When translating Clinical Outcome Assessments for different languages, having the correct level of formality and politeness for the audience in question is crucial for obtaining accurate and meaningful data. Expectations regarding the use of formal and polite language differs from culture to culture. This is particularly so in a medical and research context, where inappropriate uses of ‘you’, or honorifics, for example, can lead to discomfort, offense, misinterpretation or disengagement on behalf of participants in the clinical study.

Over the years, I have noted numerous misalignments or inconsistencies in formality levels in COA translations, which we then had to test as they were. This is particularly common in eCOA translations where standardised text, such as ‘please go back and edit’ or ‘save to submit your entry’, is translated with the formal ‘you’ in multiple European languages, even though the questionnaires themselves were intended for children. This typically occurs when the ‘formal you’ text was originally translated for adult-focused COAs, stored in a translation memory and then reused to an inappropriate audience without proper care or adaptation.

This misalignment raises two key questions:

1.       How do the inconsistencies in ‘you’ formality escape the translation process, especially in such a labour-intensive activity as the Linguistic Validation localisation steps?

2.       What repercussions are they likely to have on the usability of the measures and the eventual data quality?

Formality Conventions as Basic Audience Design

Audience Design is a concept in sociolinguistics, introduced by Allan Bell (1984). It proposes that variation in language is shaped by the speaker’s, writer’s, translator’s perceptions of their audience. It means thinking about who they are communicating to and how they should shape this communication.

In clinical practice, especially COAs, language should be intentionally tailored to the patient and who they are: children, adults, older patients, people with cognitive impairments, people from diverse cultural backgrounds, and so on. If the language is mismatched, then that is going to put the patients off, causing them to disengage, or to misunderstand questions or to feel uncomfortable. This can lead to compromised data quality.

Inconsistent formality levels in COAs constitute a failure in Audience Design, even when the needs of the audience have been very carefully considered in the development of the COA. The translation process can damage all of this, where the translated text does not consistently reflect the expected relationship between the questionnaire and the participant. Examples of this can include:

·       A child-friendly COA that uses informal ‘you’ also contains standardised screen messages that remain formal, intended for an older audience. This would cause an unnatural disconnect.

·       Conversely, a COA intended for an older audience with formal phrasing that contains occasional informal language can come across as unprofessional and disrespectful.

Therefore, COAs need to align its forms of address with audience expectations. The COA, in its development and in translation, must ensure that all elements of the questionnaire, including instructions, main items, response choices and system prompts, are consistent in tone and formality. The translation process can undo this.

How do these inconsistencies escape the translation process?

·       Pre-translated system prompts are being reused during the translation processes of multiple assessments. Typically, this is where screen prompts were originally translated for adult-focused COAs. When used in a questionnaire intended for children, whilst keeping the formal language intended for adults, this results in Audience Design failure.

·       There could be a lack of context in the translation process. This is when COA translations occur in isolated segments that are pulled from a Translation Memory system, rather than the full questionnaire context. The Translation Memory tool may automatically apply pre-approved translated text to new COAs, even when the target audience is different. Also, the forward translator is working on isolated segments. Here, the translator uses appropriate and expected formality conventions, whereas the pre-translated content does not.

·       Nobody is pulling the pre-translated and newly translated content together, leading to inconsistent reviews of the target version.

How does this affect usability and data quality?

Inconsistencies in formalities in COAs – Audience Design failure – can influence the usability of the tool and the accuracy of the data.

It can lead to confusion and disengagement. For children, the effect might be that formal messages are confusing or alienating. For older participants, they may find informal content unprofessional, unclear and offensive. This can reduce patient engagement, as the credibility of the tool is put into question.

Of course, it can lead to increased costs and delays in fixing the inconsistencies. This may require multiple additional rounds of retranslation and validation.

Solutions

To prevent inconsistencies in forms of address, the principles of Audience Design could be applied to localisation in the following ways:

·       Do not translate the COA content in discrete segments.

·       Provide content to translators by signposting, e.g. in concept elaboration and/or translatability assessments who the audience is from the outset and what the formality expectations should be.

·       Update Translation Memory Systems to store both formal and informal translations of screen instructions and prompts and select them based on the audience context.

Inconsistencies in formality in COAs are a failure in Audience Design. This can lead to confusion, disengagement and compromised data quality. Language style and use in COA content must be aligned with audience expectations. This alignment is vital in bringing about an improvement in how well clinical research works amongst diverse populations.

 Thank you for reading,

Mark Gibson

Leeds, United Kingdom, March 2025