The McGill Illness Narrative Interview-MINI : translation and cross-cultural adaptation into Portuguese

This paper presents the process of translation and cultural adaptation into Portuguese of the McGill Illness Narrative Interview – MINI, an interview protocol that is used to research meanings and modes of narrating illness experiences, tested, in the Brazilian context, for psychiatric and cancer-related problems. Two translations and their respective back-translations were developed. In addition, semantic equivalence was evaluated, a synthesis version and a final version were prepared, and two pre-tests were administered to the target populations (people with auditory verbal hallucinations or breast cancer). A high degree of semantic equivalence was found between the original instrument and the translation/back-translation pairs, and also in the perspective of referential and general meanings. The semantic and operational equivalence of the proposed modifications was confirmed in the pre-tests. Therefore, the first adaptation of an interview protocol that elicits the production of narratives about illness experiences has been provided for the Brazilian context.


Introduction
This paper presents the process of translation and cross-cultural adaptation into Portuguese of the McGill Illness Narrative Interview -MINI, an interview protocol for eliciting narratives of experiences and meanings concerning illness and symptoms.The translation was developed by researchers working at the Laboratory of Psychopathology and Subjectivity Studies of the Psychiatry Institute of UFRJ (Federal University of Rio de Janeiro).The interview was designed by Danielle Groleau, Allan Young and Laurence Kirmayer, with the Division of Social and Transcultural Psychiatry of the McGill University (Montreal, Canada) and was originally published in English in 2006 1 .
McGill MINI is a semi-structured, qualitative interview that enables the production of narratives about illness experiences related to any health problem, condition or event, including symptoms, set of symptoms, syndromes, biomedical diagnoses or popular labels.Depending on the research question, it can be used to investigate the illness experience of one individual or a group, to compare individual experiences, to survey shared cultural aspects, health behaviour categories or narrating modes of certain cultural groups 1 .
The McGill MINI is sequentially structured.It has three main sections and two supplementary ones, and aims to elicit: 1 -an initial and temporal narrative of illness experience, organized according to the sequence of events.
2 -a narrative of other previous experiences of the interviewee, family members, friends, experiences found in the media, and other popular representations that served as a model for the signification of the illness experience.These experiences emerge as prototypes related to the studied health problem.
3 -narratives in the form of explanatory models of the symptom or illness, including labels, causal attributions, treatment expectations, course and result.
4 -narratives related to search for help, reports on paths taken to receive care and on the experience of treatment and adherence.
5 -narratives about the impact of illness on identity, self-perception and relationships with others.
Multiple representational schemes and different meaning attribution modalities are used to produce narratives that are complex and, sometimes, internally inconsistent or contradictory.The utilization of the McGill MINI enables the examination of these multiple meaning attribution modes based on the identification of: i -explanatory models grounded on causal opinions that may involve conventional models, causal attributions or more elaborate models that involve specific processes or mechanisms similar to the biomedical model.
ii -prototypical models that involve meaning attribution modes based on episodes or events that emerge from one's own life or from the life of others and enable individuals to attribute meaning to their experience through analogy.
iii -chain complexes in which past experiences are metonymically linked to present symptoms through a sequence of events around the symptoms, without any explicit causal connection or evident pattern.
Medicine and narrative have always walked together if we consider the patient/illness tension that is inherent in medical practice 2,3 .However, the strength and visibility of this association have reached different expressions throughout the history of medicine.Today, we see the prevalence of the biomedical model, a model that values neither history nor context in the understanding of illness.In the field of mental disorders, the classification manuals induce professionals to make diagnoses based on a list of symptoms.Despite this, bibliographic reviews about narrative and medicine have indicated that the importance of the former in medical literature has increased in the last two decades 2,4 .The study of narratives has been valued in discussions about ethical and epistemological aspects of the clinical method and in medical education 5 .Trisha Greenhalgh 6 , a reference in narrative-based medicine, highlights that this perspective is fundamental in times of evidence-based medicine because: "Similarly (but for different reasons), the "truths" established by the empirical observation of populations in randomised trials and cohort studies cannot be mechanistically applied to individual patients (whose behaviour is irremediably contextual and idiosyncratic) or episodes of illness." Connelly 7 argues that: "If the patient's narrative is not heard fully, the possibility of diagnostic and therapeutic error increases, the likelihood of personal connections resulting from a shared experience diminishes, empathic opportunities are missed, and patients may not feel understood or cared for" Consequently, first-person reports 8,9 have become important tools to the understanding of the experienced illness process, to adequate clini-cal judgment and to the design of the therapeutic project.Diagnosis and treatment protocols, independently of the health problem under scrutiny, are insufficient to instrumentalize clinical judgment and the conduction of a therapeutic project.Clinical judgment and the definition of the therapeutic project require an interpretative work that takes into account the characteristics of the experience: the consideration of the way in which the subject lives and experiences his illness in his relationship with himself and to his environment.According to Kleinman et al. 10 , learning about the human illness experience allows knowing how the patient, the members of his family or the nearest social network perceive, interact with and respond to the symptoms and to the incapacity that can derive from them, as well as to the monitoring of body processes.Without this dimension, the possibility of success of any therapeutic intervention project becomes limited.
Providing the Brazilian scientific community with an interview protocol that elicits narratives about illness experiences and enables the investigation of the multiple ways individuals use to attribute meanings to them is undoubtedly a relevant enterprise.A brief survey in national journals has indicated that there is no standardized tool in Brazil to access the experiential dimension of illness or the process of experiencing symptoms.In spite of this, studies about narrative and medicine have become increasingly frequent in the literature 2,4,5 , as mentioned above.They reveal both the current relevance of the study of narratives and the growing importance of qualitative research in the field of health.In this context of increasing interest in the modes of meaning construction about illness experiences, the emergence of an interview script that approaches three central questions in the field of qualitative health research is extremely relevant.The three questions are the following: -How does a subject construct his knowledge about his illness experience?
-What types of knowledge support narratives of illness experience?How are they organized and structured?
-Is it possible to develop reliable studies about narratives?
These are the reasons that explain our interest in translating and validating the McGill Illness Narrative Interview -MINI to the scientific community of qualitative health research in Brazil.
In the next sections, we present the process of translation and cross-cultural adaptation of the McGill MINI, as well as the final version in the Portuguese language for current use in Brazil.

Methodology
The process of translation and cross-cultural adaptation was based on the method proposed by Herdman et al. 11 , which has already been used in Brazil by authors like Reichenheim et al. 12 , Moraes et al. 13 , Fizman et al. 14 , and Mattos et al. 15 , among others.Overall, the process lasted approximately one year: it started at the beginning of the second semester of 2007 and ended in the second semester of 2008.
The method has seven stages: 1) translation of the original interview, 2) back-translation, 3) evaluation of semantic equivalence, 4) preparation of the synthesis version, 5) pre-test in the target population, 6) preparation of the final version, 7) second pre-test in the target population with final impressions provided by specialists in the area.
In the first stage, two translations of the original instrument in English into Portuguese were carried out, independently, by two professionals in the field of health, experienced and fluent in the English language (T1 and T2).In this stage, we considered operational equivalence -which is the possibility of using the interview script with the same organization and mode of administration as those of the original instrument 16 -with the purpose of maintaining the characteristics of the original interview script, preserving its reliability and validity.The same number of questions was maintained, as well as the same division of sections, the same introduction and the same instructions to each one of the 46 questions.
In the second stage, the two translations (T1 and T2) were back translated into English, also independently, by two bilingual translators, native speakers of English.
Two evaluations constituted stage 3: an evaluation of referential meaning and an evaluation of general meaning.The evaluation of semantic equivalence, performed by two researchers, took into account the referential and general meanings.In the perspective of the referential meaning of words, the equivalence between the original instrument and each back-translation was evaluated.Referential meaning is related to the ideas and objects of the world to which one or more words refer.That is, whether one word in the original instrument has the same referential meaning of the corresponding word in the back-translation.
The second evaluation in stage 3 was related to the general meaning of each item of the original instrument compared to the corresponding item in each Portuguese version.The general meaning takes into account not only literal correspondence between words, but also more subtle aspects, such as the impact they have in the cultural context of the target population.Divergences between the equivalence analyses in this stage were the focus of discussions, which conducted the group to the decisions made in the next stage.
Stage 4 was characterized by the preparation of a synthesis version.Some items were incorporated from one of the two versions, in full or modified by the group, while others resulted from the junction of the two versions.The content of this junction underwent some modifications to better meet the criteria of semantic equivalence.
Stage 5 required a pre-test of the synthesis version in a sample of the target population to detect incongruences of meanings between this version and the original instrument.
The following stage -stage 6 -consisted of a discussion about the acceptability of this version in the evaluated population and the proposition of new modifications that guided the preparation of the final version.
In the seventh and last stage, the final version was administered to a set of people whose sociodemographic and diagnostic characteristics were similar to those of the people who participated in stage 5. Two expert researchers listened to the recordings of the interviews administered in stage 7.They used an analysis card to make notes on necessary adjustments.
The synthesis and final versions of the Mc-Gill MINI were administered to a set of 28 people who were experiencing at that moment or had experienced a symptom or disease.This initial number was open to revision, in case clues on the process of translation and cross-cultural adaptation of the interview emerged, indicating, for example, that the questions were not clear or were difficult to understand, or differences in the apprehension and effect of the questions deriving from differences between the two languages, considering the questions' objective in the original version.The target population was defined taking into account the study interests of the researchers involved.People who experienced auditory verbal hallucination and people with breast cancer constituted the studied group.In addition to the researchers' study interests, another factor that contributed to the constitution of the target population's profile was the interest in validating the Portuguese version in diverse experiences of illness and symptoms, so as to guarantee the generic character of the translated version, one of the main purposes of the original English version.
In stage 5, the synthesis version was administered to eight patients who attended the voice hearing group of the Daily Care Center of the Psychiatry Institute (IPUB) of UFRJ, a clinical and research service, and to six patients with breast cancer recruited by the research project Characterization of BRCA1 and BRCA2 Gene Mutations in a Population of Women with Breast Cancer in Rio de Janeiro; applications to prophylactic interventions and studies on psychosocial impact, developed at the Clementino Fraga Filho University Hospital (HUCFF) of UFRJ.The final version was administered to people whose sociodemographic and diagnostic characteristics were similar to those of the individuals who participated in stage 5 and constituted a universe of 14 people, composed of two sub-sets.Sub-set 1 was constituted of eight people with auditory verbal hallucination and sub-set II was composed of six individuals with breast cancer.In this stage, the interviewed population was not necessarily involved in the studies mentioned above.All the participants voluntarily accepted to answer the questions of the interview script and signed a consent document.
The interview was administered by researchers, Master's and undergraduate students linked to the Laboratory of Psychopathology and Subjectivity Studies, IPUB/UFRJ.All of them were duly trained in a workshop conducted directly by one of the authors of the interview or by researchers trained by her.

Results and discussion
The result of stage 1 (translation of the original interview from English into Portuguese), performed by two authors, and of stage 2 (back-translation into English), performed by two native speakers of English, constituted the material to be analyzed in stage 3 (evaluation of semantic equivalence), performed by two other authors in two steps, A and B. In step A, the referential meaning of the questions that composed the script of the original interview was compared to the referential meaning of the questions of the back-translations, and scores from 0% to 100% were attributed.In step B, general meaning was evaluated through a comparison between the questions of the original script and those of the translations, which were classified in four levels: unaltered, little altered, much altered or completely altered.
The results of stages 1, 2 and 4 are exemplified on Chart 1 for four questions of the McGill Illness Narrative Interview -MINI.
When we evaluated semantic equivalence, the two translations and back-translations obtained reasonable measures of equivalence of general and referential meaning, respectively, in relation to the original interview.The steps described on Chart 2 were followed in the conduction of the analysis developed in this stage.In the evaluation of semantic equivalence -general meaning, which takes into account literal correspondence between words and also the impact they have on the cultural context of the population, 4 questions were classified as completely altered and 6 questions as much altered for the first pair.For the second pair, only 1 question was considered completely altered and 1 question much altered (the translation of the verb 'experienciar' and of verb tenses varied importantly).In the evaluation of semantic equivalence -referential meaning, which regards the ideas and objects of the world to which one or more words refer and which observes whether a word in the original instrument has the same referential meaning of the corresponding word in the back-translation -, the concordance in the first pair varied between 40% and 100% for the set of questions of the interview, and reached an average of 88%.The concordance in the second pair also varied between 40% and 100%, with an average of 90%.
To elaborate the synthesis version (stage 4), the content of the two translations was joined.Whenever necessary, small modifications were made with the aim of ensuring greater clarity, greater fidelity to the original version, and also to guarantee the elicitation of narratives -the purpose of the original interview.In all cases, we decided to choose words that could be understood by people belonging to a broad range of levels of schooling, an instruction that had not been provided for the translators.Thus, words like to experience, very important in the original version, were replaced by others that ensured the same meaning, but whose use was more frequent in colloquial language.Although 'experienciar' is the correct translation of the verb to experience, it is not a verb that is frequently used in Portuguese when we speak, and it would present a great risk of not being understood or generating confusion.The verb 'sentir' adopted by the authors in this stage was understood by the participants and could ensure the meaning desired by the origi-nal interview.Likewise, we attempted to achieve correspondence between perception and the impact of different words 11 .For example, the words helper and healer (If you went to see a helper or healer of any kind, tell us about your visit and what happened afterwards.)might be adequately translated as 'ajudante' and 'curandeiro'.Although the word 'ajudante' is common in Portuguese, it is not used in this context.The word 'curandeiro', in turn, is more frequently used in contexts of care and cure, but it often acquires, in certain contexts, a negative connotation.In this case, the authors decided to describe what these words intended to indicate (some kind of spiritual help or treatment, alternative treatment or of any other type) in order to ensure the purpose of the question.A similar situation happened with the use of the verb to go, which has many meanings in the English language.In the case of the question quoted above -If you went to see a... -, with the purpose of ensuring the highest degree of clarity, we decided to use the verb 'procurar' and split the question into two (Você procurou algum tipo de ajuda, tratamento espiritual, tratamento alternativo ou tratamento de qualquer outro tipo?Nos fale como foi e o que aconteceu depois).
In stage 5 -administration of the synthesis version -, there were no great difficulties to understand the questions.Due to this, stage 6 required only some adjustments.For example, changing verb tenses so that the questions favored the desired understanding and were closer to colloquial language, or changing the form of construction of some questions in order to guarantee that their use elicited the type of narrative desired by the original interview.
One example is that the more literal translation of Question 1 in Section I -Quando você sentiu que estava com o seu problema de saúde ou dificuldades pela primeira vez? -tied the interviewees to the temporal aspect, but the question aimed to make them talk about circumstances that were present at that moment, too.In this case, we decided not to use the conjunction 'quando' and constructed the question in another way, to guarantee that it stimulated the desired narrative: Fale sobre a primeira vez que você sentiu que estava com o seu problema de saúde ou dificuldade (PS) (Chart 1).
Stage 7 elapsed almost with no problems of understanding.The only suggested change was in question 37 (Section III): Que outra terapia, tratamento, ajuda ou cuidado você buscou?, in which the verb 'buscou' was replaced by 'procurou': Que outra terapia, tratamento, ajuda ou cuidado você Chart 1. Examples of results of the stages of translation (T1 and T2), back-translation (B1 and B2) and synthesis version (Synthesis).procurou?.The aim was to ensure greater clarity, as the verb 'buscar' , in this context, was hard to understand and the utilization of the verb 'procurou' , which is literally closer to the verb used in English, 'sought out' , proved to be more adequate.

Original
Finally, we suggest that section III, which aims to elicit narratives about the explanatory model adopted by the interviewee, should be integrally administered, even if the interviewee does not have a popular term to describe his health problem.When we listened to the audio-recorded interviews, we realized that, even when the interviewee does not have a popular term to indicate his health problem and uses a medical term to describe it, asking questions 21 to 27 helps the interviewer to investigate the meanings that the interviewee attributes to the medical term and how he uses the term to describe and understand his health problem.
The final version of the cross-cultural adaptation of the McGill Illness Narrative Interview -MINI into Portuguese proposed in this paper is presented on Chart 3.

Conclusion
This work provides the first adaptation to the Brazilian context of a specific instrument to elicit illness narratives, with generic character, useful to experiences related to indisposition, symptoms or diseases.The translation and validation of this interview protocol were performed to two different populations: people with psychiatric symptoms and people with physical problems.The analysis of the seven stages of the cross-cultural adaptation of the McGill Illness Narrative Interview -MINI met the criteria of semantic equivalence and indicated that this interview script can be used, in our environment, to access the same type of narrative about illness experience that it proposes to elicit in its culture of origin.

EVAL.2 Work Methodology of Step B:
Organize, on a table, the items of the original text, the versions into Portuguese (V1 and V2) and the evaluation by two evaluators (EVAL.1 and EVAL.2) of the general meaning.