Text & LLM data · Medical

Medical record chronology review data

Medical chronologies of the kind built for personal injury, malpractice and insurance claims, checked line by line against the source records by registered nurses. Every missed event, wrong date, unsupported statement and wrong page citation is marked and corrected. The records are synthetic, written by clinicians to read like real charts, so no patient’s data is in them.

Reviewed by certified medical professionals

Specification

Records
Synthetic charts: emergency and admission notes, progress notes, imaging and lab reports, therapy notes, bills and medication lists
Chronologies
Dated entries with provider, event, findings and a page citation, as used in injury, malpractice and claims work
Labels
Missed events, wrong dates, wrong providers, statements the records don’t support, wrong citations and pre-existing conditions
Quality check
Registered nurses check every entry against the page it cites; a second nurse reviews each correction
Privacy
Synthetic records written by clinicians: no real patient records, no personal data
Format
Records as PDF and text; chronologies and corrections as CSV or JSON, each tied to its page

Example of the data

What you see, and the file you get.

Preview
20 Feb 2024Clinic visit: neck stiffnesspre-existing · missed
14 Mar 2025ER visit after a car crashcitation p. 12 → p. 14
02 Apr 2025MRI: C5–C6 disc bulgemissed event · added
chronology_review.csvCSV · JSON · Excel
record_setentry_dateprovidereventcited_pagefindingcorrection
case_00012024-02-20Primary careNeck stiffness after lifting, ibuprofen advisedp. 3missed_pre_existingAdded: earlier neck complaint
case_00012025-03-14Emergency departmentNeck pain after rear-end collision, CT negativep. 12wrong_citationSource is p. 14
case_00012025-03-21PhysiotherapyFirst assessment, reduced neck rotationp. 31wrong_dateVisit was on 2025-03-24
case_00012025-04-02RadiologyMRI cervical spine: C5–C6 disc bulgep. 40missed_eventAdded to the chronology
case_00012025-05-09OrthopaedicsPatient reports full recoveryp. 52unsupportedNote says improving, not recovered

Sample rows. Also as JSON or Excel.

Contact us to get the dataset

Tell us what you need. We walk you through the data, its formats and how it is checked on a short call.

Contact us

What it is used for

  • Medical chronology AI

    Train models that turn a stack of records into a dated, cited chronology a nurse only has to check.

  • Citation and hallucination checks

    Measure how often a model invents events, shifts dates or cites the wrong page.

  • Case and claims review

    Teach models to flag pre-existing conditions and gaps in care for injury, malpractice and insurance reviewers.

  • Record summarisation benchmarks

    Compare models and releases on the same records, with answers checked by nurses.

Questions

Are the records real?

No. The records are synthetic, written by clinicians to read like real charts, so they hold no patient’s data.

Who checks the chronologies?

Registered nurses whose licences Avenzoar has verified. They check every entry against the page it cites and mark what is missing or wrong.

Can I see samples first?

Yes. Request samples with the form on this page and we will send them with the full specification.

Tell us what your model needs.

Request samples and the full specification. We reply within a couple of business days.

Request samples