• Project title: Reactive management of MULTILATERAL semantic NORMalization of Clinical Terminology with SNOMED CT
  • Acronym: NORMA CT
  • File Number: IMIDCA/2023/4
  • Project Location: Paterna, Valencia (Spain)
  • Start date: 01/04/2023 – End date: 30/06/2024


SNOMED CT Semantic Normalization

SNOMED CT (Systematized Nomenclature of Medicine – Clinical Terms) is the most extensive, precise and important comprehensive, multilingual and codified clinical terminology developed in the world. SNOMED CT is also a terminology product that can be used to encode, retrieve, communicate and analyze clinical data allowing healthcare professionals to represent information appropriately, accurately and unambiguously. The terminology is constituted, in a basic way, by concepts, descriptions and relationships. These elements are intended to accurately represent clinical information and knowledge in the healthcare field.

SNOMED CT is widely used by professionals around the world and has become a de facto standard. It is used to consistently record, retrieve, exchange and analyze clinical data. It is used worldwide and is updated monthly. Its use worldwide demonstrates the possibilities and benefits of SNOMED CT to improve semantic interoperability and the effective exchange of clinical information.

SNOMED CT offers professionals a stable, quality-assured multilingual clinical technology with a consistent release cycle that allows clinical data to be entered once, but used many times, preserving, better than other Clinical Terminologies, the SEMANTIC VALUE, if applied correctly from the first link in the Clinical Data chain.

Disadvantages of SNOMED CT

However, SNOMED CT is very difficult to implement in the first link of the Clinical Data chain, especially “in Patient Consultation” (EHR). Which negatively impacts the normalization at source of the majority of Clinical Data and the risk of loss of semantic meaning in the rest of the value chain.

The main reason given by clinicians in this phase is that the learning curve of the “proactive management” of SNOMED CT is very high and long for the little benefit they obtain, if the return of added value reaches them again.

General objectives of the project

 NORMA CT will leverage SNOMED CT as a complete clinical terminology, highlighting TWO GOALS:

  • On the one hand, NORMA CT focuses exclusively on the early registration of appropriate CONCEPTS (ideas) and their RELATIONSHIPS between them, facilitating the obtainment of all the benefits of SNOMED CT in the first link of the Clinical Data value chain.
  • On the other hand, it pursues a MULTILATERAL or POLIVALENT presentation of the SNOMED CT concepts that guarantees exact comparability in semantic terms between multiple linguistic versions at an INTERNATIONAL level that mitigates the risk of iatrogenic errors.

Specific technical objectives of the project

In NORMA CT we will focus NOT on the morphological-syntactic translation, but on the semantic translation that, in our case, tries to capture the SINGLE UNIVERSAL CONCEPT of SNOMED CT, beyond the FSN itself (Fully Specified Name) and the descriptions and synonyms that surround you.

The different forms, qualities and difficulties of obtaining this information are one of the greatest challenges in AI training and its predictive power, posing great difficulties in the implementation of this type of technologies in the health sector, both at the Primary Care and at the Specialized Care level.

NORMA CT aims to influence, through this research and experimental prototyping project, one of the main problems of the sector in the field of the use of Artificial Intelligence to support the SEMANTIC NORMALIZATION of the immense clinical expressiveness available at an INTERNATIONAL level.

This project will allow obtaining a complete standardized identification prototype of CLINICAL DATA, expressed by clinical professionals in both structured and unstructured text (in free text), through a novel distributed system based on classification services and semantic translation of CONCEPTS SNOMED CT and its RELATIONSHIPS in the cloud.


NORMA CT TECHNOLOGICAL INNOVATION pivots around the “hybrid” system in the cloud based on the following order:

  1. semantic TRANSLATION services (if applicable)
  2. NER (Name Entity Recognition) CLASSIFICATION services
  3. RE (Relation Extraction SDP Shortest Dependency Path).

The proposed system is based on Deep Learning NLP-Transformers and the main challenge to be overcome by NORMA CT is the LOSS OF INFORMATION between the links of the value chain while it is transmitted, (both going and returning) and the beginning of the capture in a late phase that does not “connect” with the original context and meaning of the Clinical Data.

The innovation of NORMA CT proposing the reactive standardization of SNOMED CT will facilitate the positive achievement of this challenge or, at least, mitigate the currently existing adverse effects.