SPINE Volume 25, Number 24, Pp 3186–3191 2000,

7m ago
48 Views
0 Downloads
738.66 KB
6 Pages
Transcription

SPINE Volume 25, Number 24, pp 3186 –3191 2000, Lippincott Williams & Wilkins, Inc.Guidelines for the Process of Cross-Cultural Adaptationof Self-Report MeasuresDorcas E. Beaton, BScOT, MSc, PhD,*†‡§ Claire Bombardier, MD, FRCP,*§ ¶#Francis Guillemin, MD, MSc,** and Marcos Bosi Ferraz, MD, MSc, PhD††With the increase in the number of multinational andmulticultural research projects, the need to adapt healthstatus measures for use in other than the source languagehas also grown rapidly.1,4,27 Most questionnaires weredeveloped in English-speaking countries,11 but evenwithin these countries, researchers must consider immigrant populations in studies of health, especially whentheir exclusion could lead to a systematic bias in studiesof health care utilization or quality of life.9,11The cross-cultural adaptation of a health status selfadministered questionnaire for use in a new country,culture, and/or language necessitates use of a uniquemethod, to reach equivalence between the originalsource and target versions of the questionnaire. It is nowrecognized that if measures are to be used across cultures, the items must not only be translated well linguistically, but also must be adapted culturally to maintainthe content validity of the instrument at a conceptuallevel across different cultures.6,11–13,15,24 Attention tothis level of detail allows increased confidence that theimpact of a disease or its treatment is described in asimilar manner in multinational trials or outcome evaluations. The term “cross-cultural adaptation” is used toencompass a process that looks at both language (translation) and cultural adaptation issues in the process ofpreparing a questionnaire for use in another setting.Cross-cultural adaptations should be considered forseveral different scenarios. In some cases, this is moreobvious than in others. Guillemin et al11 suggest fivedifferent examples of when attention should be paid tothis adaptation by comparing the target (where it is goingto be used) and source (where it was developed) languageand culture. The first scenario is that it is to be used in thesame language and culture in which it was developed. Noadaptation is necessary. The last scenario is the oppositeextreme, the application of a questionnaire in a differentculture, language and country—moving the Short Form36-item questionnaire from the United States (source) toFrom the *Institute for Work and Health; †St. Michael’s Hospital;Departments of ‡Occupational Therapy and Medicine and the§Clinical Epidemiology and Health Care Research Program, Universityof Toronto ¶The University Health Network, Toronto General Hospital; and #Mt. Sinai Hospital, Toronto, Ontario, Canada; **Ecole deSant, Publique, Facult, de M,decine,Vandoeuvre-les-Nancy, France;and the ††Division of Rheumatology, Department of Medicine, EscolaPaulista de Medicina, Universidade Federal de Sõo Paulo, SõoPaulo, Brazil.Supported in part by the American Academy of Orthopaedic Surgeonsand the Institute for Work & Health who co-sponsored the developmentof these guidelines. DB was supported by a Medical Research Council ofCanada PhD Fellowship during the preparation of this work.3186Japan (target)7 which would necessitate translation andcultural adaptation. The other scenarios are summarizedin Table 1 and reflect situations when some translationand/or adaptation is needed.The guidelines described in this document are basedon a review of cross-cultural adaptation in the medical,sociological, and psychological literature. This reviewled to the description of a thorough adaptation processdesigned to maximize the attainment of semantic, idiomatic, experiential, and conceptual equivalence betweenthe source and target questionnaires.13. Further experience in cross-cultural adaptation of generic and diseasespecific instruments and alternative strategies driven bydifferent research groups18 have led to some refinementsin methodology since the 1993 publication.11.These guidelines serve as a template for the translationand cultural adaptation process. The process involves theadaptation of individual items, the instructions forthe questionnaire, and the response options. The textin the next section outlines the methodology suggested(Stages I–V). The subsequent section (Stage VI) presentsa suggested appraisal process whereby an advisory committee or the developers review the process and determine whether this is an acceptable translation. Althoughsuch a committee or the developers may not be engagedin tracking translated versions of the instrument, thisstage has been included in case there is a tracking system.Records of translated versions not only can save considerable time and effort (by using already available questionnaires) but also avoid erroneous comparisons of results across different translated versions.The process of cross-cultural adaptation tries to produce equivalency between source and target based oncontent. The assumption that is sometimes made is thatthis process will ensure retention of psychometric properties such as validity and reliability at an item and/or ascale level. However, this is not necessarily the case: Forinstance, if the new culture has a different way of approaching a task that makes it inherently more or lessdifficult compared with other items, it would change thevalidity, certainly in terms of item-level analyses (such asitem response theory, similar to Rasch). Further testsshould be conducted on the psychometric properties ofthe adapted questionnaire after the translation is complete.3,10,26,20 This will be discussed briefly at the end ofthe guidelines. In fact, the translation process outlined inthis article is the first step in the three-step processadopted by the International Society for Quality of LifeAssessment (IQOLA) project.8,25,26 The other two steps

Cross-Cultural Adaptation of Self-Report Measures Beaton et al 3187Table 1. Possible Scenarios Where Some Form of Cross-Cultural Adaptation is RequiredResults in a Change in . . .Wanting to use a questionnaire in a new populationdescribed as follows:ABCDEUse in same population. No change in culture,language, or country from sourceUse in established immigrants in source countryUse in other country, same languageUse in new immigrants, not English-speaking,but in same source countryUse in another country and another languageAdaptation RequiredCultureLanguageCountry of UseTranslationCultural Adaptation————— —— — ——— Adapted from Guillemin et al.4are, first, verification of the scaling requirements (itemperformance, item weights) and, second, the validationof and establishing normative values for the new version.Guidelines for the Cross-Cultural Adaptation ProcessFigure 1 outlines the cross-cultural adaptation processbeing recommended. It is the method currently used bythe American Association of Orthopaedic Surgeons(AAOS) Outcomes Committee as they coordinate thetranslation of the different components of their outcomesbattery. The written documentation of each step helps torecord that it was performed but can also serve as amemory aid at later stages. For instance, if an item is notworking in the field testing, there will be a record showing whether the translators had difficulty with that item,and how they resolved it. Sample forms have been designed for one questionnaire19 so that the worksheetsused for the translation can formulate the written reportFigure 1. Graphic representation of the stages of cross-cultural adaptation recommended.

3188 Spine Volume 25 Number 24 2000as well. The forms are available through the authors, orthrough the AAOS. Each stage in the recommended protocol is described in detail in the following sections.Stage I: Initial TranslationThe first stage in adaptation is the forward translation.Many recommend that at least two forward translationsbe made of the instrument from the original language(source language) to the target language. In this way, thetranslations can be compared and discrepancies that mayreflect more ambiguous wording in the original or discrepancies in the translation process noted. Poorer wording choices are identified and resolved in a discussionbetween the translators.Bilingual translators whose mother tongue is the target language produce the two independent translations.Translations into the mother tongue, or first language,more accurately reflect the nuances of the language.13The translators each produce a written report of thetranslation that they complete. Additional comments aremade to highlight challenging phrases or uncertainties.Their rationale for their choices is also summarized in thewritten report. Item content, response options, and instructions are all translated in this way.The two translators should have different profiles,or backgrounds.Translator 1. One of the translators should be aware ofthe concepts being examined in the questionnaire beingtranslated (functional disability or neck and shoulderdisorders). Their adaptations are intended to providedequivalency from a more clinical perspective and mayproduce a translation providing a more reliable equivalence from a measurement perspective.Translator 2. The other translator should neither beaware nor informed of the concepts being quantified andpreferably should have no medical or clinical background. This is called a naive translator, and he or she ismore likely to detect different meaning of the originalthan the first translator. This translator will be less influenced by an academic goal and will offer a translationthat reflects the language used by that population, oftenhighlighting ambiguous meanings in the original questionnaire.11Stage II: Synthesis of The TranslationsThe two translators and a recording observer sit down tosynthesize the results of the translations. Working fromthe original questionnaire as well as the first translator’s(T1) and the second translator’s (T2) versions, a synthesis of these translations is first conducted (producing onecommon translation T-12), with a written report carefully documenting the synthesis process, each of the issues addressed, and how they were resolved. It is important that consensus rather than one person’scompromising her or his feelings resolve issues. The nextstage is completed with this T-12 version of the questionnaire.Stage III: Back TranslationWorking from the T-12 version of the questionnaire andtotally blind to the original version, a translator thentranslates the questionnaire back into the original language. This is a process of validity checking to make surethat the translated version is reflecting the same itemcontent as the original versions. This step often magnifiesunclear wording in the translations. However, agreement between the back translation and the originalsource version does not guarantee a satisfactory forwardtranslation, because it could be incorrect; it simply assures a consistent translation.18 Back translation is onlyone type of validity check, highlighting gross inconsistencies or conceptual errors in the translation.Once again, two of these back-translations are considered a minimum. The back-translations (BT1 andBT2) are produced by two persons with the source language (English) as their mother tongue. The two translators should neither be aware nor be informed of theconcepts explored, and should preferably be withoutmedical background. The main reasons are to avoid information bias and to elicit unexpected meanings of theitems in the translated questionnaire (T-12),11,18 thusincreasing the likelihood of “highlighting the imperfections.”18Stage IV: Expert CommitteeThe composition of this committee is crucial to achievement of cross-cultural equivalence. The minimum composition comprises methodologists, health professionals,language professionals, and the translators (forward andback translators) involved in the process up to this point.The original developers of the questionnaire are in closecontact with the expert committee during this part ofthe process.The expert committee’s role is to consolidate all theversions of the questionnaire and develop what would beconsidered the prefinal version of the questionnaire forfield testing. The committee will therefore review all thetranslations and reach a consensus on any discrepancy.The material at the disposal of the committee includesthe original questionnaire, and each translation (T1, T2,T12, BT1, BT2) together with corresponding written reports (which explain the rationale of each decision atearlier stages).The expert committee is making critical decisions so,again, full written documentation should be made of theissues and the rationale for coming to a decisionabout them.Decisions will need to be made by this committee toachieve equivalence between the source and target version in four areas11:Semantic Equivalence. Do the words mean the samething? Are their multiple meanings to a given item? Arethere grammatical difficulties in the translation?Idiomatic Equivalence. Colloquialisms, or idioms, aredifficult to translate. The committee may have to formu-

Cross-Cultural Adaptation of Self-Report Measures Beaton et al 3189late an equivalent expression in the target version. Forexample the term “feeling downhearted and blue” fromthe SF-36 has often been difficult to translate, and anitem with similar meaning would have to be found bythe committee.Experiential Equivalence. Items are seeking to captureand experience of daily life; however, often in a differentcountry or culture, a given task may simply not be experienced (even if it is translatable). The questionnaire itemwould have to be replaced by a similar item that is in factexperienced in the target culture. An example might be inan item worded: Do you have difficulty eating with afork? when that was not the utensil used for eating in thetarget country.Conceptual Equivalence. Often words hold different con-ceptual meaning between cultures (for instance themeaning of “seeing your family