Improving Cardiac Insights: Harnessing Privacy-Preserving Record Linkage (PPRL) to Obtain, Link, and Enhance Data for a Healthier Australia
Bibliographic record
Abstract
ObjectiveGeneral Practices (GPs) collect vital information on health and disease in Australia, however much of this sensitive data is difficult to share for linkage purposes due to confidentiality and privacy concerns. PPRL techniques have presented an opportunity to work with GP’s, and better understand health conditions such as negative cardiac outcomes in Australia. ApproachWe utilised two algorithms, GHRANITE and Bloom Filters, to hash 2.5 million records from over 100 GPs. Both approaches utilised the same basic principles: data custodians pass identifying data through an irreversible hashing algorithm, and generate a unique key for each record. The hashing is done at the practice level ensuring that original identifying patient information never leaves GP's premises. Hashed data can be sent elsewhere, and if the same algorithm is used on a different set of data, linkage can be performed on both sets to identify individuals without exposing identifying information. ResultsThe PPRL process involved encoding over 150 million records from diverse datasets, encompassing hospitals, housing, and other sources. Initial results indicated 70.90% linkage utilising GHRANITE, and 84.10% linkage utilising Bloom Filters. Moving forward, the goal is to provide the linked data back to GPs, enabling them to identify high-risk patients and implement targeted interventions or additional measures. ConclusionThrough the successful implementation of PPRL, the project has made strides in overcoming data sharing barriers while safeguarding confidentiality. By harnessing innovative algorithms, the initiative has paved the way for more insights into cardiac outcomes and the potential for proactive healthcare interventions.
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 | 0.058 | 0.143 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.004 | 0.005 |
| Science and technology studies | 0.002 | 0.002 |
| Scholarly communication | 0.008 | 0.012 |
| Open science | 0.003 | 0.017 |
| Research integrity | 0.002 | 0.003 |
| Insufficient payload (model declined to judge) | 0.003 | 0.002 |
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".