2024-10-23 - Translation User Group Meeting Business Meeting Seoul on-site and online

Frank Geier : The linguistic issues are the most important from the framework, such as finding the preferred term, false friends, etc. Regarding conceptual and ontological challnges, these are more like quality assurance (QA) issues. Regarding cultural challanges: Is SNOMED CT an international terminology? Residuals from an Anglocentric past? Should there be an agreement on what is really international?

STEFAN SCHULZ Endorses the framework. Important to give feedback to Snomed Int. when the meaning of concepts is ambigous. In this sense, the translator doesn´t have a traditional role, but more of QA. The lack of text definitions is a weakness, not only for translators, but for SNOMED CT as a product.

Hynek Kruzik  Poor support of tooling makes translation difficult, both initial translation and maintainance. EHDS will requiere translations in many European languages, at least of relevant value sets.

Frank Geier Support the tooling issue. It´s a huge issue. Would like to ask Snomed Int. if there is a new tool.

Katharine Priest No knowledge of a new translation tool. Will try to find out.

Ole Våge Snomed Int. has made a presentation about a AI translation tool. But what we need is a tool which covers the whole process of translation, not just one of the steps. But the issue of tooling should be raised in Member Forum.

Marie-Alexandra Lambot This brings back the issue of translation templates, which should be picked up again. These exits for English when making concepts. These also help consistency and AI. And also of interface terms and forbidden terms.

Elze de Groot The Netherlands have tried the new AI tool.

Examples:

Emma Hultén Wound and ulcer: one of the lingusitic challenges. The languages don´t relate to eachother 1:1. The translation into Swedish would render "sår" in both cases. So the problem is not unclear meaning of the concept. The etiology is not included in the word. One strategy would be to use the preferred term as a kind of FSN which states the etiology. However, this would depart from the rule that the preferred term should be the term used by clinicians, so it´s not a good solutoin. Now, there are duplicates in Swedish translations because of this problem. So it´s an issue of knowledge of SNOMED CT and picking the right concept for a value set, for example. Just picking a word like "sår" when creating a value set is problematic, because you don´t distinguish between the underlying difference between traumatic and non-traumatic. 

Mikko Härkönen How to translate "condition"? In Finnish, there are at least two different terms.

Marie-Alexandra Lambot The problem of insufficiency and failure, which goes back to ICD-9 and ICD-10 with renal insufficiency and renal failure. In French, we don´t really express the difference even though we understand the difference. Previously, I´ve suggested to forbid one of the concepts.

Camilla Wiberg Danielsen The tooling should be helpful in cases as these, by adding more context in different languages. We can´t change the languages, and the concepts are valid concepts.

Hynek Kruzik This highlightes the missing text definitions. In Czech, there are different approaches, such as adding the English term.

Irena Rubešová A different challenge are concepts such as 'gastrontestinal tract'. In Czech, 'liver' is a part of it, in English not.

Elisabeth Serrot There are ways of measuring the difference between renal insufficiency and failure, and this information would be useful to add to the concept.

Ole Våge At least four different approaches: 1 Avoid the concept if possible 2 Add text definitions 3 English term in brackets 4 Term represent meaning (FSN-like)

STEFAN SCHULZ Another approach: linking to an authoritative source

Emma Hultén The question of renal insufficiency and failure is quite different from ulcer and wound. The latter are very clear.

@Sophie Klopfenstein Support the idea of linking to sources. Sometimes ignoring one of the concepts would be a solution.

Frank Geier We have translators, we have domain experts, but sometimes the challenges in translated are related to very specific concepts which clinicians might not know.

We are terminologists and we lack clinical knowledge. Where do we go then? Do we need an international forum for areas with these kind of difficulties?

Hynek Kruzik There are European associations for the different specialities. Not necessary to create new structures.

STEFAN SCHULZ Or use students. Or Medline or PubMed, where there are bilingual abstracts and healines.

Frank Geier Maybe we need a joint effort of finding good sources. Also, as translators, we take a skeptical approach towards the concepts we are translating. This makes the process more time consuming, but is still necessary.

Ole Våge Accessing domain experts is a great topic for future discussing and experience sharing.

But right now, it would be nice to know what kind of challenge do you struggle most with? Challenges of meaning of concepts/equivelance or challenges of finding the right term, choosing the preferred term, identifying synonyms, i.e. linguistic challenges?

Emma Hultén Finding the right term.

David Op de Beeck The synonyms, the number of synonyms and which synonym to choose as the preferred term. It will be influenced by the use cases and user groups: Who do you translate for and what for. Preferred terms for experts of a specialized part of the terminology or preferred terms which are more understandable? There is not one sole truth. Do we want to create search terms?

Frank Geier We also want to add synonyms, but the question is how many? Also the wish to be consistent.

STEFAN SCHULZ Also the risk of false friends. And the case when you translate something and then find out it´s wrong, like the case of "finger". In SNOMED CT there are four fingers and one thumb, in German there are five fingers. The challenge in SNOMED CT is that you have to go up and down the hierarchy to find out. This is not even clear in English, so this might SNOMED CT idiosyncracies.

Hynek Kruzik Is there a list of the problems which translators could access? And a collection of language guidelines would be helpful.

Ole Våge Should be add something to the general guidelines for translation

Frank Geier No, no more documents and pdfs. It would be useful to acccess common challanges and mistakes, but not in a document. Maybe a spreadsheet or power point.

Elze de Groot The Jira page of the Translation User Group would be helpful.

Camilla Wiberg Danielsen In Denmark, the experience is that the clinicians don´t use the more trivial words. They use "ulcus", they use "first finger", "second finger", etc.

Marie-Alexandra Lambot The number of the fingers were also added in the French translaiton. 

Camilla Wiberg Danielsen If we had clinicians from the beginning, several of the problems could have been resolved, like the distinction between digit and finger.

Regarding general words like "service" we had a look at several languages, like Swedish, Norwegian, German and Dutch and compare how words like "service" have been translated. In fact, they have been translated into many, many things.

Hynek Kruzik We had problems when translating "convulsion" with the synonym "seizure". When translating into Czech, "convulsion" would render a translation with a meaning different from "seizure"., also looking at the children of the concept. Is this a problem of the international edition?

Emma Hultén Regarding "service", it has a very broad meaning in English, so it is difficult to have one translation. It should be translated in diffrent ways, and tt depends on the context.

Ole Våge As a follow up: Collect these examples, put into categories and present them in a spread sheet or similar. And add strategies to solve them.

Joint meeting with Joint Advisory Group for item #5 (scroll down to find the Joint Advisor Group agenda) . See Joint Advisory Group for minutes.