Medical Terminologies MCP

Look up the WHO ATC (Anatomical Therapeutic Chemical) classification(s) for a drug by name. Use this tool to: - Find the ATC code for a medication (e.g., "metformin" → A10BA02) - Identify the therapeutic and pharmacological class hierarchy - Cross-reference drugs with their international ATC codes Returns one entry per ATC code the drug belongs to. A single-ingredient drug typically maps to one substance-level code; combination products map to multiple. ATC codes are international (WHO Collaborating Centre); this tool retrieves them via NLM RxClass. Input is a drug NAME (brand or generic, English/US naming as in RxNorm). A name RxNorm does not recognize, or a drug without an ATC mapping, returns an empty `matches` list with an explanatory note — not an error; try the generic name, or resolve the name first with rxnorm_search. Which ATC tool: start here when you have a drug name; use atc_lookup when you already have a class code (level 1-4, e.g. "A10BA") and want its name; use atc_members to list the drugs inside a class.

Look up a specific CID-10 code and return its Portuguese name. Use this tool to: - Resolve a code to its Brazilian description ("I21" → "Infarto agudo do miocárdio") - Confirm a 3-char category or 4-char subcategory exists in CID-10 - Retrieve gender / cause-of-death restriction flags when applicable Accepts both dotted ("A00.1") and undotted ("A001") forms; returns the canonical display.

Map a LIST of free-text clinical terms to standard codes in one call, with ranked candidates and a confidence label for each — the building block of a reviewable crosswalk. Use this tool to: - Harmonize a column of diagnoses, drugs or lab names from a dataset to ICD-11 / RxNorm (+ ATC) / LOINC - Triage which terms map cleanly (exact / strong) and which need a person (needs_review) - Build a crosswalk you can audit: every row keeps its candidates, scores and sources Give each term its `domain`: diagnosis → ICD-11; drug → RxNorm concepts (ingredients first) plus the ATC classes of the term; lab → LOINC. Up to 50 terms per call — a longer list is refused with a validation error: split it into batches of 50. Repeated term+domain pairs are looked up once. `max_candidates` keeps 1-5 per term (default 3). Every candidate carries `match_score` (lexical, 0-1, the find_equivalent formula) and `match_type`: exact = same words after normalization; strong = every term word is in the title (or the matched synonym) and score ≥ 0.85; needs_review = anything else. A one-word term is exact or needs_review, never strong ("Tylenol" vs "Tylenol PM" is a different product). Synonyms, abbreviations ("MI", "HbA1c") and misspellings land in needs_review or no_candidates — the label errs toward asking a person. Candidates sharing no word with the term are dropped, as are LOINC codes named "Deprecated". For diagnoses, a candidate is also scored against the synonyms WHO matched (e.g. "hypertension NOS" for Essential hypertension), reported in `matched_label`; postcoordinated clusters (codes with "/" or "&") are left out — build those with icd11_postcoordination. Lab names are ambiguous without specimen and property: "glucose" matches over a thousand LOINC codes, so write "glucose serum" or expect needs_review. One failed lookup does not fail the batch: that row comes back with status "error". Terms are searched in English and sent to the WHO and NLM APIs — de-identify the list first. For Brazilian Portuguese diagnoses use cid10_search; to check codes you already have, use validate_codes; for one term across every terminology, use find_equivalent. Record the vocabulary versions with the provenance blocks (one per source) and terminology_versions.

Navigate the ICD-11 hierarchy to find parent or child entities. Use this tool to: - Find broader categories (parents) of a condition - Find specific subtypes (children) of a condition - Understand the classification structure Name the entity by `code` (a leaf code like "5A11", or a block range like "5A10-5A2Y" — blocks come back from 'parents' with an empty code and a code_range) or by `uri` (the URI any previous answer returned). Direction 'parents' returns ancestor categories, 'children' returns subcategories. ICD-10 codes (like "E11") are not ICD-11 codes: convert them first with map_icd10_to_icd11.

Get detailed information about ONE ICD-11 entity you already have a code or URI for. Use this tool to: - Get the full definition of a disease - Retrieve coding notes, inclusions and exclusions - Get the official title and index terms (synonyms) Provide `code` (e.g., "BA00") or `uri` (any URI a previous answer returned) — at least one is required; calling with neither returns a validation error naming both. Set `language` for WHO's official translations (e.g. `language: "pt"` for official Portuguese). Returns a single entity (no pagination). A code WHO does not know comes back as a "not found" error, never an empty record. When NOT to use: to find a code from a disease name, use icd11_search first; to walk parents/children, use icd11_hierarchy; for an ICD-10 code (like "E11"), convert it with map_icd10_to_icd11 — ICD-10 codes are not ICD-11 codes.

List the postcoordination axes WHO allows for one ICD-11 stem code (MMS linearization, pinned release). Postcoordination means attaching extra detail to a stem code — severity, laterality, anatomy, causing agent, etc. — to build a composite (cluster) code. Use this tool to: - See which axes a stem code accepts before building a composite code - Check which axes are REQUIRED vs optional - See whether an axis takes one or several values, and how many values it offers Provide an ICD-11 `code` (e.g., "BA00"). Returns one entry per axis with `axis_name`, `required`, `allow_multiple` and `value_count` — the count of allowed values, not the values themselves. A code with no postcoordination, or one WHO does not know, returns an empty `axes` list (not an error), so check the code with icd11_lookup if the list is unexpectedly empty. When NOT to use: this does not build or validate a composite code, and it does not list the allowed values; to get a code from a disease name, use icd11_search.

Authoritative ICD-10 → ICD-11 mapping using WHO transition tables (release 2025-01, bundled with the server). Returns the primary 1:1 ICD-11 category for the ICD-10 code plus any alternative ICD-11 candidates that WHO documents (some ICD-10 concepts split into multiple ICD-11 entities). For each mapping, includes the ICD-11 code, title, chapter, and the Foundation URI / Linearization URI for navigating to the full entity definition. Use this for clinical coding, billing migration, retrospective analysis, and any workflow that needs authoritative mapping rather than text-search candidates. Coverage: 11,243 ICD-10 categories (excludes chapters and blocks like "A00-A09" which aren't used in clinical coding). Provide a code like "E11" (Type 2 diabetes), "I21" (Acute MI), or "A07.8" (4 alternatives in WHO's table). Both dotted ("A07.8") and undotted ("A078") forms are accepted. Returns "no mapping" when the code isn't in the WHO category-level table — that's the honest answer rather than a fuzzy search fallback.

Map between RxNorm concepts and National Drug Codes (NDC). Use this tool to: - Get all NDC codes for a drug (by RxCUI) - Find the RxCUI for an NDC code - Cross-reference between coding systems Provide either an RxCUI to get NDCs, or an NDC to get the RxCUI.

Look up an ATC code at level 1-4 to get its name and hierarchy level. Use this tool to: - Resolve an ATC code (e.g., "A10BA") to its class name ("Biguanides") - Confirm a code exists in the current ATC index - Identify the level (anatomical / therapeutic / pharmacological / chemical) Accepts codes 1-5 characters long: "A" (anatomical), "A10" (therapeutic), "A10B" (pharmacological), "A10BA" (chemical). Substance-level codes (7 chars, e.g., "A10BA02") are not exposed by this endpoint — use atc_classify with the drug name to retrieve the substance code.

List the drugs (substances) that belong to an ATC class. Use this tool to: - Enumerate all members of a therapeutic class (e.g., "A10BA" → metformin, phenformin) - Build a list of drugs sharing a pharmacological mechanism - Explore an ATC subtree at any level Each member includes its substance-level (7-char) ATC code via source_atc_code, useful for disambiguation when the queried class is at level 1-4. RxNorm's catalog is US-centric; the ATC class names and codes themselves are international.

Get one CID-10 chapter and its constituent groups (e.g., "Chapter IX → I00-I02 Febre reumática aguda, I05-I09 Doenças reumáticas crônicas do coração, ..."). Use this tool to: - Drill from a chapter into its groups (code ranges with Portuguese titles) - Build hierarchical browsers - Find which group contains a code range Provide the chapter number `num` as an integer 1-22 (chapter I = 1, IX = 9). Numbers outside 1-22 are rejected with a validation error. Returns the chapter (title and code range) plus ALL its groups in one response — no pagination; the 22 chapters hold 275 groups in total. Answered locally from the bundled CID-10 V2008; no network call. Which CID-10 tool: use cid10_chapters first if you do not know the chapter number (it lists all 22 with code ranges); use cid10_lookup for one specific code and cid10_search to find codes by Portuguese text. For the international ICD-11, use icd11_chapters.

List the 22 chapters of CID-10 with their code ranges and Portuguese titles. Use this tool to: - See the top-level structure of CID-10 (chapters I-XXII, e.g., "I. Algumas doenças infecciosas e parasitárias", "IX. Doenças do aparelho circulatório") - Map a code to its chapter by code range (e.g., I00-I99 → chapter IX) - Build a navigable table of contents for downstream tooling Returns 22 entries — CID-10 V2008 has not been updated since 2008.

Search the Brazilian CID-10 (Classificação Estatística Internacional de Doenças, 10ª Revisão) by Portuguese text. Use this tool to: - Find CID-10 codes for Brazilian SUS / ANVISA contexts ("infarto", "diabetes", "tuberculose") - Look up the official Portuguese (CBCD/USP) translation of a clinical term - Locate codes for billing, epidemiology, and clinical documentation in Brazil Returns matches from CID-10 categories (3-char) and/or subcategories (4-char). Search is diacritic-insensitive: typing "infeccoes" matches "infecções". Every word must match (AND), and everyday Portuguese is resolved to the CID-10's own wording (câncer→neoplasia maligna, AVC→acidente vascular cerebral, pressão alta→hipertensão, suicídio→lesão autoprovocada, aids→HIV); when that happens the response says so in vocabulary_notes. This tool searches the Brazilian Portuguese CID-10 V2008 — for the international ICD-11 (current WHO revision, in English by default), use icd11_search.

Returns the full document for an id obtained from `search`, as { id, title, text, url, metadata }: `text` is the readable content (Markdown) and `url` the canonical public page to cite. Companion of `search` in the OpenAI Deep Research contract, over the medical terminologies (CID-10 categories and chapters, ICD-11, LOINC, RxNorm, MeSH, terminology version records) catalog. Only ids returned by `search` are valid; an unknown id returns an error. The terminology tools (`icd11_*`, `cid10_*`, `loinc_*`, `rxnorm_*`, `mesh_*`, `atc_*`, `map_*`, `find_equivalent`, `validate_codes`) remain the tools for data queries. Behavior: read-only and idempotent — a live GET against the public source when the document needs it.

Ranked unified search for equivalent terms across multiple medical terminologies. Use this tool to: - Find the same concept in different coding systems - Compare how terminologies represent a concept - Support terminology mapping and data integration Searches across: ICD-11, LOINC, RxNorm, and MeSH. Set `target_terminologies` to limit which are searched, or set `source_terminology` to exclude one (e.g. when you already have a code from that terminology and want equivalents elsewhere). The two combine: source is subtracted from targets. `limit` caps candidates per terminology (default 5, max 10). Every candidate carries `match_score` (lexical similarity to the search term, 0-1) and `rank` (global position across all searched terminologies) — both computed by this server, since upstreams don't expose comparable relevance scores. Candidates from different terminologies whose titles are lexically identical are clustered in `groups` — a strong same-concept signal (absence of a group is NOT evidence of non-equivalence). Searches upstreams in English. For official pt-BR content, use the dedicated tools: `icd11_search`/`mesh_search` accept `language: "pt"`, and `cid10_search` is natively Portuguese.

List all ICD-11 chapters (top-level categories) of the pinned WHO release. Use this tool to: - Get an overview of ICD-11 structure - Find which chapter covers a body system or condition type - Get chapter URIs to drill down with icd11_hierarchy (direction 'children') Returns 28 entries in one response, no pagination — chapters 01-26 plus the supplementary sections V (functioning) and X (extension codes) — each with number, code, title and URI. Each chapter is fetched separately from WHO; if one fetch fails, that entry keeps its URI and carries an `error` instead of a title, and the rest still come back. Set `language` for WHO's official translations (e.g. `language: "pt"`); the result is cached, so repeated calls are cheap. When NOT to use: to find a specific disease, use icd11_search; for the Brazilian CID-10 (ICD-10) chapters, use cid10_chapters.

Search for medical conditions, diseases, and health problems in ICD-11 (International Classification of Diseases, 11th Revision). Use this tool to: - Find ICD-11 codes for diagnoses - Search for diseases by name or keyword - Look up conditions in multiple languages Set `language` for WHO's official translations — e.g. `language: "pt"` searches and returns the official Portuguese (pt-BR) ICD-11 labels. Never machine-translated. Returns matching entities with codes, titles, and relevance scores.

Get the list of valid answers for a LOINC questionnaire item. Use this tool to: - Find valid response options for survey questions - Get answer codes (LA…) for data entry validation - Get the item scores of scored instruments (e.g. PHQ-9: "Not at all" = 0 … "Nearly every day" = 3) Returns each answer with its LOINC answer code (LA…), text, display order and score (null when the list has none). A valid LOINC code without a defined answer list (e.g. a numeric lab such as 2339-0) returns an empty list; a code LOINC does not know returns a "not found" error. Only applicable to LOINC codes that represent questions with defined answer sets. For the questions of a whole questionnaire, use loinc_panels; to find a code by name, use loinc_search.

Get detailed information about a specific LOINC code. Use this tool to: - Get the full name and description of a LOINC code - Find the component, property, timing, and system - Check the scale type and method Provide a LOINC number in format "XXXXX-X" (e.g., "2339-0" for Glucose).

Get the structure of a LOINC panel or form. Use this tool to: - See all tests included in a panel (e.g., CBC, metabolic panel) - Get the structure of assessment forms - Find related observations grouped together Returns the list of LOINC codes that make up the panel.

Search for laboratory tests, clinical observations, and measurements in LOINC (Logical Observation Identifiers Names and Codes). Use this tool to: - Find LOINC codes for lab tests (e.g., "glucose", "hemoglobin") - Search for clinical measurements and vital signs - Look up diagnostic observations Returns matching LOINC codes with names, components, and properties.

Get detailed information about a MeSH descriptor by ID. Use this tool to: - Get the full definition (scope note) of a MeSH term - View tree numbers showing hierarchy location - See related concepts and synonyms Provide a MeSH Descriptor ID like "D015242" (Ofloxacin). Set `language` to request NLM's official translations where they exist (e.g. `language: "pt"`).

Get allowed qualifiers (subheadings) for a MeSH descriptor. Use this tool to: - Find which qualifiers can be combined with a descriptor - Build precise MeSH search queries - Understand aspects that can be specified Qualifiers refine descriptors (e.g., "Diabetes Mellitus/drug therapy").

Search for MeSH (Medical Subject Headings) descriptors. Use this tool to: - Find MeSH terms for indexing medical literature - Look up subject headings for PubMed searches - Find controlled vocabulary terms Set `language` to request NLM's official translations where they exist (e.g. `language: "pt"` for Portuguese labels); content is never machine-translated. Returns matching descriptors with MeSH IDs and labels.

Get the tree hierarchy location(s) for a MeSH descriptor. Use this tool to: - See where a term fits in the MeSH hierarchy - Understand broader/narrower relationships - Find related terms in the same branch MeSH tree numbers show the hierarchical path (e.g., C14.280.647 for Myocardial Infarction).

Get therapeutic and pharmacologic classes for a drug. Use this tool to: - Find the drug class (e.g., "Beta-blockers", "NSAIDs") - Identify therapeutic categories - Look up mechanism of action classifications Returns class IDs, names, and classification sources.

Get detailed information about a specific RxNorm concept by RxCUI. Use this tool to: - Get the full name and synonyms for a drug - Check the concept status (active, remapped, etc.) - View related concepts (ingredients, brands, forms) Provide an RxCUI (RxNorm Concept Unique Identifier) like "161" — as a string of digits or as an integer.

Get active ingredients for a drug by RxCUI. Use this tool to: - Find the active ingredients in a medication - Check for single vs. multiple ingredient products - Identify the generic components of brand drugs Returns ingredient RxCUIs and names.

Search for drugs in RxNorm (Normalized names for clinical drugs). Use this tool to: - Find drug concepts by brand or generic name - Look up medications for prescribing - Search for drug formulations Returns matching drugs with RxCUI identifiers, names, and term types.

Searches the medical terminologies (CID-10 categories and chapters, ICD-11, LOINC, RxNorm, MeSH, terminology version records) catalog and returns up to 10 matching documents as { id, title, url }, ordered by relevance (an empty list means nothing matched). This tool exists for the OpenAI Deep Research contract: ChatGPT deep research, company knowledge and research workflows over the Responses API require exactly the tools `search` and `fetch`. Pass one of the returned ids to `fetch` to read the document. For direct questions and for data (values, series, rankings) prefer the terminology tools (`icd11_*`, `cid10_*`, `loinc_*`, `rxnorm_*`, `mesh_*`, `atc_*`, `map_*`, `find_equivalent`, `validate_codes`), which return the actual data with provenance — this is a catalog index, not a data query. Query: natural language or keywords, Portuguese or English; accents and case are ignored. Behavior: read-only and idempotent — the catalog comes from the public source and is cached in memory.

Report what diff data is available between two versions of a terminology. For most terminologies this is **guidance only** — the server doesn't ship historical snapshots, so the tool points at the publisher's official changelog and explains the cadence. `bundled_versions` lists the version(s) this server actually has on hand. For **ICD-10 vs ICD-11** specifically, the tool surfaces a real cross-revision summary from the bundled WHO transition tables (the ICD-10 → ICD-11 case is a structural diff between two WHO revisions). Use `terminology: "icd10"` with no `to_version` to get the cross-revision summary: total mapped ICD-10 categories, how many are 1:1 vs split into multiple ICD-11 codes, and the average number of alternatives when split. Inputs: - `terminology` (required): which terminology to report on. - `from_version` (optional): the version you have data from. If omitted, the tool reports against the currently-bundled version. - `to_version` (optional): the version you want to compare to. If omitted, the tool reports against the publisher's latest known release. This tool is intentionally a metadata + guidance layer, not a diff engine — for terminologies that change frequently (LOINC, RxNorm, MeSH), the publisher's official changelog is the authoritative source.

List the current version, release date, publisher, source URL, and update cadence of every terminology this server queries against. Useful for pipeline maintainers who need to: - Confirm which release of ICD-11 / LOINC / RxNorm / MeSH / ATC the server is querying before a batch run. - Verify the bundled CID-10 (frozen at V2008) and ICD-10 → ICD-11 transition tables (currently 2025-01) match expectations. - Cite the data version in research artifacts. Pass `terminology` to filter to a single entry; otherwise the full set of 7 is returned. The ICD-10 → ICD-11 version reads live from the bundled dataset; everything else is metadata maintained alongside the project release.

Validate a mixed batch of medical codes against their source terminologies. Useful for retrospective analysis of legacy databases — flag codes that no longer exist, surface ICD-10 → ICD-11 replacements, and grade activity status where the terminology exposes it. For each input `{ code, terminology }`, returns: - **valid**: whether the code exists in the source terminology. - **active**: whether the code is currently active. Null when the source doesn't expose an explicit active/inactive distinction at category level (CID-10, ATC, ICD-11, RxNorm, MeSH all return null today; LOINC returns a real boolean). - **title**: the official label/name when available. - **replaced_by**: a successor code, populated today only for ICD-10 codes that have a primary ICD-11 mapping in the bundled WHO transition tables. - **source**: human-readable provenance of the validation (terminology + release/version). - **error**: non-null only when validation couldn't be performed (network error, upstream outage, etc.). `valid: false` + `error: null` means "code not found"; `valid: false` + `error: set` means "couldn't validate". Terminology is **required per code** — auto-detection isn't supported because category codes like "A00" exist in both ICD-10 and CID-10. Accepted values: `icd11`, `icd10`, `loinc`, `rxnorm`, `mesh`, `atc`, `cid10`. Hard cap of 50 codes per call; codes are validated in parallel through their respective clients, so total wall time scales with the slowest upstream + its rate limit (worst case ~10 s for a full batch hitting ICD-11).