Synthetic Data Digital Twins and Data Trusts Control for Privacy in Health Data Sharing
Bibliographic record
Abstract
Health data sharing is very valuable for medical research since it has the propensity to improve diagnostics, policy, medication, and so on. At the same time, sharing health data needs to be done without compromising the privacy of patients and stakeholders. However, recent advances in AI/ML and sophisticated analytics have proven to introduce biases that can easily identify patients based on their healthcare data, which violates privacy. In this work, we sort to address this major issue by exploring two emerging topics that are gaining attention from industry, academia, and governments, i.e., digital twins and data trusts. First, we proposed the use of digital twins (DTs) to generate synthetic records of patient's heart rate data. DTs are virtual replicas of the actual data and were created using two synthetic data generative models - Gaussian Copula (GC) and Tabular Variational Autoencoder (TVAE). The GC and TVAE achieved a maximum data quality score of 88% and 96% respectively. Next, we posit that the DTs should be shared with a data trusts layer. Data trusts are fiduciary frameworks that govern multi-party data sharing. The data trusts enforce access controls (based on metrics such as location, role-based, and policy-based) to the synthetic health data and reports to the data subject. The preliminary evaluations of the work show that merging the two techniques (i.e., synthetic data digital twins and data trusts) enforces better privacy for health data access. The synthetic data ensures more anonymization while the data trusts provide easy auditing, tracking, and efficient reporting to the patient or data subject. The paper also detailed the architectural design of the data trusts and evaluated the efficiency of the access control techniques.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.009 | 0.036 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.002 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.001 | 0.003 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".