Your question is excellent and speaks to a very important issue and limitation of studies where outcomes utilize self reported (ie Patient Reports Outcomes - as the NIH and other now utilize this jargon).  The very "science of self report measures" continues to evolve.

Kate Lorig PhD Nursing affiliated with Stanford has validated instruments I believe they utilized in Spanish.  John Ware's original group for outcome self report measures had, as I recall, begun trying to obtain validated SF-36 among other in languages other than English.

Our recruitment and geographic focus, although racially and ethnically diverse, all were multilingual. Even if English was their secondary language, they had a proficiency and mastery of writing, reading, comprehension over the 6th grade level. IRB requires all documents to be written at that level even if English is primary and only language and we studied a group of nuclear physicists at Brookhaven National Labs.

Suffice it to say, that extrapolations and external validation of Patient Reported Outcomes or the "Science " of self reports is very limited in different cultures and languages and one should try to obtain a validated instrument for the population/sample of interest.

Hope this helps!