Healthcare · Clinical text and coding

Everything your model needs is in the note.So is everything the law says must not leave it.

Clinical text structured against your schema, coded against the vocabulary you name, and de-identified under a method we name out loud, with its cost in clinical signal and its residual risk written down rather than waved through.

Scope the text work See the two lawful methods

Three questions decide whether a note can become training data.

Structuring clinical text is the easy half. The hard half is proving what left the building, under which method, and who was qualified to read it on the way out.

  1. Can it lawfully leave?

    Mr Berg

    The rule gives two methods and no third. Safe Harbor removes eighteen categories of identifier and needs no statistician. Expert Determination keeps more clinical signal and needs a qualified expert to determine the risk is very small and to document how. Your engagement names which one it uses before a single note is touched.

  2. Did someone who reads notes for a living read it?

    shortness of breath

    An entity is only a code once a person agrees it is. Extraction runs with native-speaker linguists, clinical judgement calls go to reviewers credentialed for the domain, and disagreement is measured on the delivered material rather than averaged away inside the process.

  3. Will it load into what you already run?

    chest pain

    The delivery schema is agreed at scoping, not discovered at handover. FHIR R4 is one of the shapes we export into, delivered as files against the profile your engagement names.

Four engagements, scoped to one corpus and one schema.

Each is independently purchasable and configured per engagement. Four things are settled before the first note is opened.

Agreed before the first note: the delivery schema , the vocabulary and its version , the de-identification method , and who adjudicates a disagreement .

  1. Clinical note structuring

    The part of the record no query can reach.

    Extraction, classification and structuring of clinical free text against an agreed schema: conditions, medications, dosages, procedures, findings, and the negations and uncertainty markers that change the meaning of all of them. Scanned material is handled layout-aware, so a table survives as a table.

  2. Coding against your vocabulary

    The code your systems already speak, at the version they speak it in.

    Entities linked to ICD-10 and ICD-11, SNOMED CT, RxNorm or MedDRA at the version your engagement names. Cases where two codes can both be defended go to an adjudicator rather than to a majority vote, and the disagreement is kept in the record.

  3. De-identification with the method named

    A record you can defend to a privacy officer who has read the rule.

    Direct and indirect identifiers removed under the method your engagement selects, with what that method costs you in clinical signal stated up front, and residual re-identification risk documented rather than declared absent.

  4. Evaluation sets from clinical text

    Test material your model has not already read.

    Held-out clinical sets for the tasks you actually ship: extraction, coding, summarisation and question answering, with inter-annotator agreement measured on the delivered material using a coefficient agreed at scoping.

Scope the first corpus

There are two lawful methods. There is no third.

Both are named in the same rule, and they trade the same thing against each other: how much clinical signal survives, against how much the method demands of you. The engagement picks one, in writing, before anything is touched. Here is one sentence under both.

Safe Harbor 45 CFR 164.514(b)(2)

Mr Berg, 72, seen 14 March at Gjøvik with chest pain since Tuesday.
Denies shortness of breath. Started aspirin 75 mg.

On this sentence 4 removed · 0 held
Expert Determination 45 CFR 164.514(b)(1)

Mr Berg, 72, seen 14 March at Gjøvik with chest pain since Tuesday.
Denies shortness of breath. Started aspirin 75 mg.

On this sentence 1 removed · 3 held
Synthetic specimen, not a record
Removed
Kept, under the expert's documented analysis

Safe Harbor

Remove the eighteen categories. Hold no actual knowledge.

Eighteen categories of identifier come out: names, geography below state level, all dates tied to an individual, contact details, record and account numbers, biometric identifiers, and the rest. The party de-identifying must also hold no actual knowledge that what remains could still identify someone. The age stays: only ages over eighty-nine come out.

What it gives you
Mechanical and auditable. No statistician required.
What it costs you
Takes real clinical signal with it. Every date, the place, the weekday.
What it leaves behind
The identifier inventory, and what was removed from which field.

Expert Determination

A qualified expert determines the risk is very small, and documents how.

A person with appropriate knowledge of generally accepted statistical and scientific principles determines that the risk of re-identification is very small, and documents the methods and results of that analysis. The rule sets no mathematical threshold. The judgement is the expert's, made against the context of the release.

What it gives you
Keeps clinical signal that Safe Harbor would strip out. The dates can stay.
What it costs you
Requires a qualified expert, and their analysis takes its own time.
What it leaves behind
The expert's documented methods and results, signed by the expert.

Four stations. The note only moves when a person says it can.

A note arrives as it was written, entities are marked in place, codes are linked at a pinned version, and the package is delivered in the shape agreed at scoping. Each station turns green only when a person has confirmed it, never when a model has finished.

  1. Note

    Received as it was written

    Free text, dictation output or a scan. Where the source is an image it is handled layout-aware, so headings, tables and columns survive as structure instead of collapsing into a wall of words.

    Source logged

  2. Entities

    Marked by someone who reads them

    Conditions, medications, dosages, procedures and findings marked in place, with negation and uncertainty marked alongside them, because a ruled-out diagnosis is not a diagnosis and a suspected one is not a confirmed one.

    Spans confirmed

  3. Codes

    Linked at a pinned version

    Each entity linked to a code in the vocabulary the engagement named, at the version it named. Where two codes can both be defended, the case goes to an adjudicator and the disagreement stays in the record rather than being resolved by a vote.

    Code adjudicated

  4. Package

    Delivered in the shape you asked for

    The structured records, the vocabulary version, the agreement figure, and the de-identification method with its stated limits, in the delivery schema agreed at scoping.

    Schema matched

Mr Berg, 72, seen 14 March at Gjøvik with chest pain since Tuesday. Denies shortness of breath. Started aspirin 75 mg.

schema
as agreed at scoping
vocabulary
version-pinned
spans
character offsets kept
de-identification
Safe Harbor · 4 removals · residual risk documented
agreement
coefficient named, measured on this set

Synthetic specimen, not a record

Synthetic specimen, not a record
  1. Note

    Mr Berg, 72, seen 14 March at Gjøvik with chest pain since Tuesday.
    Denies shortness of breath. Started aspirin 75 mg.

    Source logged
  2. Entities

    Mr Berg, 72, seen 14 March at Gjøvik with chest pain since Tuesday.
    Denies shortness of breath. Started aspirin 75 mg.

    Spans confirmed
  3. Codes

    Mr Berg, 72, seen 14 March at Gjøvik with chest pain since Tuesday.
    Denies shortness of breath. Started aspirin 75 mg.

    Code adjudicated
  4. Package

    Mr Berg, 72, seen 14 March at Gjøvik with chest pain since Tuesday.
    Denies shortness of breath. Started aspirin 75 mg.

    schema
    as agreed at scoping
    vocabulary
    version-pinned
    spans
    character offsets kept
    de-identification
    Safe Harbor · 4 removals · residual risk documented
    agreement
    coefficient named, measured on this set
    Schema matched

  1. Note

    Received as it was written

    Free text, dictation output or a scan. Where the source is an image it is handled layout-aware, so headings, tables and columns survive as structure instead of collapsing into a wall of words.

  2. Entities

    Marked by someone who reads them

    Conditions, medications, dosages, procedures and findings marked in place, with negation and uncertainty marked alongside them, because a ruled-out diagnosis is not a diagnosis and a suspected one is not a confirmed one.

  3. Codes

    Linked at a pinned version

    Each entity linked to a code in the vocabulary the engagement named, at the version it named. Where two codes can both be defended, the case goes to an adjudicator and the disagreement stays in the record rather than being resolved by a vote.

  4. Package

    Delivered in the shape you asked for

    The structured records, the vocabulary version, the agreement figure, and the de-identification method with its stated limits, in the delivery schema agreed at scoping.

The shape is agreed before the work, not after.

What arrives, in which form, and what that form actually commits us to. Nothing here is discovered at handover.

Structured records
Your agreed schema, or JSON On the specimen one condition, one drug with its dose, one finding kept as ruled out
The field names are the ones written into the engagement, not ours. A schema change mid-engagement is a scope change, and we say so.
Codes
ICD-10 · ICD-11 · SNOMED CT · RxNorm · MedDRA On the specimen R07.4 · 29857009 · RxNorm 1191
Named and version-pinned per engagement, because a code that moved between versions is a different code.
Clinical exchange
FHIR R4 resources On the specimen Condition · MedicationStatement
Delivered as files, against the profile your engagement names. Loading them into your systems stays inside your own change control.
Spans and provenance
Character offsets against the source On the specimen chest pain · characters 42 to 52
Every code can be traced back to the words that produced it, which is what makes a disputed label reviewable months later.
De-identification evidence
Method, removals, residual risk On the specimen Safe Harbor · 4 removals · residual risk documented
The record a privacy officer asks for: which method, what came out, and what risk remains.
Agreement
A named coefficient on the delivered set
Measured on this material, not asserted about the process. The coefficient is agreed at scoping so the number means the same thing to both sides.
hub Image 09 · the ribbon at rest, the single green LED (shared)

Clinical text is the most identifying material we handle. It stays where it belongs.

A Norwegian company under GDPR. Clinical project data is stored in Europe by default and processed in the EEA where required, with the paperwork to prove both. US healthcare work is handled HIPAA-compliant, under a signed Business Associate Agreement where protected health information is involved.

Jurisdiction
Norway · GDPR-native
Storage
European by default
Processing
EEA where required
Languages
Nordic and European clinical text · native-speaker linguists
Annotators
Named, contracted, never crowdsourced
Reviewers
Credential-verified per engagement
Evidence
Method, removals, spans, agreement
US engagements
HIPAA-compliant handling · BAA where PHI is involved
Contracts
Standard DPA terms · SCCs available
Erasure
30-day end-of-contract SLA

The healthcare hub Clinical data collection Clinical AI systems Clinical model evaluation Medical imaging annotation De-identifying clinical audio

Start with one corpus, one schema, one vocabulary.

We scope against the material as it actually looks, agree the schema and the de-identification method before anything is touched, and prove the method on a bounded first batch.

  1. Scope
  2. Schema
  3. Method
  4. First batch
  5. Acceptance

A bounded first batch carries its own acceptance criteria. You judge the structured output against your own records, not against a benchmark.

Briefs are treated as confidential. We are used to corpora that cannot leave the EEA and schemas that cannot be described outside the engagement.

hub Image 10 · the bench edge, the channel as a floor line (shared)

The brief Tell us what the material is, what schema it has to land in, and which method your privacy officer expects. We reply with a feasibility read.

What the engagement needs (optional)