Clinical Meaning Integrity · From term to code
Terminology Mapping Review
Before your coded data is reused, make sure the codes still mean what was recorded.
An independent written assessment of how local terms and codes map to SNOMED CT or ICD, and whether each map preserves the clinical meaning of its source.
A mapping can be technically valid and still be clinically wrong for its intended use.
The review is not a coding audit. It is a mapping-by-mapping assessment of what survives, and what is lost, when a recorded term becomes a code.
A map is only right for a purpose. A map that serves billing can be wrong for clinical reuse, and a map that serves analytics can be wrong for decision support. Every review starts from the purpose you state, and judges each map against it.
Meaning is lost most often in predictable places: negation, family history, uncertainty, temporality such as a past condition read as a current one, and qualifiers that change a term's meaning, in any language. Each of these failures still produces a valid code. Technical validation confirms that a code exists and is well formed. It does not establish that the target still means what the source meant.
Who it is for
Health-AI companies that turn clinical text or local codes into SNOMED CT or ICD before their models use them. Digital health platforms and healthcare organisations preparing to exchange or reuse coded data can commission the same review.
What you send
- The mapping table: source term or local code, target code, and any map attributes you hold.
- The purpose the map serves.
- The terminology versions in use.
No patient data is needed or accepted. The review works on the map, not on records.
How the sample is drawn
Up to 100 mappings are reviewed. They are drawn in four parts: every source entry carrying context, such as negation, family history, uncertainty, temporality or a qualifier; the most frequent entries; every entry with no mapping, more than one, or other ambiguity apparent in the table; and a random draw from the rest. The rule and the final composition of the sample are disclosed in the report.
The review is an assessment of the reviewed sample. It is not a certification of the complete mapping set.
The standard review covers one stated purpose and one terminology. Larger sets, more than one terminology or purpose, or extensive post-coordination are scoped and priced separately, in writing, before work begins.
What you receive
A written report with a numbered map register. For each reviewed mapping: the source entry and the meaning it carries, the target and the meaning it carries, the classification, and the reason for it, including what changed where the mapping is not equivalent.
The report closes with the patterns behind the gaps: where meaning is lost most often, which kinds of source entries are most affected, and whether the losses concentrate in particular mapping patterns.
The four classifications
- Equivalent: the target carries the meaning of the source for the stated purpose.
- Broader: the target says less than the source. Detail is lost.
- Narrower: the target says more than the source. Detail is added that was never recorded.
- Meaning lost: the target does not carry the source's meaning. A negation, a context, an uncertainty or a qualifier did not survive.
Fee and turnaround
A fixed fee, agreed in writing and paid in full before the review begins. The fee and terms are sent on request.
Delivered within seven business days of receiving complete materials and payment. The work is asynchronous; no meetings are required.
What it is not
The review examines the map. It is not:
- a remediation plan: it shows where meaning is lost, not how to rebuild the map;
- a certification of the map, of coding compliance, or of billing accuracy;
- a clinical safety assessment of any product that uses the map;
- legal or regulatory advice.
Use of the review
The review is prepared independently, for the commissioning client.
- The reviewer's name may not be used to endorse, promote or validate the map or any product that uses it.
- The review may not be presented as certification or validation.
- Unless agreed in writing, the report may not be published or distributed, and its findings may not be quoted in marketing.
See a sample
A full sample review of a simulated outpatient problem list, in English and colloquial Arabic, is published in Writing. Read the sample review →
How to begin
Request the fee and terms. The reply explains what to send and how the review begins.
Terminology Mapping Review · Up to 100 mappings · Seven business days · Written report · No patient data · No meetings required