Commercial or industrial use of mental health data for research: primer and best-practice guidelines from the DATAMIND patient/public Lived Experience Advisory Group
Bibliographic record
Abstract
BACKGROUND: Routinely collected health data, such as that held by the United Kingdom (UK) National Health Service/Health and Social Care (collectively "NHS"), has important research uses, but its appropriate use requires public trust and transparency. Commercial/industrial access to routinely collected health data is especially controversial and sensitive for the public, and particular concerns may relate to mental health (MH) data. Existing best-practice MH data science guidelines do not cover commercial uses specifically, but emphasise the importance of patient/public co-development of data science. OBJECTIVES: To develop patient/public-led guidelines for the commercial/industrial use of MH data for research, and to capture relevant background information required by patient/public participants. The focus was on the UK and its constituent nations, but the principles may have wider applicability. METHODS: A patient/public lived experience advisory group (LEAG) was set up within DATAMIND, the Health Data Research UK data hub for MH informatics research development. Initial training and discussion yielded a requirement for definitions and explanations of concepts and processes relating to MH data research, developed iteratively. Subsequently, the LEAG developed guidelines via a qualitative and iterative quasi-Delphi approach. The agreed scope excluded data provided for research with informed consent, data processing arrangements such as companies hosting electronic health records or e-mail systems on the instruction of health services, or compliance with legal minimum requirements. The scope included the use of routinely collected MH data (e.g. NHS data) for research by commercial/industrial organisations without explicit consent, and aspects of MH data collection directly by industry with consent. RESULTS: Alongside the primer in MH data research concepts, the LEAG provide recommendations and best-practice guidelines relating to commercial/industrial research use of MH data, for organisations controlling MH data (such as NHS bodies) and for commercial applicants seeking to use MH data for research. Alongside principles of transparency, patient rights, patient/public involvement in research, stringent governance, and statistical disclosure control, the guidelines recommend a risk-benefit approach to assessing applications for data use, within limits that include avoiding the export of unconsented patient-level data outside NHS-controlled secure data environments, and not providing access to unconsented free-text MH data to commercial applicants. We also provide some recommendations for NHS executive and regulatory bodies, relating to public choice and transparency, clarity of guidance to research-active NHS organisations, and support for de-identification. CONCLUSIONS: Patient/public involvement and understanding is central to MH data research. The primer materials developed here constitute information requested by public advisers prior to considering best practice. The guidelines reflect the views of people with personal or family experience of mental ill health. We hope they are of practical use to the wider MH research community and serve to increase public transparency and trust.
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 | 0.467 | 0.357 |
| Meta-epidemiology (narrow) | 0.002 | 0.004 |
| Meta-epidemiology (broad) | 0.003 | 0.004 |
| Bibliometrics | 0.008 | 0.009 |
| Science and technology studies | 0.009 | 0.020 |
| Scholarly communication | 0.022 | 0.026 |
| Open science | 0.014 | 0.036 |
| Research integrity | 0.031 | 0.032 |
| Insufficient payload (model declined to judge) | 0.008 | 0.009 |
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".