Jack, I’ll see your 2 cents and raise you one penny more (remember when typewriter’s had a cent key?). My personal comments are not as important as us brainstorming on our own jargon, and settling on definitions we can work with and share outside the group for more clarity.
I am not a big fan of ‘code system’ as I (perhaps too emotionally) feel it leads us to forget that we are dealing with systems of concepts, and the relationships between concepts. We speak too easily of selecting or mapping codes when we are really doing things with concepts, along with their attributes and metadata, and the concept code, cui, is the handy currency we use to exchange and otherwise operate on concepts. People new to knowledge representation confuse concepts and their codes, and lose an appreciation of the former, because our language is not crisp.
Applying this concern to Jon’s good questions is difficult, and I think there may not be a universally accepted single answer to each question.
I am better with the ‘Authored’ bucket being call a Terminology System (and might use Ontology if complex relationships are incorporated) than Code System, and I feel Controlled Vocabulary is also accurate, but is more an operational or functional description of how the System operates. Unacceptable, but maybe accurate, would be Concept System.
The ‘Groups’ category does not specify that the members of the group are tightly related as all being valid answers to the same information question, so for me, Value Set is not a good label as I infer that role to Value Set. Collection works better as I feel it allows for a perhaps arbitrary gathering of concepts from one or more sources. I would suggest Value Set is a real good label for a Collection that comprises the valid answers to a particular question - maybe another bucket for Jon. When concepts are not specified as being answers, there may not be an a priori expectation that they are answers and not questions. For example, and this is admittedly a superficial example, Hemoglobin could be the answer to what lab tests were done, or Hemoglobin could be a question to which a numeric result is the answer.
When we speak generically, ‘terminology’ works fine for most situations I have been in. I would not upper case the ‘t’ when using it in this context.
On Wednesday, September 4, 2013 1:39:20 PM UTC-4Jack, paynejd wrote:
I have a question about the best terminology to use for terminology, which I suppose is meta-terminology, or maybe meta-semantics. Whatever we come up with should be added to the OpenHIE TS dictionary.
I’m looking for the best term for these 3 entities:
"bucket for authored concepts, terms, clinical quality measures, and indicators" - e.g. SNOMED CT, ICD-10, LOINC, ICPC-2, etc. but also IMO interface terminology, World Development Indicators, or WHO Indicator & Measurement Registry (IMR). (I have been using “source”)
"group of concepts from one or more other sources of authored content" - ordered or unordered list of references to concepts/terms/measures/indicators, possibly with additional annotation or other metadata (I have been using “collection”)
"word/phrase that generically refers to both of the above" - it is useful to be able to generically refer to any list of concepts, whether a comprehensive list of original content or a small subset (not sure what works here?)
"word/phrase that generically refers to terms, concepts, clinical quality measures, and indicators" - (does “concept” cover all of these?)
Here is an initial list of candidate words/phrases with my attempt at definitions for each:
- For authored content
- *source *- name for original, authored content, such as SNOMED, LOINC, etc., or IMO Interface Terminology, WHO Indicator Registry, etc.
codeset - ??
reference terminology - used as a reference, which often means original content (i.e. LOINC), but could also just mean the gold standard for a particular implementation (i.e. the Rwanda maternal-child health subset)
- *dictionary *- list of concepts plus definitions; can be authoritative or not
- *vocabulary *- same as dictionary??
registry - such as WHO Indicator Registry
- For lists of references to concepts
collection - ordered or unordered list of concepts from one or more sources, with or without a specific use case and other annotation
value set - unordered list of concepts from a single source with a specific use case
subset - unordered list of concepts from a single source
- For the individual phrases and their definitions
- *clinical quality measure *- count or summary of the data associated with a particular term
indicator - numerator plus denominator