SNOMED codes are standardized numeric identifiers assigned to clinical concepts, from diseases and symptoms to procedures and body structures, so that computers and health systems can share medical information in a consistent, precise way. The full name is SNOMED Clinical Terms, usually abbreviated SNOMED CT. It contains hundreds of thousands of concepts and is maintained by SNOMED International (formally called the International Health Terminology Standards Organisation). In the United States, it is one of the required standards for electronic exchange of clinical health information across federal government systems.
What Problem SNOMED CT Solves
A doctor in one hospital might document “heart attack,” while another records “myocardial infarction,” and a third writes “MI.” All three mean the same thing, but a computer searching patient records wouldn’t automatically know that. SNOMED CT solves this by assigning a single unique numeric code to each clinical concept, then linking every synonym for that concept back to the same code. This lets electronic health records, lab systems, and research databases treat those three entries as identical.
That consistency is called interoperability: the ability for different health IT systems to exchange data and actually understand it. Without a shared terminology, transferring a patient’s record from one hospital to another can introduce errors or gaps. SNOMED CT fills that gap by providing a common language that machines can process and humans can read.
How a SNOMED Code Is Structured
Each SNOMED code, called an SCTID, is a numeric identifier with a built-in structure. It contains an item identifier (the unique number for that particular concept), a partition identifier (which tells you whether the code refers to a concept, a description, or a relationship), and a check digit at the end that catches data entry errors. Some codes also include a namespace identifier that indicates which organization created the concept. The codes are purely numeric, with no letters or special characters.
Three Core Building Blocks
SNOMED CT is built from three types of components that work together:
- Concepts represent individual clinical ideas, from “abscess” to “zygote.” Each concept gets a unique numeric identifier and sits within a hierarchy that moves from general to specific. This structure lets you record a very detailed diagnosis and later pull reports at a broader level.
- Descriptions are the human-readable terms linked to each concept. A single concept can have multiple descriptions, each one a synonym. The concept for a sore throat, for example, might have descriptions like “pain in throat,” “pharyngeal pain,” and “sore throat” all pointing to the same code.
- Relationships connect concepts to each other. They define things like “pneumonia IS A type of lung disease” or “appendectomy has a procedure site of appendix.” These logical links are what allow computers to reason about medical data, powering features like clinical decision support alerts.
How Concepts Are Organized
All SNOMED CT concepts branch out from a single root into major categories called top-level hierarchies. Each hierarchy groups similar types of concepts together. Some of the most commonly used ones include:
- Clinical finding: results of clinical observations, assessments, or judgments, including diagnoses like asthma, headache, or normal breath sounds.
- Procedure: activities performed in healthcare, from appendectomies and imaging scans to physiotherapy sessions and medication administration.
- Body structure: anatomical locations and structures referenced by other concepts.
- Pharmaceutical/biologic product: drug products, such as “amoxicillin 250mg oral capsule.”
- Specimen: samples collected for examination, like a urine specimen or a needle biopsy from the prostate.
- Observable entity: questions or assessments that can produce a result, such as blood pressure or body weight.
- Staging and scales: assessment tools and tumor staging systems like the Glasgow Coma Scale.
There are additional hierarchies covering things like occupations, qualifier values (such as “left,” “right,” or “severe”), and situations with explicit context (like “past history of myocardial infarction” or “family history of glaucoma”). This breadth is what makes SNOMED CT useful well beyond diagnosis coding.
SNOMED CT vs. ICD Codes
People often wonder how SNOMED codes differ from the ICD codes they see on medical bills. The two systems serve fundamentally different purposes and are designed to complement each other, not compete.
ICD (the International Classification of Diseases) is a classification system. It groups diseases into standardized categories using alphanumeric codes, primarily for billing, public health statistics, and mortality tracking. It is optimized for counting and comparing: how many people in a region were diagnosed with diabetes last year, for instance.
SNOMED CT is a clinical terminology. It captures the fine-grained detail of what’s actually happening with a patient, supporting documentation, clinical decision-making, and computer-based analysis. Where ICD might have one code for “type 2 diabetes,” SNOMED CT can represent that same condition along with its severity, associated complications, and the specific body sites involved, all as coded, computable data.
The two systems are formally linked. A collaboration between SNOMED International and the World Health Organization has produced cross-maps connecting approximately 19,000 SNOMED CT concepts to ICD-10 codes. In practice, this means a clinician can document in rich SNOMED detail at the point of care, and the system can automatically generate the appropriate ICD code for billing and reporting without anyone re-entering information.
Where SNOMED Codes Are Used
SNOMED CT shows up in electronic health records, lab information systems, clinical decision support tools, and public health reporting. In the U.S., the Departments of Health and Human Services, Defense, and Veterans Affairs have all selected it for use in laboratory results, diagnoses, procedures, anatomy, and nursing documentation. The FDA adopted a SNOMED CT “Problem List” subset for electronic labeling of prescription drug products, improving how drug information is exchanged in FDA-approved package inserts.
Because SNOMED CT maps to other medical terminologies already in use, organizations can avoid entering the same data twice. A hospital can code clinical data once in SNOMED CT and then derive billing codes, public health reports, and quality measures from that single entry.
How It Works With Health Data Standards
SNOMED codes plug into broader health data exchange standards like HL7 FHIR, the framework increasingly used for sharing health records between systems. In FHIR, SNOMED CT is identified by the URI http://snomed.info/sct. When a system sends or receives clinical data, it references SNOMED concept IDs within that namespace. The preferred display term for each concept is determined by the relevant language or dialect, so the same code can surface the correct term whether a system operates in English, Spanish, or another supported language.
FHIR also supports SNOMED CT expressions, which combine multiple concepts using a structured grammar to represent complex clinical ideas that don’t have a single pre-existing code. This composability is one of the things that sets SNOMED apart from simpler code systems.
Updates and Releases
The SNOMED CT International Edition is released monthly, on the first of each month. Individual countries may also publish their own national editions on different schedules, adding concepts specific to local healthcare needs. This frequent update cycle means new diseases, procedures, and clinical concepts can be incorporated relatively quickly.
Licensing and Access
SNOMED CT is free to use in member countries, which include the United States, the United Kingdom, Australia, Canada, and dozens of others. You do need a license, but there is no fee. In the U.S., the National Library of Medicine manages distribution. Non-member countries can also obtain access, though fees may apply, and organizations must register through the SNOMED International Member Licensing and Distribution Service before using the terminology. Low-income countries and qualifying research projects in any country also get free access.