PackagedProductDefinition.description - cardinality and language

[rlindstrm]

Current specification allows only one description without specifying the language. For multilingual countries this might be a problem, as the standard translations might not be clear/fair enough.

EMA implementation uses extension to indicate the language.

				<description value="Üks vaginaalpehmekapsel on blistris, mis koosneb läbipaistvast kolmekordsest PVC/PVdC/PVC laminaatkilest ja on suletud alumiiniumfooliumist kattega. Blister ja polüpropüleenist aplikaator on pakendatud kartongkarpi. ">
					<extension url="https://ema.europa.eu/fhir/extension/language">
						<valueCoding>
							<system value="https://spor.ema.europa.eu/v1/lists/100000072057"/>
							<code value="100000072172"/>
							<display value="Estonian"/>
						</valueCoding>
					</extension>
				</description>

In MedicinalProductDefinition.name, FHIR spec presents a different approach with country and name clearly indicated.

Should be discussed whether the current solution satisfies users or if a more consistent approach should be considered.

[gcangioli]

Any resource may define the language in the metadata.
Any human readable element may use the translation extension to convey additional translations see https://build.fhir.org/languages.html#ext

Is this not enough ?

[rlindstrm]

I don't know. Noticed by EE and SE, and we don't really need it anyway, because we only have one official language.
We just noticed that the product name has a special language attribute while package description does not, so EMA has created their own extension for it.

I guess for multilingual countries it might be a bit of a political issue to choose which one of your languages is the 'real' one and which are given as translations. But you're right, technically it's not much of a problem.

My next question is, is the EMA implementation (see xml example above) good? Using a self-defined extension to indicate the language of a text block? (EMA IG, ch 4.2.1)

[gcangioli]

My next question is, is the EMA implementation (see xml example above) good? Using a self-defined extension to indicate the language of a text block? (EMA IG, ch 4.2.1)

I don't know .. we should ask them...I suppose there is a reason if they did it

the only thing I can see is that translation uses a code for the language, maybe they need to have a CodeableConcept.. but I'm guessing...

[costateixeira]

https://jira.hl7.org/browse/FHIR-39128