Healthc Inform Res Search

CLOSE


Healthc Inform Res > Volume 29(4); 2023 > Article
Park: Review of the 2023 Korean Society of Medical Informatics Summer Camp on SNOMED CT
The Korean Society of Medical Informatics (KOSMI) Summer Camp on SNOMED CT took place in early August 2023 in Busan. The camp covered a range of topics, including the importance of standard terminology, an introduction to SNOMED CT, the implementation of SNOMED CT in Electronic Health Records (EHRs), the customization of SNOMED CT, and the retrieval of SNOMED CT content.
Healthcare professionals record patient information in the EHR using either coded data or free text. Terminology standards can be used to capture coded data at the point of care, or to convert free text data into a structured format at the point of use. Structured data can then be utilized for patient care, clinical research, or public health.
SNOMED CT is the world’s most comprehensive, multilingual clinical healthcare terminology, encompassing a wide array of concepts including clinical findings, procedures, observable entities, pharmaceutical/biological products, organisms, social contexts, family history, and past medical history. It facilitates a consistent, processable representation of clinical concepts, thereby enabling the capture, structuring, and indexing of patient data in EHR. SNOMED CT is also mapped to other international standards such as the International Statistical Classification of Diseases and Related. Health Problems (ICD), Logical Observation Identifiers Names and Codes (LOINC), Medical Dictionary for Regulatory Activities Terminology (MedDRA), and International Classification of Nursing Practice (ICNP). Currently, it is in use in over 80 countries.
SNOMED CT can be implemented in an EHR as a code system for storing clinical information, an interface terminology for capturing and displaying clinical information, an indexing system for efficient retrieval of clinical information, a common terminology for meaningful communication and integration of heterogeneous data, a dictionary for querying, analyzing, and reporting, and as a link between health records and knowledge resources. It can also serve as an extensible foundation for representing new types of clinical concept. SNOMED CT comprises more than 360,000 concepts, each representing a unique clinical meaning, 1,500,000 human-readable terms known as descriptions linked to concepts, and 3,000,000 relationships that represent connections between concepts. These relationships are used to define concepts within SNOMED CT. For instance, 31978002 |Fracture of tibia| is defined as:
414292006|Fracture of lower leg|:363698007|Finding site|=12611008|Bone structure of tibia|
A structured combination of one or more concept identifiers used to represent a clinical meaning is called an expression. There are two types of expressions: pre-cooperated and post-cooperated. A pre-cooperated expression is one that includes a single SNOMED CT concept, while a post-cooperated expression includes more than two SNOMED CT concepts. Both types of expressions must be represented using compositional grammar in accordance with the concept model. This model outlines which attributes (or relationship types) and values can be applied to each type of concept. Post-coordinated expressions are particularly useful for creating concepts that are not already included in SNOMED CT. For instance, a “spiral fracture of the tibia” can be post-coordinated as:
31978002|Fracture of tibia (disorder)|:116676008|Associated morphology (attribute)|=73737008|Fracture, spiral (morphologic abnormality)|
SNOMED CT can be customized to improve its usability. Customization use cases encompass extensions, which involve adding terms used in a local language or incorporating components absent in the international edition. Subsets are another customization option, allowing for the addition of terms used by a specific user group or the inclusion of concepts relevant only to a local context. Mapping is another customization strategy, which involves integrating local codes with SNOMED CT and merging statistical classification systems with SNOMED CT. Lastly, the terminology can be configured to create specific groups of concepts for data entry, reporting, or analytics, and to link components to clinical knowledge resources. Examples of these customizations include the Korean extension of SNOMED CT with new concepts or new descriptions in Korean, a cardiovascular disease reference set for a disease registry, an ICNP and SNOMED CT map reference set, and an anesthesia type reference set.
A reference set is a standard format used for maintaining and distributing a set of references to SNOMED CT components. This enhances the usability of SNOMED CT by providing a standardized method for its customization. Reference sets can be published as part of the SNOMED CT content. The members of a reference set can be viewed using the “Member” function of the SNOMED CT browser or the ECL operator ^. For instance, the members of the International Patient Summary reference can be viewed by searching for the concept and listing the members using the “Member” function or the Expression Constraint Language (ECL).
^816080008|International Patient Summary|
Creating a reference set involves four steps. First, select the candidate data items. Ensure that these items have a consistent level of granularity, no duplication of meaning, and no missing data items when you select them. Second, map the data items to SNOMED CT concepts. If there is no equivalent SNOMED CT concept, consider dividing the meaning across multiple data elements, using a post-coordinated expression, creating a new extension concept, or using a more general concept. Third, create a reference set by selecting SNOMED CT concepts that are semantically equivalent to the data items. Fourth, publish the reference set as a derivative of SNOMED CT.
SNOMED CT content can be retrieved using an expression constraint. An expression constraint is a computable rule employed to define a set of clinical meanings, represented by either pre-coordinated or post-coordinated expressions. Expression constraints have various use cases, including terminology binding, which provides links between the information model and the terminology; intensional reference set definitions, which support the customization of SNOMED CT content; SNOMED CT content queries, which facilitate search using hierarchies and formal concept definitions; and the SNOMED CT concept model, which constrains the domain and range of each attribute. Expression constraints can be formulated by determining the domain of the focus concept, the set of valid attributes for the given domain, and the valid range for each attribute. The ECL employs operators such as “descendant of,” “descendant or self of,” “child of,” “child or self of,” “ancestor of,” “ancestor or self of,” “parent of,” “parent or self of,” “member of,” “any,” “conjunction,” “disjunction,” “exclusion,” and “reverse flag” to represent expression constraints. The use of ECL was demonstrated with a research question: “Should we prescribe a tumor necrosis factor (TNF) inhibitor to a patient with autoimmune arthritis and a chronic lung disease with an outcome of pneumonia?” To answer this research question, four subsets (autoimmune arthritis, chronic lung disease, TNF inhibitor, and bacterial pneumonia) and three cohorts (autoimmune arthritis + chronic lung disease, TNF inhibitor vs. no TNF inhibitor, and % bacterial pneumonia in TNF and no TNF inhibitor groups) were created. For instance, the autoimmune arthritis subset was created as follows using ECL:
<<64572001|Disease|:      <<363698007|Finding site|      =<<39352004|Joint structure|,<<116676008|Associated morphology|      =<<707496003|Inflammation and consolidation|,<<370135005|Pathological process|      =<<769247005|Abnormal immune process|
With the widespread adoption of EHRs, healthcare organizations have recognized some of the challenges associated with aggregating and sharing clinical data in a machine-readable format. For this reason, many national initiatives are underway to support the meaningful sharing of health data. Standard terminologies within health information technology systems is a fundamental component of these projects. SNOMED CT serves as one of the solutions. Healthcare professionals play a critical role in the adoption and advancement of SNOMED CT in EHRs. A training and education session on SNOMED CT, such as the KOSMI Summer Camp, will help healthcare professionals develop their competencies in SNOMED CT, enabling them to adopt and use standard terminology in their practice.

Notes

Conflict of Interest

No potential conflict of interest relevant to this article was reported.



ABOUT
ARTICLE CATEGORY

Browse all articles >

BROWSE ARTICLES
FOR CONTRIBUTORS
Editorial Office
1618 Kyungheegung Achim Bldg 3, 34, Sajik-ro 8-gil, Jongno-gu, Seoul 03174, Korea
Tel: +82-2-733-7637, +82-2-734-7637    E-mail: hir@kosmi.org                

Copyright © 2024 by Korean Society of Medical Informatics.

Developed in M2community

Close layer
prev next