-
Notifications
You must be signed in to change notification settings - Fork 0
Concepts
The vocabulary this client speaks, and why each idea matters. Read this once and the method names in Guides and API-Reference read themselves.
A single clinical idea: a drug, a lab test, a diagnosis. Each concept has a stable numeric conceptId that stays the same no matter which coding system originally named it.
Why it matters: concepts are the anchor everything else hangs off. Once you hold a conceptId, every other namespace (interactions, mappings, reference ranges) takes it as input.
A source coding system such as SNOMED CT, RxNorm, or LOINC. The same idea often has a different code in each.
Why it matters: HOLON normalizes across vocabularies so you work with one concept instead of many codes. When you already hold a raw code from one system, concepts.getByCode(code, vocabulary) resolves it to the concept.
What kind of thing a concept is: a Drug, a Measurement, a Condition.
Why it matters: domains let you narrow a search to the category you mean. A search for a term that names both a drug and a condition returns fewer, more relevant hits once you pass a domain.
A mapping links the same idea across two vocabularies. Translation is the act of moving a code from one vocabulary into its equivalent in another.
Why it matters: translating a RxNorm code into its SNOMED equivalent is how you move a patient's data between systems that speak different codes. See the mappings namespace in Guides.
The normal range for a lab test, narrowed by age and sex, so a value can be read as high, low, or in range for that specific patient.
Why it matters: a raw lab number means little without context. The same value can be normal for one demographic and abnormal for another. Reference ranges give you the boundary to compare against.
A known, clinically significant effect between two drugs.
Why it matters: the interactions API screens a pair or a whole medication list, so you catch a risky combination before it reaches a patient.
A score for how alike two sets of phenotype terms are.
Why it matters: it is how you compare patients, or match a presentation to a candidate condition, when exact codes will not line up. You get a similarity signal where a code-for-code match would return nothing.