Are you involved in health research with linguistically diverse populations? Do you rely on back-translation to validate your research instruments? A groundbreaking article by Ozolins, Hale, Cheng, Hyatt, and Schofield (2020) provides a comprehensive critique of the widely adopted, yet often poorly implemented and documented, back-translation methodology in medical research. Published in the Expert Review of Pharmacoeconomics & Outcomes Research, this paper is essential reading for researchers seeking to enhance the methodological robustness and cultural appropriateness of their studies.
The Problematic “Gold Standard”
While often hailed as a “gold standard” in medical research, particularly for surveying non-English-speaking populations, back-translation is frequently adopted uncritically. The authors reveal how many studies treat translation as a mere technical or outsourced activity, an afterthought rather than an integral part of research design. This approach often overlooks the complex, arduous, and time-consuming nature of adapting questionnaires for different cultural and linguistic groups, leading to significant methodological failings.
Key issues identified include:
- Lack of Documentation: Many studies fail to adequately document the intricate translation process, obscuring how translation difficulties are resolved.
- Contextual Blind Spots: Back-translators are typically given limited context or briefing, leading to interpretations that may diverge from the original research intent.
- Challenges with Colloquialisms and Idioms: Direct translation of culturally specific phrases (e.g., “butterflies in the stomach” or “feeling blue”) can result in comical or misleading responses, requiring careful cultural adaptation.
- Focus on Literal over Pragmatic Equivalence: There’s an overemphasis on semantic equivalence rather than achieving pragmatic equivalence, which aims for the translated text to have the same effect on the target reader as the original would have on the source reader.
- Ambiguities in Source Texts: The translation process itself can unexpectedly uncover inherent ambiguities or unwieldy formulations within the original, validated research instruments, as demonstrated by issues with the Supportive Care Needs Survey (SCNS-SF-R).
Insights from the INFORM Study
The article draws on findings from the Australian INFORM project, a study on migrant cancer patients, where translation experts were integrated into the research team. This hands-on experience provided critical insights:
- STAI-S Example: The translation of “I feel upset” from the State-Trait Anxiety Inventory (STAI-S) produced varied back-translations like “agitated,” “worried,” “insecure,” and “disturbed” across four languages. This highlighted how the back-translator’s lack of briefing led to a choice (“insecure” as a long-term condition) that contradicted the research team’s intent to measure acute anxiety.
- Impact on Forward Translators: The awareness of an impending back-translation can inadvertently lead forward translators to strive for more literal versions, potentially sacrificing the most suitable and authentic phrasing in the target language.
Towards More Robust Translation Protocols
Ozolins et al. argue for a paradigm shift in how medical research approaches translation. Their expert opinion and analysis lead to crucial recommendations for improved protocols:
- Integrate Translation Experts: Translation experts and professional translators must be part of the research team from the very beginning of project design and throughout all stages.
- Explicit Documentation: Research protocols and reports need to explicitly detail translation processes, making it transparent how translation issues are discussed and resolved.
- Briefing for Forward Translators: Provide comprehensive briefings to forward translators on the purpose and intent of questionnaire items, while acknowledging the inherent limitations of briefing back-translators to maintain objectivity.
- Prioritize Pragmatic Equivalence: The ultimate goal of back-translation discussions should be to ensure the chosen terminology in the target language achieves optimal pragmatic equivalence to the source text, not merely literal correspondence.
- Reconsider Source Text Wording: Embrace the translation process as an opportunity to reflect on and potentially refine the wording and intent of original source language items.
- Consider Alternatives: Explore alternative methodologies, such as a translation-plus-revision model, where translators and revisers are fully briefed and collaborate closely with the research team.
This paper serves as a vital call to action, urging medical researchers to move beyond superficial adherence to back-translation and engage deeply with translation theory and practice. By integrating expert linguistic knowledge and fostering transparency, health research can truly ensure equity of health outcomes across diverse multilingual and multicultural populations.
Reference:
Ozolins, U., Hale, S., Cheng, X., Hyatt, A., & Schofield, P. (2020). Translation and back-translation methodology in health research – a critique. Expert Review of Pharmacoeconomics & Outcomes Research. https://doi.org/10.1080/14737167.2020.1734453
