The sharing of vocabularies between different entites and communities in healthcare- that is, the use of standard terminologies- is a very powerful instrument within the clinical framework. It enables data sharing aimed at facilitating both continuity of care for patients and research studies.

This practice is beneficial, in that it allows a correct interaction between softwares and smooth aggregation/integration of data coming from sources belonging to different clinical and geographical contexts (as it is often the case within national and international research studies), limiting data elaboration to very specific tasks and bjectives, thus eliminating the risk of of information loss.

The most popular and widespread healthcare terminologies are typically created and curated by big no-profit companies or governmental agencies, and are focused on specific aspects of the clinical framework (diagnoses, drugs, lab exams etc). SNOMED, LOINC, ICD and ATC are egregious examples of such vocabularies. They represent the de-facto standards for representing concepts and events in the clinical framework.

It is not uncommon for local governmental agencies to exploit these shared terminologies as a foundation to create personalized classifications aimed at satisfying specific demands and needs by users.

In fact, localized versions of the common terminologies exist, where concepts appear in languages different from the original one- typically english. Sometimes the “basic” set of concepts is also extended by introducing new elements in the classification.

All of this happened when italian regional agencies needed to create lists of health services/procedures to be provided for free to patients with specific chronic conditions: most of the list items are concepts taken straight from ICD9-CM, but their descriptions are translated in italian. Also, new procedures are introduced, identified by codes complying with the ICD9 structure but not existing in the original classification []

A Vocabulary aggregators like UMLS o come i concetti OMOP [ e LINK A NOSTRA PAGINA] consentono di risolvere problematiche di interoperabilità e integrazione delle sorgenti operando un mappaggio in formato comune delle terminologie di riferimento (SNOMED, LOINC ecc).

Further information regarding the OMOP CDM can be found here […]

Below, a short overview of the most common standard terminologies.

common standard terminologies

SNOMED (Systematized Nomenclature of MEDicine) CT
Maintained and supported by the non-profit SNOMED International

The first iteration of SNOMED, created by the College of American Pathologists (CAP), is created in 1973 in the USA, as an extension of the Systematized Nomenclature of Pathology (SNOP), in use since 1965. It’s a multi-axial system (featuring 11 axes, way more than the 2- Toography and Morphology- featured in the SNOP) of hierarchical classification including terms regarding diagnoses, procedures, anatomical sites, organisms, drugs, specimens etc. [Roger A. Côté (1986). “Architecture of SNOMED”. Proceedings of the Annual Symposium on Computer Application in Medical Care: 74–80]

SNOMED CT, the only version that is currently supported and curated, is born from the union of the previous versions – RT, Reference Terminology, by the CAP, and CTV3, Clinical Terms v.3, by the National Health System (NHS). It underwent a radical change in the classification strategy: the multiaxial approach was abandoned in favor of a more complex, relationship-based hierarchical representation.

Despite non-CT versions of SNOMED being officially deprecated and discontinued, it is not uncommon to find those codes inside hospital DBs, especially for older data and/or very specific contexts, like registries of pathology [].

Currently SNOMED CT includes more than 350.000 terms.

LOINC (Logical Observation Identifiers Names and Codes)
Maintained and supported by the non-profit Regenstrief Institute

It was created in 1994 in response to a specific need for standardization of terminology regarding lab analysis. While the use of a standard (HL7) for inter-laboratory communications was already established, the same couldn’t be said about test identifiers, which were typically based on internal classifications belonging to the sender clinical center.

In time, this vocabulary expanded, getting to currently include also terms related to diagnoses, procedures, clinical and anthropometric parameters.

The latest release from August 2021 (LOINC 2.71) includes 96635 codes, belonging to both clinical and lab contexts.


ICD (International Classification of Diseases)
Maintained and supported by the World Health Organization (WHO)

A very early version of this classification can be said to have appeared back in the end of the XIX Century, when french statistician Jacques Bertillon created his ” Classification of Causes of Death “, presented in 1893 and later adopted worldwide.

Starting from its sixth revision the terminology changed its name to “International Statistical Classification of Diseases, Injuries and Causes of Death (ICD)”. Since 1948 the task of preparing and publishing periodic revisions of ICD is assigned to WHO.[]

In its current configuration, ICD is a system for diagnostic classification created to organize diseases and injuries in groups based on specific criteria. It includes terms regarding signs and symptoms, as well as accidents and causes of death.

ICD11, to be officially released in January 2022, includes 17.000 categories, 80.000 concepts and 130.000 terms.


ATC (Anatomical Therapeutic Chemical Classification)
Maintained and supported by the World Health Organization (WHO)

During a WHO Symposium in 1969 the need emerged for a shared system of drugs classification. In Norway the Drug Utilization Research Group (DURG) created the ATC classification as a revision and extension of the European Pharmaceutical Market Research Association (EphMRA).

In 1981, the ATC/DDD (ATC/Daily Drug Dose) was officially recommended by WHO as international standard for drug studies. In 1982 the WHO Collaborating Centre for Drug Statistics Methodology was created and assigned the task of coordinating development and use of the ATC/DDD.

This vocabulary is structured in 5 levels that classify drugs based on their anatomic, therapeutic/farmacological and chemical group, focusing on their active ingredients.


Maintained and supported by the United States National Library of Medicine (NLM).

In 2001, in response to emerging issues with non-correctly identified synonyms inside the UMLS system, the NLM started a project called RxNorm, with the aim of creating a new drug classification capable of facilitating interoperability of drug-related data. []

It’s a US-based terminology that classifies all the drugs available on the US market. It should be stressed that, compared to ATC, RxNorm focuses on commercial forms of pharmaceutical products: for each element, information about doses, modes and ways of administration, as well as box contents, commercial names and producer company are specified. For this reason it is selected as a standard within the OMOP CDM.


MedDRA (Medical Dictionary for Regulatory Activities)
Maintained and supported by MedDRA MSSO (Maintenance and Support Services Organization)

In the late ‘90s, the International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use developed MedDRA, a standard medical terminology organized in 5 hierarchical granularity levels with the aim of facilitating information exchange in the context of international regulations regarding medical devices.

Products covered by the scope of MedDRA include pharmaceuticals, biologics, vaccines and drug-device combination products.[]



AUTHOR: Anna Alloni