v2+ Details - Local Development build (v0.3.2). See the Directory of published versions

Problem with possible Terminology solutions

​This page is intended to provide explanations for terminology solutions that seem not adequate. It will also give answers for a why or why not.

FHIR ValueSet Resource

​The ValueSet Resource in FHIR is an aggregation of all items that are somehow related to using a value set in its different flavors. In whether and in which way it can be profiled is explained on a terminology2.html. However, by that means a binding can only be done to a value set - without taking care of the semantic meaning, i.e. definition vs. expansion.

FHIR Value Set Instances

In FHIR value sets are used “as is”. They must be treated as a stand-alone information item.

​Providing some terminology.html on using value sets introduces some good reasons to weigh between a theoretical and pragmatic use. Also, in which way value set definitions are used (for example) with some parameters to help with and simplify maintenance. As such it can be expected that the value set resource is used quite frequently.

Basic Model: Direct Association of Codesystems

​Beside the foundational details as described in terminology the basic elements can be reduced to:

• data element that should be coded
• concept domain that describes the domain the information should come from
• the value set that seecifies the codes that should be used
• the codesystem(s) providing the codes

These elements are now arranged to fulfill the basic requirements to associate codes with a data element:

The primary question is: what mechanisms shall be provided or are necessary?

The green lines document the relationships that seem to be out of question. The assocation from a data element to the concept domain is to define which codes should be used. The same can be done for value sets. Furthermore, a codesystem also points to a set of concept domains that are applicable.

Option 1: direct specification

A simple solution could be to directly associate a code system to a data element (blue line). That could work if the complete codesystem is to be used. In case it sohuld be reduced to a subset, a value set becomes reasonable.

Hence, the question arises, what the benefit is in comparison to the next option?

Option 2: indirect specification

This way, each association of a codesystem, if any, is done via a value set. That eliminates the blue assocation (line).

How to profile?

A direct association of a codesystem to a data element (blue line) cannot be profiled. Then a value set must be introduced.