> ## Documentation Index
> Fetch the complete documentation index at: https://developer.eka.care/llms.txt
> Use this file to discover all available pages before exploring further.

# Medical Entity Codification

## Overview

**Medical Entity Codification** is Eka Care's named entity linking (NEL)
technology. It maps a free-text medical entity — a symptom, diagnosis, lab test
or drug — to the matching code in a standardised clinical ontology.

Free text is how clinicians naturally record information, but it is hard for
machines to aggregate, compare or exchange. Codification bridges that gap:
`"chest pain"` becomes SNOMED CT `29857009`, `"hemoglobin"` becomes LOINC
`718-7`, and `"type 2 diabetes mellitus"` becomes ICD-10-CM `E11`. The result is
structured, interoperable health data.

## Supported ontologies

| Ontology   | Domain                                    | Guide                                                         |
| ---------- | ----------------------------------------- | ------------------------------------------------------------- |
| SNOMED CT  | Symptoms, findings, disorders, procedures | [SNOMED CT](/eka-medai/technologies/codification/snomed)      |
| LOINC      | Lab tests and observations                | [LOINC](/eka-medai/technologies/codification/loinc)           |
| Medication | Branded and generic drugs                 | [Medication](/eka-medai/technologies/codification/medication) |
| ICD-10-CM  | Diagnoses from clinical text              | [ICD-10-CM](/eka-medai/technologies/codification/icd-10-cm)   |

## How the pipeline works

Each codification request runs through three stages:

1. **Query understanding** — the input text is parsed for the target ontology.
   For LOINC this extracts the test name, unit and specimen; for medication, the
   brand, salt, form and volume; for ICD-10-CM Comprehend, the clinical entity
   and its attributes.
2. **Retrieval** — candidate terms are fetched from the ontology index, using
   semantic vector search, a curated knowledge base, or AWS Comprehend Medical
   depending on the ontology.
3. **Linking** — candidates are scored and the best single match is flagged
   with `is_linked: true` when the pipeline is confident.

## The role of context

Some ontologies need more than a name to resolve correctly. A lab test like
`hemoglobin` is ambiguous until you know the unit and specimen — blood and urine
hemoglobin are different LOINC codes. The API accepts an optional `metadata`
field to carry this context, which materially improves accuracy for LOINC and
medication. The per-ontology guides explain exactly what to send.

## Use Cases

* **Structured EMRs** — code symptoms, diagnoses and prescriptions consistently.
* **Lab standardisation** — link extracted test names to LOINC for longitudinal tracking.
* **Interoperability** — produce the coded entities required for FHIR resources.
* **Analytics** — normalise clinical entities for reliable population-level insights.
* **Medical coding** — map free-text diagnoses to ICD-10-CM for claims workflows.

## Try it out

1. Explore the [API Reference](/api-reference/health-ai/medical-entity-codification/overview)
   to get started.
2. Call [List Registry](/api-reference/health-ai/medical-entity-codification/registry) to
   see the live ontologies and versions.
3. Link your first entity with [Link Entity](/api-reference/health-ai/medical-entity-codification/link-entity).

Ready to structure your clinical data? [Get in Touch](mailto:ekaconnect@eka.care) today.