SNOMED CT — Concepts, Relationships, and Expressions

What SNOMED CT Is

SNOMED CT (Systematized Nomenclature of Medicine — Clinical Terms) is the most comprehensive, multilingual clinical terminology in use today. Where a simple code list can only label a fixed set of things, SNOMED CT aims to represent the meaning of what clinicians record across the full breadth of care: clinical findings and disorders, procedures, body structures, organisms, substances, specimens, and more ¹. Its purpose is to capture clinical meaning at the point of care in a form that is precise for humans to read and stable for software to process.

That breadth is only useful because SNOMED CT is structured. It is not a flat list of terms but a network of formally defined concepts linked by relationships. This structure is what lets a record entered in one system be understood, aggregated, or reasoned over in another.

The Building Blocks

SNOMED CT rests on three kinds of components: concepts, descriptions, and relationships.

Concepts. Each concept is a single clinical idea with a unique numeric identifier called an SCTID. The identifier carries no meaning of its own; its meaning comes from how the concept sits in the terminology. Crucially, an SCTID is machine-stable: the number 73211009 denotes the idea of diabetes mellitus and will continue to, even as the words used to describe it evolve ². Software can therefore compare, count, and reason over concepts by identifier without parsing text.

Descriptions. Descriptions are the human-readable terms attached to a concept. Every concept has exactly one Fully Specified Name (FSN), an unambiguous label that names the concept and, in parentheses, its semantic tag — for example, “Diabetes mellitus (disorder)” or “Myocardial infarction (disorder)” ². Each concept also has a Preferred Term for everyday display (“Myocardial infarction”) and may carry any number of Synonyms (“heart attack”). The key idea is a one-to-many split between meaning and words: many human terms can point to one stable concept, so a clinician can search with familiar language while the record still resolves to a single unambiguous meaning ¹.

Concept (SCTID)  22298006
  FSN            Myocardial infarction (disorder)
  Preferred Term Myocardial infarction
  Synonym        Heart attack

Relationships. Relationships connect concepts to one another. The most important is the is a relationship, which builds a subtype hierarchy: a more specific concept is a kind of a more general one. Because a concept may have more than one parent — it can be a subtype of several broader ideas at once — the hierarchy is a polyhierarchy rather than a simple tree ¹. Diabetes mellitus, for instance, acts as a parent for more specific kinds of diabetes that sit beneath it via is a links ². Beyond is a, concepts carry defining attribute relationships that state their meaning formally — a disorder may be linked through a finding site to the body structure it affects and through an associated morphology to the kind of change involved. These defining relationships are what allow software to recognize, for example, that one disorder is logically a subtype of another.

Diabetes mellitus (disorder)            73211009
   ^ is a
Type 1 diabetes mellitus (more specific child concept)

Expressions and Post-Coordination

Sometimes the precise clinical idea a clinician needs is not available as a single ready-made concept. SNOMED CT distinguishes pre-coordinated concepts — a single SCTID that already packages the full meaning — from post-coordinated expressions, where several concepts are combined at the moment of recording to express something finer-grained ¹. An expression names a focus concept and may refine it with attribute-value pairs (for example, refining a finding with a severity or a laterality). A laboratory result such as “No growth (finding)” can likewise be combined with other concepts to state what was tested and how ². Post-coordination gives the terminology open-ended expressiveness without forcing every conceivable combination to be authored in advance.

What SNOMED CT Is For — and What It Is Not

SNOMED CT is built to record clinical meaning richly and consistently as care happens: what was found, diagnosed, done, or observed ¹. Its depth and polyhierarchical structure make it well suited to clinical documentation, decision support, and retrieval.

It is not designed for billing, reimbursement, or statistical reporting. Those jobs belong to classifications such as ICD, which deliberately group conditions into a smaller set of mutually exclusive categories tuned for counting and payment ¹. A useful mental model: SNOMED CT captures meaning at the point of care, while a classification like ICD summarizes that meaning for administrative and epidemiological use. The two are complementary, and a later lesson on ICD will examine how detailed SNOMED CT records are mapped onto those coarser categories. Keeping the distinction clear is essential — using a clinical terminology for statistics, or a billing classification for clinical detail, defeats the purpose of each.