Translation into Portuguese of the M-CHAT Scale for early screening of autism

Abstracts

INTRODUÇÃO: A escala Modified Checklist for Autism in Toddlers (M-CHAT) é um instrumento de rastreamento precoce de autismo, que visa identificar indícios desse transtorno em crianças entre 18 e 24 meses. Deve ser aplicada nos pais ou cuidadores da criança. É auto-aplicável e simples, e apresenta alta sensibilidade e especificidade. Foi desenvolvida no idioma inglês e ainda não está disponível uma versão em português. Na literatura não existe consenso quanto à técnica de tradução, sendo a adaptação transcultural uma das formas possíveis. O objetivo do presente estudo foi realizar a tradução do inglês para o português do Brasil da escala M-CHAT para rastreio precoce do autismo, respeitando a equivalência transcultural. MÉTODO: Foi obtida permissão da autora da escala, realizada tradução, seguida de sua respectiva back-translation, avaliação da equivalência referencial, primeiro pré-teste em amostra da população-alvo, avaliação da equivalência geral, avaliação de especialistas em autismo infantil, elaboração da versão preliminar, segundo pré-teste em pais de crianças de ambulatório de pediatria do SUS e elaboração da versão final. RESULTADOS: A avaliação da equivalência referencial demonstrou que 78% das questões eram semelhantes, 13% aproximadas e 9% diferentes. Das 20 pessoas interrogadas no primeiro pré-teste 9 entenderam 100% das questões. O segundo pré-teste demonstrou boa aceitação e entendimento pela população-alvo, com 70% dessa sem nenhuma queixa. Após as avaliações pormenorizadas foi elaborada a versão final. CONCLUSÃO: O estudo torna disponível a versão em português da escala M-CHAT, considerada adequada por especialistas e compreensível pela população.

Transtorno autístico; escalas; tradução (produto)

ORIGINAL ARTICLE

Translation into Portuguese of the M-CHAT Scale for early screening of autism^*

Mirella Fiuza Losapio^I; Milena Pereira Pondé^II

^IMedical student, Escola Bahiana de Medicina e Saúde Pública (EBMSP), Salvador, BA, Brazil

^IIPhD. Associate professor, Pharmacology and Psychiatry, EBMSP

Correspondence

ABSTRACT

INTRODUCTION: The Modified Checklist for Autism in Toddlers (M-CHAT) is a screening instrument for autism that can be applied to parents of children aged 18-24 months. It is self-applicable, simple, and has high sensibility and specificity. It was developed in English and it is still not available in Portuguese. In the scientific literature there is no agreement as to the translation technique, cross-cultural adaptation being one of the possibilities. The objective of the present study was to translate the M-CHAT scale from English into Brazilian Portuguese to scan early autism, preserving cross-cultural equivalence.

METHOD: After obtaining permission from the author of the scale, its translation was performed, followed by its back-translation; evaluation of reference equivalence; first pretest in a sample of the target population; evaluation of general equivalence; evaluation by specialists in infantile autism; development of a preliminary version; second pretest with parents of autistic children treated in a Pediatric Ambulatory of the Brazilian Public Health System; and development of its final version.

RESULTS: Evaluation of reference equivalence showed that 78% of the questions were similar, 13% approximate, and 9% different. Of the 20 people interviewed in the first pretest, nine understood 100% of the questions. The second pretest demonstrated good acceptance and understanding by the target population, with no complaints in 70% of the sample. After detailed evaluations, the final version was developed.

CONCLUSION: The study makes the Portuguese version of the M-CHAT scale available, which was considered adequate by specialists and understood by the population.

Keywords: Autistic disorder, scales, translation (product).

Introduction

The current definition of autism, both in the 10th review of the International Classification of Mental Diseases (ICD-10)¹ and in the Diagnostic and Statistic Manual of the American Psychiatric Association (DSM-IV),² refers to a complex developmental disorder, characterized by impairments in three behavioral areas: social interaction, communication and patterns of interest and repetitive and stereotyped behaviors.³ Abnormalities in the functioning of each area should be present at around 3 years of age. Clinical manifestations are widely varied in terms of severity levels. Diagnosis of autistic disorder requires at least six behavioral criteria, one in each area of altered behavior.⁴

Use of screening questionnaires provides the best method for early identification of children with autism.⁵ Many instruments have been used to identify autism, such as the Childhood Autism Rating Scale (CARS)⁶; the Social Communication Questionnaire (SCQ)⁷; the Screening Tool for Autism in Two Years Old (STAT)⁸; the Developmental Behavior Checklist (DBC)⁹; the Checklist for Autism in Toddlers (CHAT)¹⁰; and the Modified Checklist for Autism in Toddlers (M-CHAT).¹¹ One of the most widely used assessment scales is the CARS, which consists of a 15-item structured interview with parents or guardians of an autistic child older than 2 years.⁶ The SCQ scale is comprised of 40 questions and distinguishes between autism and mental retardation in children older than 4 years.⁷ STAT is an interactive assessment of children aged 24-35 months of age, administered by a trained physician.⁸ DBC is a 96-item scale aiming to evaluate emotional and behavioral problems in children aged 4-18 years, with intellectual deficit.⁹ CHAT was developed to identify children with risk for autism at 18 months of age.¹² A 6-year study conducted in the United Kingdom including more than 16,000 children screened by CHAT at 18 months of age showed only 0.40 sensitivity and 0.98 specificity, with positive predictive value of 0.26.¹³

M-CHAT is a screening scale that can be used in all children during pediatric visits, with the aim of identifying traits of autism in children at an early age.¹¹ Screening instruments are useful to assess individuals that are apparently well, but that have a disease or risk factor for disease, different from those that do not have symptoms.¹⁴ M-CHAT is extremely simple and does not need to be administered by physicians. Response to scale items takes into account parents' observations toward the child's behavior, takes only a few minutes to be completed, is not dependent on previous scheduling, has low cost and does not cause discomfort to patients.¹¹ This scale is an extension of CHAT. It consists of 23 yes/no questions that should be self-administered by parents of children aged 18-24 months of age. Parents should be literate and escorting their child in the pediatric visit. CHAT format and the first nine items were maintained. The other 14 questions were developed based on a list of symptoms that are often present in children with autism.¹¹ The authors of M-CHAT performed a validation study in the USA, with a sample of 1,122 children aged 18 months and that were having routine pediatric visits and with another sample of 171 children participating in early screening in specialized services.¹¹ In that study two criteria were used to determine M-CHAT sensitivity and specificity. The first used any of the 23 questions, with 0.97 sensitivity and 0.95 specificity. The second used two of the best six questions, with 0.95 sensitivity and 0.99 specificity.¹¹ The M-CHAT scale has been translated into the following languages: Icelandic, Japanese, Arabic, Chinese, Cingalese, Dutch, French, Spanish (from Puerto Rico and Spain), German, Hindi, Tamil, and Turkish, and it is being translated into Thai.

This study aims at translating the M-CHAT scale for early screening of autism into Brazilian Portuguese, respecting cross-cultural equivalence.

Method

Development of the Brazilian Portuguese translation of the M-CHAT questionnaire was based on a model proposed by Reichenheim & Moraes¹⁵ and consists of the following steps: obtaining permission from authors, translation, back-translation (technique that consists of translating a previously translated document back to its original language; such technique enables improvement in translation quality),¹⁶ evaluation of reference equivalence, first pretest, evaluation of general equivalence, evaluation by experts, development of a preliminary version, second pretest, and development of the final version.

The first step was contacting M-CHAT author Diana Robins by e-mail. Translation of the original questionnaire into Brazilian Portuguese was performed by a Brazilian psychiatrist fluent in English. Next, the text translated into Portuguese was sent to an American professional translator fluent in Portuguese that was not aware of the original scale in English. Back-translation was performed based on the translation into Portuguese.

Evaluation of reference equivalence consisted of a comparison between the back-translated version (post-back-translation) and the original scale. An external rater fluent in English, whose original language is Portuguese, evaluated the literal correspondence between the terms in the original version and the back-translated version, classifying them as similar, approximate or different.

The first pretest was performed by a medical student that applied the preliminary version to 20 parents of children receiving care at the pediatric outpatient clinic of the Brazilian Unified Health System (SUS), Escola Bahiana de Medicina e Saúde Pública (EBMSP), in Salvador, Brazil. Inclusion criteria were being a father or mother of a child aged 18 years or younger in good health condition, being literate, having Brazilian Portuguese as native language, and waiting for first visit or return visit. In this step the person in charge of applying the questionnaire read the instrument along with the interviewee in the clinic waiting room and, at the end of each question, asked him to provide an example of the behavior mentioned in the M-CHAT. Next, the interviewer judged the patient's understanding and, in case it was not considered good enough, marked the item for further analysis.

A psychiatrist and the interviewer evaluated the terms that are closer to the description of the characteristics of a child with autism in the Brazilian cultural setting. Pretest results were also taken into account, showing how the population understood each question. After a discussion, semantic adjustments were performed and then the first preliminary version was developed.

This version was e-mailed to a neuropediatrician and a psychologist specialized in infantile autism in Salvador with the aim of evaluating semantic equivalence and pertinence to the new context to which it was being adapted. The specialists criticized the translation, also via e-mail, and suggested some changes. Each suggestion was analyzed and discussed by the psychiatrist that translated the scale and by the interviewer. Reference and general meanings of the expressions were taken into account. At that moment the second preliminary version was developed. The second pretest consisted of applying the second preliminary version to 20 parents of children receiving care at the pediatric outpatient clinic of SUS, Escola Bahiana de Medicina e Saúde Pública (EBMSP), in Salvador, using the same criteria of the first application. According to the pretest results, the translator and the interviewer made the last changes to the scale.

The project was sent to the Ethics Committee of Fundação Bahiana para Desenvolvimento das Ciências, and was approved by protocol no. 80/2007 on December 5, 2007.

Results

Obtaining permission from the authors

Contact with Diana Robins was made in April 2007 by e-mail. Such contact served to obtain permission to translate the scale, as well as suggestions of study methodology.

Translation

The first translation was performed by a psychiatrist with more than 5-year experience working with autistic patients, who has Portuguese as his native language and is fluent in English. The back-translation was performed by a translator that is fluent in Portuguese and has English as her native language.

Evaluation of reference equivalence

This evaluation was based on a comparison between terms and phrases of the original version and the back-translated version in each question. Each item was classified as similar when both versions were completely equal; approximate when words or phrases were different, but with no change in context; and different when there was change in the literal meaning between the original and the back-translated version. Of 23 questions, 18 were considered similar (78%); three were approximate (13%); and two were different (9%). Based on that classification, the two questions classified as different were modified.

In question 8, based on the comparison between the original and back-translated versions , "...sem simplesmente ficar murmurando sozinho, mexendo no brinquedo..." was replaced by "...sem simplesmente ficar colocando na boca, remexendo no brinquedo..." In the other question considered different (number 11) "...já se mostrou hipersensível ao barulho..." was replaced by "...já pareceu hipersensível ao barulho..." Table 1 shows the comparison of each question between the original and back-translation version, in addition to the translated version and its correction.

Table 1 - Click to enlarge

Pretest

Acceptance and understanding of the first pretest were reasonable, as nine out of 20 interviewees understood 100% of the questions. With regard to schooling, 55% of the study population was comprised by individuals that concluded high school, 30% had incomplete high school, and 15% had incomplete elementary school.

Evaluation of general equivalence

Considering the results from the first pretest, along with the clinical observations of which terminology is more commonly used by the Brazilian population relative to autism, the general meaning of the scale was evaluated and the following changes were made:

In question 2, "Seu filho se interessa por outras crianças?", the phrase "ter interesse por" was understood by 10% of the population in the pretest as having a feeling of affective relationships between a couple. Therefore, it was replaced by "se interessa em estar com outras crianças."

In question 3, "Seu filho gosta de subir em coisas como escadas?" was added to the example "móveis," with the aim of not restricting the question, whose purpose is to evaluate whether the child has the behavior of climbing things, independent of what they are. Such question was not understood by 5% of the population.

Questions 6 and 7 had the same problem in the translation. By asking if "Seu filho já usou o seu dedo indicador para apontar..." the question was ambiguous, as it could be understood as using the index finger of the child, mother or father. The solution was to add the pronoun "dele" after the word "indicador" in both questions.

Question 8, "Seu filho consegue brincar corretamente com brinquedos pequenos (ex. carros ou tijolos) sem simplesmente ficar colocando na boca, remexendo no brinquedo ou jogando-o para o alto e soltando?" was not understood by 10% of the target population. Those who did not understand it had no specific complaint, they simply could not provide an example to the question due to complete lack of understanding. Analysis of raters showed that the phrase was too long, using gerund and changing the order of the pronoun, which might be elements making comprehension difficult by low-schooling individuals. The following changes were performed: "corretamente" was replaced by "de forma correta;" "simplesmente" by "apenas;" "ficar colocando" by "colocar;" "remexendo" by "remexer;" and "jogando-o para o alto e soltando" by "jogar para cima e soltar."

Question 16, "Seu filho anda?," was replaced by "Seu filho já sabe andar?," as it was not widely accepted, leaving 5% of the target-population surprised, thinking that there was a second meaning, which should be inferred. Such change was made to make it as objective as possible.

In question 18, "O seu filho faz movimentos incomuns com os dedos perto da face dele?," raters decided to replace "incomuns" by "estranhos" and "da face" by "do rosto" to make the language more colloquial, as it was not understood by 20% of the population.

According to the raters, question 22, "O seu filho às vezes fica encarando o nada ou vagando sem direção?," did not express exactly what it intended to. Thus, they chose to add the term "aéreo" and replace "encarando o nada ou vagando sem direção" by "olhando sem uma direção definida."

Question 23 was changed with the aim of making it simpler. "O seu filho olha para o seu rosto para conferir a sua reação quando está diante de algo que não lhe é familiar?" was replaced by "O seu filho olha para você para conferir a sua reação quando vê algo estranho?"

Expert evaluation

Advice was asked from two experts in infantile autism. After their suggestions, the following changes were made:

Question 2 was changed again because the experts disagreed with the change described in the previous step. They considered "Seu filho tem interesse por outras crianças?" as the ideal format, since "se interessa em estar" restricted the question to physical presence.

Question 4, "Seu filho gosta de brincar de esconder o rosto e depois mostrar/esconde-esconde?," was replaced by "Seu filho gosta de brincar de esconder e mostrar o rosto ou de esconde-esconde?," as they claimed that "esconde-esconde [hide and seek]" is not the same game as hiding and showing one's face.

They also suggested that in question 8, "Seu filho consegue brincar de forma correta com brinquedos pequenos (ex. carros ou tijolos), sem apenas colocar na boca, remexer no brinquedo ou jogar para cima e soltar?," the word "tijolos [bricks]" shoould be replaced by "blocos [blocks]" because in Brazil brick is a type of construction material, and not a toy, opposed to blocks. They also suggested replacing "jogar para cima e soltar" by "deixar o brinquedo cair" as it represented more faithfully how children play.

Question 9, "O seu filho alguma vez trouxe objetos para você (pais) para lhe mostrar alguma coisa?," was replaced by "O seu filho alguma vez trouxe objetos para vocês (pais) para lhe mostrar esse objeto?" to make it more objective.

In question 22, they considered that the autistic behavior could be better described as follows: "O seu filho às vezes fica aéreo, 'olhando para o nada' ou caminhando sem direção definida?," instead of "O seu filho às vezes fica aéreo, olhando sem uma direção definida?"

In question 23, they preferred to preserve the original form "O seu filho olha para o seu rosto para conferir a sua reação quando vê algo estranho?" instead of "O seu filho olha para você para conferir a sua reação quando vê algo estranho?" They argued in favor of the importance of preserving the phrase "seu rosto [his face]" because it deals with eye contact, which is altered in autistic individuals.

Second pretest

Based on the changes, a preliminary version was developed, which was submitted to a second pretest. Interviewed individuals showed good acceptance of the scale, with good understanding by the target population; 70% of the population understood the instrument with no complaints. This time 30% of the population was comprised of individuals that concluded high school, 30% had incomplete high school, and 40% had incomplete elementary school.

Development of the final version

Only questions 1, 11 and 14 were changed.

In question 1, "Seu filho gosta de se balançar, de saltar no seu joelho, etc.?," the term "saltar [leap]" was replaced by "pular [jump]" as it is more used in the population;

As to question 11, "O seu filho já pareceu hipersensível ao barulho (ex. tapando os ouvidos)?," since 10% of the population did not know the meaning of "hipersensível [hypersensitive]," raters decided to replace it for "muito sensível [very sensitive]."

Question 14: "O seu filho responde quando você o chama pelo nome?" The pronoun "o" used before the verb was creating difficulties. It was then changed to "chama ele pelo nome," making it more colloquial.

The final version of the M-CHAT scale is included in the appendix.

Discussion

This study fulfills its objective, making the first version of the M-CHAT available for Brazilian Portuguese. Thus, this is an important contribution to enable early diagnosis of autism in the Brazilian context. Although the M-CHAT scale has no diagnostic value,¹¹ it is able to select suspicious cases for further assessment. Sociocultural factors are present in any individual that manifests psychiatric signs and symptoms. Culture emerges from a quite heterogeneous group of elements that include, among others, language, ethnicity, religion, traditions, beliefs, values, interpersonal relationships, and modes of production and social organization.¹⁶ Therefore, valuing cross-cultural adaptation is clearly important during translation of scales in the psychiatric area. There is no consensus on how to systematize translation of a scale. Strategies range from a simple translation of the original scale by the researchers until more detailed processes.¹⁷ The chosen model was based on the model by Reichenheim & Moraes,¹⁵ whose proposal is that the cross-cultural evaluation process should take place in a series of steps to facilitate a perfect functional equivalence. The authors stress appraisal of conceptual and item equivalence, of semantic equivalence, of operational equivalence and of measurement equivalence.¹⁵

Conceptual and item equivalence aims at verifying pertinence of the domains comprehended by the original scale within the new context to which it is being translated.¹⁵ In the translation of the M-CHAT, it was not necessary to discuss pertinence of items with a group of experts, since one of the researchers is a specialist in infantile autism and considered all items of the original scale relevant to the Brazilian context.

Semantic equivalence plays a major role, as it reflects correspondence of concept meaning between the original instrument and the translation, causing a similar effect in the population in both cultures.^12,15 This study fulfills such concept, as it uses translation and back-translation, performed by different professionals fluent in Portuguese and English. Appraisal of denotative or reference and connotative or general meanings was also performed, as well as testing in the target population.

This study does not strictly follow the steps recommended by Reichenheim & Moraes regarding semantic equivalence. The authors chose to distinctly evaluate two of the most important meanings of words and phrases - reference and general meanings. Justification is that subtle imprecision can be perceived in reference equivalence during the steps of translation and back-translation, especially because only one translation was performed. Such argument is corroborated by the results, since the external rater considered two questions between original and back-translation as "different." Thus, proper changes were made, ensuring literal correspondence for each item of the M-CHAT in Portuguese.

Two pretests were also performed. The first aimed at obtaining a preview for further evaluation of general equivalence. Thus, denotative meanings might be more carefully discussed, as they had practical background from the target population. Question 16, for example, was changed although being considered equivalent by the authors, as it did not have large acceptability during the first pretest. In this context, researchers considered application of a second pretest pertinent, with the aim of ensuring that the changes performed so far could be enough. After the importance of double testing the population was confirmed, small changes were required after the second pretest. The possibility of performing more tests in the population was suggested; however, it was discarded based on the results obtained by Perneger et al.,¹⁸ who, in a comparative study, concluded that application of two pretests is comparable with more extenuating translations.

Translation proved to be operationally equivalent, as the format of questions and the means used (printed material) were not criticized during the pretest and met the researchers' expectations, as they did not cause impatience in the target population. The administration setting - waiting room of pediatric offices - corresponds to what will be used in clinical practice. Only application form can be criticized because during the pretests the interviewer read the questions, which are self-administered, along with the interviewee. However, this could not have been otherwise, since it was necessary to verify the examples provided for each item to ensure whether understanding had been achieved.

Reduced number of translations was compensated by a detailed verification of reference evaluation. The sample chosen is representative of the target population of the M-CHAT (parents of children in waiting rooms of pediatric offices), which ensures more reliability to the results. The experts that evaluated this study corresponded to the suggestions from the literature¹⁹ as to interdisciplinarity: one neuropediatrician and one psychologist.

The authors propose further psychometric evaluation to verify measurement equivalence. It is of great interest to analyze the domain of the population included in the study, identifying whether the findings may be extended to other populations.

The importance of cultural approximation within different regions in Brazil should also be stressed. Considering that Brazil is a country with heterogeneous cultural roots, terms used in this translation might not be pertinent in other regions. Therefore, the authors invite critics and suggestions that may enhance the translation of the M-CHAT scale into Brazilian Portuguese.

The present translation of the M-CHAT was sent to the author of the original instrument, Dr. Diana Robins and considered adequate by psychiatry, psychology and neuropediatrics professionals and understandable by the patients in the target population sample. Therefore, there is an available version of M-CHAT developed into Brazilian Portuguese, observing the aspects of semantic, operational, conceptual, and item equivalence.

References

• ¹
  World Health Organization (WHO). International Classification of Diseases: diagnostic criteria for research, 10th ed. Geneva: WHO; 1993.
• ²
  American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 4th ed (DSM-IV). Washington: American Psychiatric Association; 1994.
• 3. Gray KM, Tonge BJ. Are there early features of autism in infants and preschool children? J Paediatr Child Health. 2001;37(3):221-6.
• 4. Klin A. [Autism and Asperger syndrome: an overview]. Rev Bras Psiquiatr. 2006;28 Suppl 1:S3-11.
• 5. Gray KM, Tonge BJ. Screening for autism in infants and preschool children with developmental delay. Aust N Z J Psychiatry. 2005;39(5):378-86.
• 6. Schopler E, Reichler RJ, Renner B. The Childhood Autism Rating Scale (CARS). Los Angeles: Western Psychological Services; 1988.
• 7. Berument SK, Rutter M, Lord C, Pickles A, Bailey A. Autism screening questionnaire: diagnostic validity. Br J Psychiatry. 1999;175:444-51.
• 8. Stone WL, Coonrod EE, Ousley OY. Brief report: screening tool for autism in two-year-olds (STAT): development and preliminary data. J Autism Dev Disord. 2000;30(6):607-12.
• 9. Brereton AV, Tonge BJ, Mackinnon AJ, Einfeld SL. Screening young people for autism with the developmental behaviour checklist. J Am Acad Child Adolesc Psychiatry. 2002;41(11):1369-75.
• 10. Baron-Cohen S, Wheelwright S, Cox A, Baird G, Charman T, Swettenham J, et al. The early identification of autism by the Checklist for Autism in Toddlers (CHAT). J R Soc Med. 2000;93(10):521-5.
• 11. Robins DL, Fein D, Barton ML, Green JA. The Modified Checklist for Autism in Toddlers: an initial study investigating the early detection of autism and pervasive developmental disorders. J Autism Dev Disord. 2001;31(2):131-44.
• 12. Mengyuan F. Cross-cultural Communication and Translation. US-China Foreign Language. 2004;2(6):16-22.
• 13. Baird G, Charman T, Baron-Cohen S, Cox A, Swettenham J, Wheelwright S, et al. A screening instrument for autism at 18 months of age: a 6-year follow-up study. J Am Acad Child Adolesc Psychiatry. 2000;39(6)694-702.
• 14. Georg AE, Duncan BB, Toscano CM, Schmidt MI, Mengue S, Duarte C, et al. Análise econômica de programa para rastreamento do diabetes mellitus no Brasil. Rev Saude Publ. 2005;39(3):452-60.
• 15. Reichenheim ME, Moraes CL. Operacionalização de adaptação transcultural de instrumentos de aferição usados em epidemiologia. Rev Saude Publica. 2007;41(4):665-73.
• 16. Jorge MR. Adaptação transcultural de instrumentos de pesquisa em saúde mental. Rev Psiquiatr Clin. 1998;25(5):233-9.
• 17. Moraes CL, Hasselman MH, Reichenheim ME. Portuguese-language cross-cultural adaptation of the Revised Conflict Tactics Scales (CTS2), an instrument used to identify violence in couples. Cad Saude Publica. 2002;18(1):163-176.
• 18. Perneger Tv, Leplège A, Etter JF. Cross-cultural adaptation of a psychometric instrument: two methods compared. J Clin Epidemiol. 1999;52(11):1037-46.
• 19. Guillemin F. Cross-cultural adaptation and validation of health measures. Scand J Rheumatol. 1995;24(2):61-3.

Publication Dates

• Publication in this collection
  17 Mar 2009
• Date of issue
  Dec 2008

History

• Accepted
  07 May 2008
• Received
  02 Apr 2008