/2026.09.01.748580},
url = {https://doi.org/10.64898/2026.09.01.748580},
note = {Preprint}
}Have sensitive data of your own?
SPHERE makes privacy-preserving synthetic twins of sensitive tabular data. Free for non-commercial academic research · macOS (Apple Silicon).