Tradução e adaptação brasileira do Questionnaire D'Alimentation

Purpose: to translate the French language instrument "Questionnaire D'Alimentation" to Brazilian Portuguese and culturally adapt it for adolescents. Methods: the translation was based on a protocol that consisted of translation to Brazilian Portuguese (by a professor with a Literary Arts degree and a Doctor of Dental Surgery, both fluent in French and Brazilian Portuguese); back-translation into French; revision by a Committee of specialists (two translators, one native and one sworn translator, and two university professors, one being a Doctor of Dental Surgery and a Speech Therapist) and cultural equivalence (pre-test). The version used in the pre-test consisted of 26 questions with five possible responses (5-Likert), distributed in five domains (Food-Mastication, Habits, Meats, Fruits and Vegetables). The pre-test and test-retest was performed with a sample of 20 adolescents (10 boys/10 girls) from public schools of Piracicaba (Brazil). At this stage, the alternative "I did not understand" was added to each question in order to identify those that were not understood. Test-retest reliability was assessed for each domain using intra-class correlation coefficients (ICCs). Results: in the pre-test, an excellent comprehension of the instrument was observed; in test-retest, ICCs ranged from 0.45 to 0.81 (moderate to excellent agreement). Conclusion: the Portuguese version of the Questionnaire D'Alimentationhas shown to be easy to understand by Brazilian adolescents and useful in the evaluation of the masticatory function and feeding or swallowing disorders that may affect food intake.


INTRODUCTION
Mastication is an essential part of the digestive process, because this is when food is fragmented into smaller particles and mixed with saliva, making it possible to swallow and digest 1 . During the mastication process, contraction of various muscle groups occurs, generating pressure between the tooth cusps, which breaks down the food 2 . The condition of the dentition, such as the number of teeth present and in contact, the size of functional areas, and the adolescents, thereby obtaining a questionnaire that can be used to evaluate the quality of masticatory function in Brazilian Portuguese language.

METHODS
This study was approved by the Research Ethics Committee of the Piracicaba Dental School, University of Campinas (UNICAMP), Protocol Number 108/2012. Those responsible for the subjects signed the Informed Consent. It is important to point out that before the translation processes began, the authors of Questionnaire D'Alimentation were contacted (JF).
The Questionnaire D'Alimentation consists of 38 questions, and the domains "Food-Mastication", "Habits", "Meats", "Fruits" and "Vegetables" comprise 29 questions specifically related to the frequency of and difficulty with mastication of foods of different types of consistency, in the two weeks preceding the subject's evaluation. The remaining nine questions relate to diet specificities (for example, appetite, allergies, reflux, nausea, and other). Each question offers five possible response indicators (5-Likert), according to the content of the question, whether it is about the intensity of the difficulty, (domain "Food-Mastication") or about the frequency of consuming a certain food (other domains). In addition, the domains "Meats", "Fruits" and "Vegetables" also present an alternative to be checked (not applicable -N/A) if the subject does not usually eat these foods.
Translation and cross-cultural adaptation of the Questionnaire D'Alimentation were performed in the following steps, proposed by Guillemin, et al. 16 : initial translation, back-translation, revision by a committee of specialists, and cultural adaptation ( Figure 1).
Previous studies aimed to evaluate masticatory ability by means of questionnaires related to his/ her masticatory capacity, food preference and consistency [8][9][10] . Some have used the Visual Analog Scale (VAS) to make a more generalized evaluation, in which the subject attributes a visual score that varies from 0 to 10 to questions such as "How well do you chew?", or "Do you feel any discomfort when you chew?" 6,8,11 . A simple evaluation has also been found, such as "Are you satisfied with your chewing capacity?", with dichotomous YES/ NO responses 12 . A questionnaire (Questionnaire D'Alimentation) was developed for French speaking subjects who wore partial/complete dentures living in Montreal, Quebec 13,14 . This questionnaire consists of 38 questions, with 29 questions being specifically related to the frequency of and difficulty chewing different types of foods during the two weeks before the evaluation.
Difficulty with mastication is the most probable mechanism by which poor dental health conditions or feeding/swallowing disorders may affect food intake, particularly foods with greater consistency, thereby leading to inadequate intake and having a negative impact on nutritional status. A previous study has shown that the higher the number of teeth affected by caries, the greater the chance of dissatisfaction with mastication 12 . It was also observed that subjects who have masticatory problems also present a higher DMFT (number of decayed, missing and filled teeth), fewer functional teeth, and orthodontic treatment need, both in adults and children 9,12,15 . To measure chewing ability in relation to diet, it is important to gain an understanding of how the subject prepares and chooses their foods.
To our knowledge, there are very few instruments with the aim of evaluating chewing ability specifically. Thus, the aim of the present study was to translate and cross-culturally adapt the instrument Questionnaire D'Alimentation for Brazilian

Revision by a Committee of Specialists and Cultural Adaptation
The versions V1 and V2 in Brazilian Portuguese and the Versions V3 and V4 in French, as well as the original instrument, were submitted to a Reviser Committee formed by two French language professors (one being native and the other a certified translator), a Speech Therapist -University Professor (RCB) and a Doctor of Dental Surgery -University Professor (PMC). This stage consisted of the following aspects 17 : • Semantic equivalence: This refers to the meaning of words; words that do not have a literal translation with a similar meaning were translated into terms in Brazilian Portuguese that had an equivalent meaning; • Idiomatic equivalence: The formulation of colloquial expressions equivalent to those in the original language; • Cultural equivalence of each question: experiences undergone within the cultural context of the society. In this stage, the cultural adaptation was performed, based on the target population of the research; that is, Brazilian adolescents who were not denture wearers. Therefore, three questions related to denture wearing were excluded from the domain "Habits" (questions 19, 23 and 24). Thus version V5 in Brazilian Portuguese was obtained.

Cultural Equivalence of the Instrument
Finally, to evaluate the cultural equivalence of the instrument, the version V5 in Brazilian Portuguese (Quality of Masticatory Function Questionnaire -QMFQ) was self-applied by 20 adolescents 18 from public schools of Piracicaba (SP, Brazil), 13 (n=12) and 14 years of age (n=8), of both genders (10♂/10♀), selected by means of a draw from all authorizations received, under the supervision of two researchers (MHM and TSB). In version V5 the option "I didn't understand" was added as an alternative response to all questions, as a way of identifying questions that had not been adequately understood. The percentage of "I didn't understand" responses had to be lower than 15% so that the instrument could be considered culturally adapted 19 . If the established limit was exceeded, the instrument would have to be submitted to a new cultural adaptation process, until the item "I didn't understand" had not been chosen in any question by 85% or more of the adolescents.
For assessment of test-retest reliability, the same volunteers were invited to fill out a second copy of the questionnaire one week later for Intraclass Correlation Coefficient (ICC) determination using the BioEstat 5.3 (Mamirauá, Belém, PA, Brazil) statistical software package.

Initial Translation
The version in French (original questionnaire) was initially translated to Brazilian Portuguese by a Professor with a degree in Literary Arts, as well as a Doctor of Dental Surgery and University Professor (RIF), both fluent in French and Brazilian Portuguese, and aware of the aim of this study, emphasizing conceptual translation rather than literal translation (versions in Brazilian Portuguese V1 and V2).

Back-translation
The versions in Brazilian Portuguese V1 and V2 went through back-translation into French, done by two native French teachers, who did not participate in the first stage of translation, and who had no access to the original instrument, thus obtaining the translations in French V3 and V4. The purpose of back-translation is to compare the back translationed version with the original instrument.

Final Version V5
Back-Translation V4

V5 in Brazilian Portuguese
Revision by Committee Translation V2 Translation V1

Figure 1 -Stages of the processes of translation and cultural adaptation of the instrument Questionnaire D'Alimentation
original instrument were submitted to a Reviewing Committee. Figure 2 presents the questions in the original version and their translations, as well as a synopsis of the decision making process relative to specification of the first and second version in Portuguese (V1 and V2) made by the committee. For some questions, the translations made by both translators were identical, or practically identical; for others, one or other version was prioritized; moreover, in other questions, the option taken was to combine the two versions, generating a version of consensus, with the purpose of obtaining greater clarity of the item.

Initial Translation and back-translation
Each version in Brazilian Portuguese (V1 and V2) was independently re-translated into French (back-translation) by two native French teachers, who were not aware of the purpose of the work, thereby giving rise to the origin of versions V3 and V4 in French.

Revision by a Committee of Specialists
The versions V1 and V2 in Brazilian Portuguese and the versions V3 and V4 in French, as well as the  of the difficulty (extreme, very, moderate, hardly, no difficulty) and frequency (never, rarely, sometimes, frequently, always, not applicable). For the first item, the option was to choose the modified version (V3). For the frequency scale, the option was for the V1; however, the term "à l´occasion", which was translated as "occasionally", was changed to "sometimes" by the Committee, which is the word more routinely used.
After conclusion of the process of translation and revision by the Committee, three questions related to denture wearing were excluded from the domain "Habits" (questions 19, 23 and 24).

Cultural Equivalence of the Instrument
The use of the "Quality of Masticatory Function Questionnaire" was considered easy, fast and the questions were well understood by the subjects, since the alternative "I didn't understand" was not checked by any of the participants. On the other hand, of the 20 adolescents who participated in pre-test, six checked the alternative response N/A (not applicable) to at least one question in the domains "Meats", "Fruits" or "Vegetables" ( Figure  3). Of these six participants, only one handed in the questionnaire with the responses left blank, being one question with reference to the domain "Fruits" (19: Bite into whole raw apples) and three questions of the domain "Vegetables".
The test-retest reliability of the questionnaire was assessed for each domain, and the ICCs found were: 0.79 (Food-mastication), 0.45 (Habits), 0.62 (Meats), 0.74 (Fruits) and 0.81 (Vegetables), showing moderate to excellent agreement (p<0.01) 20 . The final version of the instrument is shown in Figure 4.
At this stage, replacements were made of terms presented in V1 and V2 by synonyms, so that the terms would be better understood by the target population. Questions that sought to specify the size of the food (meats, fruits and vegetables) were those that generated the greatest difficulty. For example, the literal translation of the expression "la grosseur d´un dé à coudre" would be "the thickness of a thumb". The second translator had also suggested the term "inch", which is commonly used by the Brazilian adult population. However, the target population (adolescents) may not know the size of an "inch" and so the Committee suggested the use of the expression "small pieces".
In questions 13, 14, 15, 31, and 36 the term "croquer" was used, which was translated as "grind" in V1 and as "chew" in V2. The Committee therefore suggested the use of "bite", because this would linguistically be more accept and easier to understand, since "chew" would be translated into French as "mastiquer".
Question 17 asks whether the subject has difficult with chewing "pain croûté". This expression was translated by both translators as "toasted bread"; whereas for the Committee this translation was not acceptable, and it was replaced by "bread with a hard crust". The term "hacher" (questions 27 and 29) was translated as "chop" (V1) and "mince" (V2), and it was afterwards changed by the Committee, which proposed the term "shred" (the meat). Question 30 also generated doubts because of the expression "metre la viande em purée", and it was translated as "Boiling the meat till it was tender".
As regards the alternative scales to be chosen, there were two to be translated: the scale of intensity

Figure 3 -Pre-Test Synopsis: Description of the questions with alternative responses "not applicable" (N/A) or left blank (X) checked by six subjects
Este questionário visa avaliar sua escolha por alimentos em função de sua capacidade de mastigar nas duas últimas semanas.

DISCUSSION
An instrument may only be considered valid if it is capable of accurately capturing a certain concept 21 . Furthermore, in the culture for which it is being adapted, a translated instrument must be capable of obtaining the same effect as the original instrument has in the context in which it was created. The lack of cultural equivalence compromises the validity of the information collected, making it impossible to use the instrument to study a concept correctly 22 . This is why there are standardized instructions that are intended to minimize the loss of the original instrument characteristics that could result from the change in language 16,[22][23][24] . Therefore, the methodology of the present study followed the protocol suggested by Guillemim, et al. 16 , by which mistakes and misinterpretations in the initial translations are revealed. During these stages, semantic equivalence was appreciated, and not the literal interpretation, between the terms, since the literal interpretation is not always shown to be more advantageous to express concepts or situations of the new population one wishes to study 16,[24][25][26] .
In this process, using and comparing more than one version is relevant. In addition to making it possible to choose items to incorporate, or to allow the junction of items arising from different versions, this strategy allows to carefully examine the sequence of the procedures, including the translations themselves, their re-translation and the appreciation that follows. The importance of a general criticism made by the committee of specialists must also be pointed out; among the members of the committee there must be individuals who are specialists in the disease investigated, in the measure used and in the concept explored, and they should preferably be bilingual 16 .
The differences between definitions, beliefs and behaviors demand that the use of an instrument drawn up in other cultural contexts should, in addition to reliable translation, be preceded by cross-cultural adaptation to the country in which it will be applied, in order to keep the same concepts as those of the original 27 . At the stage of cross-cultural adaptation, the number and characteristics of volunteers involved in pre-test were chosen in agreement with methodological criteria used for cultural adaptation of questionnaires 16,18 .
Originally it was proposed that this instrument should be self-applied, being necessary for the respondent to read the instructions carefully before answering the questions; therefore, future studies will be able to show whether there is equivalence. In this study, it was considered feasible for researchers to provide initial instructions to the subjects, with the aim to ensure that the rules for filling out the questionnaire were understood by the studied sample. For responses left blank or for alternative response N/A, the missing value may be replaced by the mean/median scores for each domain or for each subject 28,29 .
The process of translation and cultural adaptation of health-related questionnaires must keep a conceptual basis and follow standardized methodologies, since differences in subject's habits and culture may lead to misunderstandings, thereby altering the psychometric properties of the instrument. Therefore, in order to fully achieve cultural adaptation, it is also necessary to conduct a study of equivalence measurement, evaluating the reliability and validity of this new version in different populations and conditions (e.g. subjects with dental caries and/or periodontal disease, patients presenting feeding and/or swallowing disorders, denture users, patients undergoing orthognathic or bariatric surgery, and others).

CONCLUSION
The Portuguese version of the Questionnaire D'Alimentation has shown to be easy to understand by Brazilian adolescents and useful in the evaluation of the masticatory function and feeding or swallowing disorders that may affect food intake.