Sharing traumatic stress research data: assessing and reducing the risk of re-identification
Bibliographic record
Abstract
Background: FAIR Data practices support data sharing and re-use and are essential for advancing science and practice to benefit individuals, families, and communities affected by trauma. In traumatic stress research, as in other health and social science research, ethical, legal, and regulatory frameworks require careful attention to data privacy. Most traumatic stress researchers are aware of basic methods for de-identifying/anonymising datasets that are to be shared. But our field has not generally made use of systematic, data analytic approaches to reduce the risk of re-identification of study participants or disclosure of personal or sensitive information.Objective: To facilitate safe and ethical data sharing by better preparing traumatic stress researchers to systematically assess and reduce re-identification risk using contemporary data analytic methods.Method: In two case studies using publicly available trauma research datasets from international, multi-language projects, we applied a systematic approach guided by the Checklist for Reducing Re-Identification Risk in Traumatic Stress Research Data.Results: For each case study dataset, we identified specific recommended actions to further reduce the risk of re-identification, and we then communicated these recommendations to the original investigators. After implementing the recommended changes, each dataset is judged to be at very low re-identification risk.Discussion: The particular nature of traumatic stress research, i.e. its content, data, and study designs, can influence the likelihood and potential impact of re-identification or disclosure. The two worked case examples in this paper demonstrate the utility of applying a systematic approach to assess and further mitigate re-identification risk in shared datasets. At each stage of the research data lifecycle, there are research practices and choices relevant to reducing re-identification risk. This paper presents practical tips for research teams to facilitate FAIR data practices while attending to data 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 | 0.595 | 0.791 |
| Meta-epidemiology (narrow) | 0.002 | 0.002 |
| Meta-epidemiology (broad) | 0.002 | 0.004 |
| Bibliometrics | 0.014 | 0.008 |
| Science and technology studies | 0.017 | 0.020 |
| Scholarly communication | 0.015 | 0.027 |
| Open science | 0.010 | 0.040 |
| Research integrity | 0.008 | 0.010 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
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; the direct Gemma label and the distilled Codex classifier agree on what is shown here.
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".