Evaluating Common De-Identification Heuristics for Personal Health Information
Bibliographic record
Abstract
BACKGROUND: With the growing adoption of electronic medical records, there are increasing demands for the use of this electronic clinical data in observational research. A frequent ethics board requirement for such secondary use of personal health information in observational research is that the data be de-identified. De-identification heuristics are provided in the Health Insurance Portability and Accountability Act Privacy Rule, funding agency and professional association privacy guidelines, and common practice. OBJECTIVE: The aim of the study was to evaluate whether the re-identification risks due to record linkage are sufficiently low when following common de-identification heuristics and whether the risk is stable across sample sizes and data sets. METHODS: Two methods were followed to construct identification data sets. Re-identification attacks were simulated on these. For each data set we varied the sample size down to 30 individuals, and for each sample size evaluated the risk of re-identification for all combinations of quasi-identifiers. The combinations of quasi-identifiers that were low risk more than 50% of the time were considered stable. RESULTS: The identification data sets we were able to construct were the list of all physicians and the list of all lawyers registered in Ontario, using 1% sampling fractions. The quasi-identifiers of region, gender, and year of birth were found to be low risk more than 50% of the time across both data sets. The combination of gender and region was also found to be low risk more than 50% of the time. We were not able to create an identification data set for the whole population. CONCLUSIONS: Existing Canadian federal and provincial privacy laws help explain why it is difficult to create an identification data set for the whole population. That such examples of high re-identification risk exist for mainstream professions makes a strong case for not disclosing the high-risk variables and their combinations identified here. For professional subpopulations with published membership lists, many variables often needed by researchers would have to be excluded or generalized to ensure consistently low re-identification risk. Data custodians and researchers need to consider other statistical disclosure techniques for protecting privacy.
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 | 0.092 | 0.278 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.005 | 0.003 |
| Science and technology studies | 0.002 | 0.003 |
| Scholarly communication | 0.004 | 0.005 |
| Open science | 0.003 | 0.003 |
| Research integrity | 0.002 | 0.002 |
| 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".