About this tool

Medical Report De-identifier

Replace common patient, clinician, and facility identifiers in pasted text, TXT, DOCX, or text-based PDF files locally in your browser.

The Medical Report De-identifier finds and replaces labelled identifiers in clinical text across nine categories — patient names, MRN and patient IDs, dates of birth, phone numbers, email addresses, addresses, other dates, clinician names and facility names — swapping each one for a numbered placeholder such as [PATIENT_NAME_1]. The same value always gets the same placeholder, so relationships in the report survive while the identity does not. Text you paste, or a TXT, DOCX or text-layer PDF you load, is processed entirely in your browser; this is a redaction aid, not a certified HIPAA de-identification service.

How to use Medical Report De-identifier

1. 1

   Open Medical Report De-identifier on AltFTool — it loads instantly in your browser.

2. 2

   Paste the report under 'Report text', or press 'Choose file' for a TXT, DOCX or text-layer PDF up to 15 MB — 'Load demo' fills in the synthetic sample.

3. 3

   Switch any of the nine identifier categories off, or use 'Select all' / 'Select none', and watch the Anonymized preview rebuild with tokens like [PATIENT_NAME_1].

4. 4

   Use 'Copy text' or 'Download TXT' for the scrubbed report, and 'Download counts JSON' for a counts-only log that carries no source text.

Why use Medical Report De-identifier

• Consistent pseudonyms, not blanket blackouts

  Repeat mentions of a value collapse to one placeholder using NFKC and case-insensitive matching, so "Dr. Rao" and "dr rao" become the same token.

• Category-by-category control

  Each of the nine categories can be switched off independently — keep visit dates for a chronology while still removing names and IDs.

• A log that leaks nothing

  The counts-only export is explicitly built without the source text or any detected value, so it is safe to attach to a ticket or audit trail.

Common use cases

• You want to share a discharge summary with a colleague for a second opinion and need the patient's name, MRN and contact details out of it before it leaves your machine.
• You are assembling teaching cases from real reports and need the same patient referred to consistently throughout a document — every mention of one name becoming the same numbered token rather than a different one each time.
• You need to record that a document was scrubbed without recording what was in it, so you export the counts-only report that lists how many identifiers were replaced per category and nothing else.

Frequently asked questions