What Trenthos IME is
Trenthos IME reads the medico-legal brief behind an independent medical examination, a file review or a supplementary report, and builds a dated chronology with a document and page reference on every entry. Behind most of these reports sits an evening of unglamorous work: reading several hundred pages of records, putting events in order, and hunting down the page where the MRI report actually said what you remember it saying. Trenthos IME does that reading and indexing, and it shows where every entry came from.
The division of labour is strict: the system reads, dates and indexes; the examiner examines, reasons, forms the opinion and writes the report. Trenthos IME does not diagnose, give a clinical opinion or rate an impairment, and it does not write the report. Doctor in, doctor out.
A brief is held as a matter, and everything the system produces from it is traceable back to the pages it came from.
How it is built
Trenthos IME is a web application: examiners work entirely in the browser, with nothing to install. The front end is a single-page application; the back end is a structured API over a PostgreSQL database. Reading the pages and building the chronology run as background jobs through a work queue, so uploads never block the interface, and documents live in dedicated object storage, encrypted at rest.
The product runs on infrastructure Trenthos controls in Australia, with every component (application, database, document storage and AI inference) onshore. Shipping
The pipeline
A matter is processed in ordered stages (ingest, reading the pages, extraction, chronology) and the state of each stage is visible per matter. Shipping
Ingest
Documents are uploaded as they arrived: PDFs, typed or scanned. There is no need to sort or split them first. Every file is validated, classified by kind (letter of instruction, clinical records, investigation, prior report, correspondence) and assigned a stable document number. That “D3” is what every later citation points back to.
Reading every page
Each page is read individually. Where a PDF carries real embedded text, that text is used directly; scanned pages are read by optical character recognition. Every page records which method read it and how confidently, and the examiner can see the method on the page. Documents are indexed page by page precisely so that citations can be checked page by page.
Extraction
The system moves a window across each document and proposes typed entries: events, investigations, treatments, prior opinions and claim documents, each with a date, a provider where the pages name one, a page number, and a short verbatim quote from the source. A date the pages do not support is left as written rather than placed on the timeline. An entry with no date of its own is placed by the date of the note or document it appears in, and says so.
Chronology
The chronology itself is assembled by plain code, not by a model: entries are ordered, duplicates are merged with every citation kept, and the system raises flags a reader would want: gaps of more than 90 days in the records, history predating the date of injury, and treatment appearing before the first recorded presentation.
Grounding and traceability
The product is built around one rule: no chronology entry without a source page. What makes that more than a slogan is that the check is done mechanically in code, never asserted by the model. When the AI proposes an entry with a quote and a page, deterministic code checks that the quote actually appears on that page; an entry whose quote cannot be found there is never silently accepted. It is kept, marked “Quote not found on page”, and left for a person to check. Verification itself stays with you: every citation opens its source page so you read the source yourself, the pathway the court practice notes require. Shipping
Honest gaps
A chronology that quietly reads as more than it is becomes the easiest kind to take apart, so the system is deliberately conservative about what it claims: Shipping
- Flagged entries stay in the chronology. An entry whose quoted words cannot be found on its page is marked and kept, so you know to look.
- Undated entries are labelled. An entry with no date of its own is labelled undated, or shows which note or document its date came from.
- Unread pages are listed. Pages the extraction could not read after its retries are named in the chronology, so nothing from them is assumed.
- No opinion. The chronology records what the pages say. It contains no diagnosis, causation, prognosis or impairment rating.
Finding things in the brief
The questions in the letter of instruction are read with the brief. Pick a question, or search a topic such as work capacity or driving, and every matching entry from every document comes up in date order, each with its page. The search matches the words and their usual variants; it does not rank or select entries by meaning, so what you see is every entry that matches. Shipping
Matters are set up under the Australian workers’ compensation schemes: WorkSafe Victoria, SIRA (NSW), WorkCover Queensland, ReturnToWork SA, WorkCover WA, WorkSafe Tasmania, NT WorkSafe, WorkSafe ACT and Comcare. The chronology is built from the records themselves, so it works the same way for an independent medical examination, a file review or a supplementary brief.
Using the chronology
The examiner opens the chronology in their Trenthos IME account, after two-factor sign-in and their PIN. Clicking any citation opens the cited page, with the quoted words highlighted. A concise view shows the entries most examiners need first, and you can add any entry to it or take one out; nothing is deleted. Shipping
The chronology downloads as a PDF or Word document, every entry dated and cited to its document and page, ready to sit alongside your report. An AI-use disclosure annexure, compiled from the system’s own records rather than written by a model, downloads with it. Shipping
Security and access
Sessions use short-lived access tokens with rotating refresh tokens and reuse detection, so a stolen token dies quickly and a replayed one revokes the whole family. Multi-factor authentication (TOTP) is required for clinical routes, an idle session locks and requires a PIN to resume, and passwords are hashed with argon2id. Authentication endpoints are rate-limited and designed not to leak which accounts exist. Shipping
Access is role-based - admin, examiner, support, reviewer - and all traffic is encrypted in transit with TLS.
Tenancy and audit
Isolation between practices is enforced in the database itself: PostgreSQL row-level security is enabled and forced on every clinical table, so a query outside the active tenant's context returns nothing - even application bugs cannot cross tenants. Shipping
Every insert, update and delete on clinical data is captured by database triggers with before-and-after snapshots; reads of clinical content are logged at the API layer; and every AI call is logged with its model and usage, and the full record of each call is kept so that it can be produced if a court requires it. Records are soft-deleted rather than destroyed mid-engagement.
Data residency and AI
Everything is onshore. Processing and storage run in Australia, and - unusually for a product in this category - AI inference currently runs on Trenthos-controlled hardware in Australia: case content is not sent to third-party AI services, and no customer data is ever used to train models. Shipping
The data model is built for de-identification: a matter carries a claimant label of the practice’s choosing rather than identity fields, so there is nowhere to put a Medicare number, a home address or a date of birth. On a trial, identifying details are redacted from the brief on Trenthos systems before the chronology is built; the identified originals are deleted within 7 days of delivering the chronology, everything else 30 days after delivery, and every backup copy within 21 days after that. Shipping
Getting a brief in
A matter starts with the letter of instruction and the records as they arrived, with no need to sort or split them. The pipeline classifies, reads and indexes what it is given, and the chronology follows.
For a trial, you create an account, accept a confidentiality undertaking and upload the brief in the app; never as an email attachment. Nothing is processed until we have checked your Ahpra registration on the public register. The brief is redacted on Trenthos systems in Australia before the chronology is built, and the identified originals are deleted within 7 days of delivering the chronology.
Status
Trenthos IME is open to Australian examiners. The pipeline is tested end to end on synthetic briefs, and the API is covered by an automated test suite run in continuous integration against real services. Shipping
Security certifications (SOC 2, ISO 27001) are not claimed until they are held.
Trying it free
The honest test is your own brief, checked against your own reading of it. Try it free on up to 50 pages: create an account on the sign-up page with your Ahpra registration number and specialty, and send your brief in the app. We confirm your registration and build your preview, with the claimant’s identity hidden. You get an email when it is ready. Questions first? Email support@trenthos.com.