Clinical coding that stays inside the classification

Manual code navigation, live DRG grouping and an AI autocoder that works the way a coder does: search the index, follow the cross-references, verify in the tables. Every code it keeps traces back to the document and the sentence it came from. Today it codes in ICD-10-CM and ICD-10-PCS and groups to MS-DRG, which is what these screens show. ICD-10-AM, ACHI and AR-DRG follow certification.

Auto coding an inpatient cholecystectomy episode: what HealthOS holds on the admission, the agent’s pre-admission pass, and the codes it kept.

From the record to the claim

Manual coding

Index search first, then code table verification, as the coding standards require.

AI autocoding

An agent that navigates the classification and refuses any code it can’t prove.

Live DRG

The grouping updates with every code, so funding impact is visible as you work.

Into the claim

A finalised code set and DRG move to the claiming queue with no re-keying.

What it does

Code faster, and prove every code

Manual coding

Search the index, verify in the tables

Type a term such as “cellulitis” and the Code Finder searches the classification index, showing lead terms, modifiers and cross-references as they appear in print. Follow a reference into the tables, drill to the character level the code needs, and read inclusions, exclusions and coding notes inline.

Diagnosis and procedure codes in the same interface
Procedure codes built axis by axis: body part, approach, device and qualifier
Every assignment timestamped and attributed in the coding timeline

Building a procedure code in the Code Finder, choosing body part, approach, device and qualifier.

AR-DRG in certification

Live DRG as you code

As codes are added the grouper runs and returns the Diagnosis Related Group that drives funding. The result updates with every change, so a coding decision and its funding impact are visible together.

The coding worksheet with an ICD-10-CM diagnosis and an ICD-10-PCS procedure, and the live MS-DRG with its cost weight.

AI autocoding

An agent that codes the way a coder does

The autocoder reads the clinical record for an episode: discharge summaries, progress notes, procedure reports and pathology. For each condition or procedure it searches the code index, follows every cross-reference, then verifies the candidate in the code tables. A code it can’t prove it navigated to is refused. Nothing is recalled from training data.

A pre-admission pass codes only from what was on file before the patient arrived
A post-discharge pass codes from the full episode record
Each pass is independent and produces its own traceable result

The autocoder’s view of an episode: documents on file before admission and from the stay, the pass, and the codes kept.

Every code traces back to its source

Each code shows the document, the sentence and the classification path behind it: the search term, the index result, the cross-reference and the tabular check. Patient identifiers are removed before any text goes to the model, and the screen shows what was removed and what was sent.

How the agent got there: each search of the index and each verification in the tabular list, step by step.

The coder decides

The autocoder’s output is a draft. The coder reviews each code, then accepts, changes or rejects it, and the grouper re-runs on the accepted set. Every run is logged: which model, which documents were read, which codes were proposed and accepted, and the final grouping. Your organisation chooses the model that powers the agent, separately from other AI features.

Eclipse in NOI testing

Coding feeds the claim

An episode with a finalised code set and a DRG appears in the claiming queue, ready for Medicare, DVA or private health insurance submission, with no re-keying between the coding screen and the claim.

Coded episodes appear in the Ready to Claim queue
DRG grouping drives the funding calculation
Remittances reconcile against the original claim

Narrated walkthroughs

The short version

All walkthroughs

1m 52sClinical CodingManual navigation and AI autocoding, from the classification.

Acute & hospitalHospital intake to discharge on one record: one referral inbox, Medicare, DVA and IHI checks before admission, pre-admission from home, structured discharge and coding. Care AdminAdmissions, discharges, transfers, bed status, billing and census surfaced read-only from the systems you already run, with clinical coding and casemix from the same record. In the back officeFinance and claiming.

See clinical coding on a real episode

We’ll walk you through the Code Finder and the autocoder on real clinical data.