Original Article
Cross-cultural adaptation of the Health Education Impact Questionnaire: experimental study showed expert committee, not back-translation, added value

https://doi.org/10.1016/j.jclinepi.2013.07.013Get rights and content

Abstract

Objectives

To assess the contribution of back-translation and expert committee to the content and psychometric properties of a translated multidimensional questionnaire.

Study Design and Setting

Recommendations for questionnaire translation include back-translation and expert committee, but their contribution to measurement properties is unknown. Four English to French translations of the Health Education Impact Questionnaire were generated with and without committee or back-translation. Face validity, acceptability, and structural properties were compared after random assignment to people with rheumatoid arthritis (N = 1,168), chronic renal failure (N = 2,368), and diabetes (N = 538). For face validity, 15 bilingual people compared translations quality with the original. Psychometric properties were examined using confirmatory factor analysis (metric and scalar invariance) and item response theory.

Results

Qualitatively, there were five types of translation errors: style, intensity, frequency/time frame, breadth, and meaning. Bilingual assessors ranked best the translations with committee (P = 0.0026). All translations had good structural properties (root mean square error of approximation <0.05; comparative fit index [CFI], ≥0.899; and Tucker–Lewis index, ≥0.889). Full measurement invariance was observed between translations (ΔCFI ≤ 0.01) with metric invariance between translations and original (lowest ΔCFI = 0.022 between fully constrained models and models with free intercepts). Item characteristic curve analyses revealed no significant differences.

Conclusion

This is the first experimental evidence that back-translation has moderate impact, whereas expert committee helps to ensure accurate content.

Introduction

What is new?

Key findings

  1. A multidisciplinary committee is necessary to ensure accurate content in the cross-cultural adaptation of a questionnaire.

What this adds to what was known?
  1. No specific experimental quantitative and qualitative evidence existed that demonstrated the relative value of a committee and back-translation in the cross-cultural adaptation, particularly using a sample sufficiently powered to detect small differences across methods.

  2. We confirmed that a committee contributes substantive improvements to the cross-cultural adaptation process, whereas we could not see clear evidence that back-translations specifically improved translation quality.

What is the implication and what should change now?
  1. Researchers and practitioners seeking to translate questionnaires should ensure that a well-informed bilingual translation committee is recruited. A back-translation can be omitted in this situation, thus reducing costs and time taken to translate questionnaires.

  2. Back-translation should no longer be mandatory but remains useful when input is required from the questionnaire developer unfamiliar with the target language or the absence of a bilingual committee.

The evaluation of patient-centered care interventions relies on tools that measure patient-reported outcomes (PROs) [1]. A PRO is designed, and its validity documented [2], for specific settings and the language in which it was developed. To be used outside its original setting, it must undergo translation and cultural adaptation [3]. There are several advantages in adapting an existing questionnaire: it is quicker, cheaper, and easier to undertake adaptation rather than creating a new one. A robust questionnaire available in several languages permits international studies and comparison across languages [3], [4].

The aim of adaptation is to generate a version of the questionnaire that returns data that are equivalent to the original. The process should avoid the introduction of bias and inaccuracies. For example, a scale could be translated in different ways: Good, fair, and poor could be translated in French to bon, moyen, and mauvais or, alternatively, bon, médiocre, and mauvais because fair has two possible translations with different intensity and would elicit different answers [5]. Challenges of translation include faux-amis (false friends) which are prone to mistranslations and a word-to-word translation of an expression can lead to a nonsense [4].

A recent review [6] identified 17 translation and cultural adaptation processes. For example, the World Health Organization or Guillemin and Beaton [3], [4], [7], [8] recommend forward translation, back-translation, and expert committees. The back-translation is the original language generated from the target language [3], [7], [9].

The purpose of the back-translation is to highlight discrepancies between the source document and the translation. It permits a greater involvement of the original author [4], [10]. The back-translation step is generally considered best practice and produces a satisfactory result despite being time consuming and expensive [11], [12]. The back-translation may, however, introduce errors and bias or correct them in situations where the back-translator identifies a forward translator error. There is currently no rigorous evidence of the value of the back-translation in questionnaire adaptation, leading to suggestions that it could be omitted [4], [5], [8], [9], [10], [13], [14].

Guillemin and Beaton advocate that a multidisciplinary committee should discuss the back-translation, synthesizing the previous phases of translation and back-translation and correcting errors. This step is considered central for cultural adaptation and producing a consensual forward translation [3], [7]. However, a multidisciplinary committee may not change the outcome, suggesting that the procedure could be further simplified [13], [14].

Few studies have compared cultural adaptation approaches [6], [9], [13], [14], and no robust experiment demonstrates the value of back-translations or committees. We therefore translated and culturally adapted a well-validated English questionnaire, the Health Education Impact Questionnaire (heiQ) [15], [16], used widely in the evaluation of chronic disease self-management and health education programs [16], a rapidly growing field.

Using an experimental procedure, we tested the relative contribution of the back-translation and the multidisciplinary committee to the content and psychometric validity of the translation of a multidimensional questionnaire.

Section snippets

Health Education Impact Questionnaire

The heiQ (see Appendix at www.jclinepi.com) was developed to measure the impact of a wide range of health education programs [15], [16], [17], [18]. It has 40 items across eight dimensions. The Likert response options (1–4) include strongly disagree, disagree, agree, and strongly agree and are summated within each domain to obtain a scale score. The items were constructed with minimal ambiguity and colloquialisms, confirmed with people from low socioeconomic backgrounds, and has been adapted to

Item and scale content

Per protocol, two forward translators directly produced the Initial translation. Sixteen changes were made to the BT-only translation, 36 changes to the Committee-only translation, and 25 to the Both translation. There were two main types of changes to the Initial: improvements in style and clarity (27 items) and changes to the meaning of the item to get closer to the original (29 items). For example, the stylistic change in item 22 “If I need help, I have plenty of people I can rely on” was in

Discussion

This study showed that the translations with multidisciplinary committees had better face validity and more accurate content. The translations that included a committee had better fit indices across CFA models; however, the differences were small and not likely to substantively change conclusions that might be drawn from analysis of clinical or public health data.

One of the primary roles of the multidisciplinary committee is to ensure that the questionnaire content stays true to the original.

Conclusion

Using a large sample of typical patients, within a randomized design applying advanced psychometric procedures, this study provides clear guidance that although the inclusion or exclusion of back-translation and consensus committees have a minor effect on psychometric structure and the ensuing quantitative accuracy, the quality (content validity) of the adapted tool is better with the inclusion of a committee.

Acknowledgments

The authors acknowledge S. Gendarme; E. Speyer and the team in Nancy, France; A.C. Rat and committee members; S. Briançon and the Quavirein team; S. Chirol and ANDAR association; O. Ziegler and Diabetology Department of Nancy; the bilingual people; and the team in IWH, Toronto, Canada.

References (26)

  • A. Leplège et al.

    The adaptation of health status measures: methodological aspects of the translation procedure

  • D.E. Beaton et al.

    Guidelines for the process of cross-cultural adaptation of self-report measures

    Spine

    (2000)
  • F. Guillemin

    Cross-cultural adaptation and validation of health status measures

    Scand J Rheumatol

    (1995)
  • Cited by (0)

    Conflict of interest: All the authors declared no conflict of interest.

    Funding: Inserm CIC-EC CIE6, Société Française de Rhumatologie, and the Région Lorraine, France.

    View full text