The process and challenges of language translation and cultural adaptation of study instruments: a case study from the NIDA CTN CHOICES-2 trial

• Published in: Journal of ethnicity in substance abuse Vol. 22; no. 2; p. 417
• Main Authors: Collins, Julia A, Lewin, Shoham, Shmueli-Blumberg, Dikla, Hoffman, Kim A, Terashima, Javier Ponce, Korthuis, P Todd, Horigian, Viviana E
• Format: Journal Article
• Language: English
• Published: England 02.05.2023
• Subjects: Humans; Language; Surveys and Questionnaires; Translating; Translations; implementation of clinical trials; Language translation; substance use disorder and treatment
• ISSN: 1533-2659
• DOI: 10.1080/15332640.2021.1952129

As the U.S. grows more diverse, researchers decide how to include non-English speakers. Budget limitations may not allow for translation of all instruments. Study teams must determine which instruments must receive certified translations. This paper describes the procedures utilized in one U.S.-based, multi-site clinical trial to decide which study instruments should undergo certified translation and discusses dialect review procedures. The team determined which instruments (n = 31) would be translated using a qualitative evaluation to determine the need to obtain a Spanish-language certified translation: 1) "Could the meaning of these questions change (and potentially elicit a different response) if the translations were not consistent?" and 2) "Is it acceptable to have potential inconsistencies in these data?" Instruments for which question 1 was "yes" and question 2 was "no" (e.g., eligibility, outcomes, safety) were marked for certified translation. A dialect committee reviewed all translated patient-reported outcome measures to ensure that the translations had accounted for different meanings of words based on respondents' countries or regions of origin and recommended changes where necessary. Fourteen interview-based instruments underwent certified forward-only translation into U.S. Spanish. The remaining 2 interview-based instruments were translated via real-time conversation with participants by bilingual staff. Six forms were administrative and not translated. Five out of 9 professionally translated patient-reported outcome measures were amended to better reflect relevant dialects. In the absence of specific guidance, it is feasible for study team members to 1) determine which instruments should undergo certified translation and 2) incorporate dialect into translations.