Trial dataset anonymisation is the removal of participant identifiers before a file is shared. It supports MHRA clinical-trial transparency requirements and ICH E6(R2). anonym.plus runs offline and keeps the measured values readable.
When this applies
The MHRA asks a sponsor to share participant-level results. Names, free-text fields, and rare dates must be hidden first.
How anonym.plus handles it
- Load the file (CSV, XLSX, PDF, or DOCX) into anonym.plus.
- The tool scans columns and free text for direct identifiers.
- Local OCR pulls text from any scanned supporting page.
- Confirm the flagged participant names, dates, and places.
- Replace each one with a steady token across the whole file.
- Save the cleaned copy on your device with no network call.
What you need to provide
- The file (CSV, XLSX, PDF, DOCX, or scan).
- An operator: Replace for tokens, Redact to drop a field.
- Optional: a token map for consistent labels across exports.
Patient data entity types detected
| Category | anonym.plus entity type | Example |
|---|---|---|
| Names | PERSON | James O’Brien → [PARTICIPANT_1] |
| Event dates | DATE_TIME | Randomised 02/02/2026 → [DATE] |
| Location | LOCATION | Manchester, England → [REGION] |
| EMAIL_ADDRESS | j.obrien@example.co.uk → [EMAIL] | |
| Free-text ID | ID | Screening SCR-0091 → [SCREEN_ID] |
| NHS number | MEDICAL_RECORD_NUMBER | NHS 512 345 6789 → [LOCAL_ID] |
Compliance achieved
- Supports the duties in the HRA's Make it Public research-transparency strategy — public registration before the first participant, and a summary result within 12 months of the end of the trial.
- Research reuse of participant-level data rests on UK GDPR Art. 9(2)(j) read with the Art. 89(1) safeguards.
- DPA 2018 s.19 conditions those safeguards: the processing must not be likely to cause substantial damage or distress, and must not support measures or decisions about a particular participant.
- Once no participant can be singled out, UK GDPR Recital 26 puts the file outside data-protection scope — judge that with the ICO's motivated-intruder test.
- Runs offline — no cloud upload, no data-processor contract; working copies stay under AES-256-GCM.
Anonymise clinical trial datasets offline — see plans & start free →
Limitations & cautions
MHRA transparency expects a risk assessment, not just field removal. The tool strips direct identifiers and flags rare values like an unusual age. You still judge whether quasi-identifiers in combination could re-identify a participant before release.
Frequently asked questions
What does MHRA transparency require for trial data?
The HRA’s Make it Public strategy expects a trial to be registered publicly before the first participant is recruited and a summary result to follow within 12 months of the end of the trial. Anything participant-level released alongside it must carry the UK GDPR Art. 89(1) safeguards, with identifiers removed and the method documented.
Can it process spreadsheet columns?
Yes. Load a CSV or XLSX and the tool scans both column values and free text. It applies the same token to a repeated name across every row.
Is the published file still useful for analysis?
Yes. Numeric outcomes and timing offsets remain. Only direct identifiers are swapped, so the file keeps its scientific value.