Codes vs scores for scales

The Glasgow Coma scale contains questions that are scored, and each score has a criterion by which to score it. E.g., the a patient scores 2 on the motor scale for “Extension to pain”; the 2 is added to the other scores for a total, which can be used for assessment & monitoring.



That component score can be conceived of as a concept, “patient exhibits extension to pain,” and given a concept code, “LA6563-6.” But the score is used in a calculation, so that magnitude is also needed. Simple implementations just use the score; more sophisticated ones might apply the code instead (as unique in the namespace), thereby requiring the receiver to look up the score. An appropriate data type would permit the sender to include both values, but a) currently prevailing standards, viz. CDA, use the R1 datatypes, with a deficient CO type, and b) even if it didn’t, the score isn’t really ordinal.



We are considering using the LOINC encoding of GCS questions, but typing the values as integers. This isn’t any less correct than the ordinal, and it supports score arithmetic.



Are we missing anything important?

We agree, typing the values as integers seems appropriate.



Jami Deckard, MS

LOINC Content Developer

Regenstrief Institute, Inc.