Cross-cultural adaptation of the Clear Communication Index to Brazilian Portuguese

ABSTRACT OBJECTIVE To perform a cross-cultural adaptation of the Clear Communication Index instrument from the Centers for Disease Control and Prevention (CDC-CCI) from English to Brazilian Portuguese. METHODS This study comprised initial discussion about the conceptual equivalence of the instrument by a committee formed by experts on health education. We performed translations, synthesis of translations, back-translations, revision by the committee, and linguistic revision. Semantic equivalence was obtained by analyzing the referential and general meaning of each item by the committee, resulting in a pre-final version of the instrument. Subsequently, thirty professionals with health sciences degrees performed a pre-test. These professionals used the pre-final version of the instrument to assess a health education material. A questionnaire was applied to evaluate the acceptability of the instrument, the understanding of each of the 20 items, as well as the individual and professional variables. We analyzed the scores attributed to the health education material, the variables related to healthcare professionals, the proportions of the acceptability of the instrument, and the comprehension of each item. RESULTS After we obtained the conceptual equivalence of the instrument, the committee of experts, the instrument’s main author, and the linguist produced the pre-final version using two translations, a synthesis of the translations, and two back-translations. A general equivalence was maintained in 15 of the 20 items (75%), four of the items were slightly altered (20%), and one item was very altered (5%). Nineteen items presented referential equivalence or near equivalence (95%). We then carried out with the pre-test, in which the professionals used the pre-final version. Two items in the domains of “risks” and “main message” were unclear and needed to be revised. CONCLUSION The process of cross-cultural adaptation of the Clear Communication Index provided an adapted version to the Brazilian Portuguese language.


INTRODUCTION
We performed a cross-cultural adaptation of the CDC-CCI to help design and assess health messages and materials 9,10 . Four translators participated in the development of this research (two Brazilians and two Americans); a committee of experts consisting of Professors from the Schools of Dentistry and Pharmacy of the Universidade Federal de Minas Gerais (UFMG); a linguist; the main author of the original instrument, Cynthia Baur (CB), Professor at the University of Maryland, USA; and 30 primary healthcare professionals in public health, all volunteers, from a small city of the state of Minas Gerais, southeast Brazil. The sample size of 30 was similar to other studies in the cross-cultural adaptation of healthcare instruments [11][12][13] and consisted of a convenience sample of health professionals with higher education degrees who work at ten Primary Health Care Units in Minas Gerais. We approached these professionals at the Primary Health Care Unit where they work, and they provided written consent to participate in this project.
The CDC-CCI instrument, in its full version, consists of four introductory open-ended questions and 20 close-ended questions, with two answer options: "Yes" (score = 1) or "No" (score = 0), which the person who performs the scoring uses to evaluate the clarity and understanding of the information. The open-ended questions have no quantitative value, and each of the other 20 items is worth 1 point. Total scores vary from zero to 20 and are converted into a score on a scale of 0-100. The recommended minimum score is 90. The 20 questions encompass seven areas: "Main Message and Call to Action," "Language," "Information Design," "State of the Science (scientific knowledge)," "Behavioral Recommendations," "Numbers," and "Risk." 6 There is a short version called the "Modified Index," with 13 questions in the same seven areas described above. For this study, we used the full version.
After consulting the author responsible for validating the instrument (CB), the cross-cultural adaptation process followed the recommendations set forth in international literature, which include conceptual and semantic equivalences 9,10 .
Conceptual equivalence refers to the validity of concepts (domains) explored in the instrument being adapted and is obtained through feedback from the group who will use the instrument; in this case, experts in health education. This equivalence establishes whether or not the instrument can be understood and accepted in the new cultural context. The conceptual equivalence establishes if the measurability of the seven domains in both the adapted instrument and the original are similar 10 .
Semantic equivalence depicts the correspondence of the meaning or correct translation of items (terms and words) 9,10 by a committee of experts. The semantic equivalence is based on the comparison of the meanings between the original instrument and the backtranslations. Semantic equivalence can be evaluated from two aspects: a) the referential meaning that refers to similarities in meanings of items and can signal vocabulary or grammatical mistakes or discrepancies and b) the general meaning of each item that refers to the similarities of the ideas transmitted by the pairs of items.
The conceptual and semantic equivalence was obtained through translations, synthesis of the translations, back-translations, revision by a committee of experts in health, a linguist, and pre-tests ( Figure 1).
The committee of experts in health education assessed conceptual equivalence by checking the applicability, time of use, and capacity to measure (in its 20 questions in seven sections) the domain of "clarity of health materials" by the CDC-CCI in the Brazilian context. We discussed how different public and communication channels would use this instrument. In this moment, the researchers did not assess public health professionals.
Once the experts established conceptual equivalence, the subsequent step consisted of two translations from the original English to Brazilian Portuguese (T1 and T2), by two independent translators, both Brazilian with fluency in English. One of the translators had knowledge and practice in health, having familiarity with the terms and concepts present in the instrument. The other had no specific knowledge regarding the instrument's technical terms.
Four researchers, experts in health sciences, compared the two translations (T1 and T2), identifying discrepancies. This comparison generated a synthesis of the translations (T1 and T2), which aimed to identify possible difficulties in understanding the instrument. They compared the meaning of words in the different languages (English and Brazilian Portuguese) so that the same results were obtained in both translations.
Following the synthesis of the two translations (T1 and T2), two separate translators, native speakers from the United States with fluency in Brazilian Portuguese, performed independently two back-translations to English. The back-translators had no knowledge of the objectives of this work and did not have access to the original instrument. We sent the back-translations to the main author responsible for the validation of the original instrument (CB) in order to evaluate the quality of the translations and suggest modifications in the instrument. After this stage, the first version of the instrument was completed.
Subsequently, we conducted a review of the back-translations and a synthesis of the translations. Thus, a committee of experts composed of the same four researchers in health sciences, all four translators, a linguist, and two health professionals with experience in health research took the instrument in the original version as a reference. The establishment of a committee of experts was necessary for the achievement of a consensus regarding the conceptual and semantic equivalence of the items.
The committee of experts received the back-translation and original version of the CDC-CCI. For referential meaning, the committee evaluated these two versions without knowing which was the original and which had been back-translated 11,14 . A visual analog scale was used for referential meaning evaluation. The committee judged the equivalence of the pairs  of statements (original and back-translated) by consensus, with a scale from zero to 100% using the following categories: "non-equivalent" (< 80%), "near equivalent" (80−89%), and "equivalent" (90−100%).
For the general meaning evaluation, the committee used a scale from zero to 100%. They evaluated each pair of statements, having to reach a consensus. They classified them as: unaltered (UA), slightly altered (SA), very altered (VA), and completely altered (CA) 10,11,14,15 . In this step, the committee was aware of the two versions (the original and the back-translated). The scales used for the semantic equivalence can be seen in Figure 2.
Following semantic equivalence (referential and general meaning evaluation), the research team approached primary healthcare professionals at their places of work in Minas Gerais to perform a pre-test of the Brazilian Portuguese draft instrument (BR-CDC-CCI) in May and June 2018. The researchers asked the healthcare professionals to use the instrument to assess the health education material "Rational Use of Drugs" 16 from the Ministry of Health. The research team chose this health education material because it is publicly available, about a common topic, and includes features that the CDC-CCI is designed to evaluate. The professionals used the draft BR-CDC-CCI instrument to score the "Rational Use of Drugs" material. Each of the 30 professionals completed questionnaires on the acceptability of the instrument as a whole, the understanding of each of the 20 items, and individual and professional profiles. For acceptability, we included a general and dichotomous question ("Yes" or "No"): "Do you think that this instrument would be acceptable for Brazilian professionals?" Regarding the understanding of each item, we included the following question: "After you read the BR-CDC-CCI items and evaluation criteria, mark those that were understood ("Yes") and those that were not understood ("No"). For the items not understood, write in the corresponding space any problem with comprehension you had." Professionals also reported the amount of time spent answering the items of the instrument. In addition, the professionals answered questions about their age, sex, time since graduation, type of health sciences degree, public service time, and whether they had completed a graduate degree.
We analyzed the scores from the 30 primary care professionals using the Statistical Package for Social Sciences (SPSS for Windows, version 25.0, SPSS Inc., Chicago, IL). The variables related to the profiles and individual characteristics of the healthcare professionals were statistically analyzed by measuring the frequency and central tendency. The statistical analyses also included the calculation of the proportions of acceptability of the instrument and comprehension of each item from the BR-CDC-CCI.

RESULTS
We carried out the cross-cultural adaptation systematically. The first assessment by the experts showed that the applicability, time of use, and capacity to obtain the necessary domains of "clarity of health materials" could be obtained in the Brazilian version of the CDC-CCI. The group considered that the instrument would be used for healthcare professionals when creating and evaluating health information materials in a wide range of communication channels and for a diverse public. They considered the instrument necessary, practical, and useful in the Brazilian context and approved its Conceptual Equivalence.
We changed some sections of the English CDC-CCI in which cities, units of measurement, or U.S. public institutions were mentioned. Examples of this first step were changes of "Spring field " to "São Paulo" (item 18), "ounces" to "grams" (item 16), and "Public health organizations (...) (ASTHO)" to Brazilian Health Organizations such as "CONASS, CONASEMS." Other considerations included the exclusion of links to U.S. government agencies and institutions, where we inserted Brazilian equivalents. The author responsible for the validation of the original instrument (CB) also highlighted the need to maintain the term "primary audience" instead of "main audience" throughout the instrument. The idea of "primary" is for educators to focus on the most important audience who will use the information, even though other "secondary" audiences may also see the information if it is posted on a website, for example. Table shows the BR-CDC-CCI adaptation process, from the synthesis of translations to the adapted version, including the conceptual and semantic equivalence evaluations. For referential equivalence, nineteen items presented equivalence or near equivalence (95%), while only question 18 was "non-equivalent." In the general equivalence between the original instrument and the back-translation, 15 of the 20 items were unaltered (UA=75%); four of the items, questions 2, 3, 5 and 18, were slightly altered (SA=20%); and one item, question 1, was very altered (VA=5%).
The committee of experts highlighted the need for a review by a linguist, who evaluated the material and made recommendations. The main changes occurred in questions 18, considered no referential equivalent, and 1, which was very altered. The changes, as suggested by consensus, are presented in Table. In accordance with the assessment of the committee, question 1 presented a misunderstanding by not specifying what the "Main Message" would be in the context of the information, and for not defining what the educational materials would be. The committee observed linguistic and grammatical errors in questions 2, 3, 5, and 18 and did not approve them as maximum equivalence. Among these were the differences between the terms "section" and "session" in question 2; "highlighted" and "emphasized" in question 3; "calls to action for the primary audience" and "calls to action directed to the public" in question 5; and "web" and "internet" in question 18. We corrected the differences with the aid of a linguist.
The majority of the primary healthcare professionals involved in the pre-test were female (87%), with an average age of 36.8 years (range 24-49), average time since completion of undergraduate degree of 13 years (range 3-31), and 53% with a graduatelevel degree. Most of the professionals were nurses (57%), 20% were dentists, and 23% were healthcare professionals in other areas (Speech Therapist, Physician, Nutritionist, and Psychologist). In relation to professionals' understanding of the BR-CDC-CCI items, six professionals (20%) did not understand question 18 regarding the nature of the risk, and five (17%) did not understand question 1 regarding the main message. One (3%) professional did not understand question 5, regarding call to action, and question 14, regarding behavioral directions. All 30 professionals understood the rest of the questions (Table).
In view of these considerations, the committee and the linguist re-evaluated questions 1, 5, 14, and 18. Items 5 and 14 were considered confusing by one professional each, since they did not understand the meaning of the words "primary audience" and the example of the nutrition behavioral recommendation using folic acid. As items 1 and 18 continued to be a problem, the committee observed that the pre-final version should be clarified, and certain sentences shortened. The committee changed the order of some words and removed others, making the final Brazilian version clearer than the previous one (Table). Não se considera mensagem principal a apresentação de apenas um tópico, tal como "doença cardíaca" ou "gripe sazonal". Se o material contiver várias mensagens e nenhuma mensagem principal, responda não. NOTA: Se você respondeu Não para a questão 1, marque 0 para a questão 2-4 e siga para a questão 5.

Is the main message at the top, the beginning, or in the front of the material?
The main message should be in the first paragraph or section. A section is a block of text between headers. For a Web material, the first section must be fully visible without scrolling.

Is the main message highlighted with visual cues?
If the main message is highlighted through the use of fonts, colors, shapes, lines, arrows, or headings, such as "What you need to know," answer yes.

Is the main message emphasized with visual prompts?
If the main message is emphasized with font, color, shapes, lines, arrows, or titles, such as "What you need to know," answer yes.

Does the material include one or more calls to action for the primary audience?
If the material includes a specific behavioral recommendation, a stimulus for more information, a request to share information with another person, or a broad call for program or policy change, answer yes. If the call to action is for someone other than the primary audience, answer no.

Does the material include one or more calls to action for the main audience?
If the material includes a specific behavioral recommendation, a stimulus to obtain more information, a request to share information with another person, or a broad call to change the program or policy, answer Yes.
If the call to action is for someone that is not from the main audience, answer No.

Does the material include one or more calls to action for the primary audience?
If the material includes a specific behavioral recommendation, a prompt to get more information, a request to share information with someone else, or a broad call for program or policy change, answer yes.
If the call to action is for someone other than the primary audience, answer no.

Does the material use lists with bullets or numbers?
If the material has a list of more than 7 items, and the list isn't broken down into sub-lists, answer "No." If the list consists only of additional information or references, or is placed at the end of the material, answer "No."

Does the material use lists with markers or numbers?
If the material contains a list with more than 7 items, and the list is not divided into sub-lists, answer no. If the list is only of additional information or references, or if it is at the end of the material, answer no.

Does the material use bulleted or numbered lists?
If the material contains a list with more than 7 items, and the list is not broken up into sublists, answer no. If the list is for additional information or references only or at the end of the material, answer no.

Does the material include one or more behavioral recommendations for the main audience?
If not, STOP here and do not answer Part B.

Does the material explain why the behavioral recommendation(s) is (are) important for the primary audience?
If the material uses only numbers to explain the importance of behavioral recommendation without offering other relevant information to the audience, answer no.

Does the material explain why the behavioral recommendation(s) is important for the main audience?
If the material has only numbers to explain the importance of the behavioral recommendation without other relevant information for the public, answer No.

Does the material explain why the behavioral recommendation(s) is important?
If you offer only numbers to explain the importance of the behavioral recommendation with no other relevant information for the audience, answer no.

Do the behavioral recommendations include specific instructions on how to carry out the behavior?
This may include stepby-step instructions or a simple description (for example: Look for cereals that have 100% of the recommended daily amount of folic acid). If the material includes information about when and how to get in touch with a physician or other healthcare professional, answer "Yes." If the material mentions when and how often to carry out a behavior, answer "Yes."

Does the material always explain what the numbers mean?
For example, "the recommended amount of meat as part of a healthy meal is 3 to 4 ounces, which is similar to the size of a playing card."

Does the material always explain what the numbers mean?
For example, "the quantity of meat recommended as part of a healthy meal is from 3 to 4 ounceswhich is similar to the size of a playing card."

Does the material always explain what the numbers mean?
For example, "The amount of meat recommended as part of a healthy meal is 3 to 4 ounces-it will look about the same size as a deck of cards."

DISCUSSION
The CDC-CCI instrument aims to contribute to the improved performance of healthcare professionals who create educational materials, since it leads them to critically analyze their own communicative capacity 17,18 . The use of the CDC-CCI can improve the development and transmission of health messages, as well as the public's orientation regarding actions and better results in health 19 .
Low health literacy is considered a social determinant of health, with low literacy being a predictor of worse health outcomes 20,21 . Professionals' use of tools to adapt health messages and materials for low literacy audiences can contribute to improvements in the public's adherence to care and therapeutic outcomes, as well as to reduce social inequalities 22 . The BR-CDC-CCI, after its final validation, can offer healthcare professionals a practical resource, guiding them in the creation and evaluation of materials and educational messages in health, following the example of other studies in the literature 18,19,23 .
The adaptation of the CDC-CCI instrument is a crucial stage, since it provides an opportunity to test its feasibility in Brazil. This study obtained the results of the conceptual and semantic equivalence using robust methods that were used in other processes for adapting instruments from English to Brazilian Portuguese [11][12][13] . One adapted instrument must be equivalent to the source instrument in such a way that its meaning is the same for the majority of the desired population 10 in their different cultural and linguistic contexts. For this, original instruments and adaptation "must dialogue with each other" according to a team of judges. These judges should have the ability to understand whether or not the representation of the original instrument is similar to the representation in its final population, which, in this context, is a final population of healthcare professionals or others involved in the development of health education materials.
In this research, two items showed that divergences need to be rigorously analyzed, and the corrected versions should be included in the final format of an instrument. The misunderstanding generated by items 1 (main message) and 18 (nature of risk) made it difficult to evaluate the domains "Main Message" and "Risk" in the Brazilian context. These items were misunderstood in the process of obtaining equivalence (referential and general) and remained critical during the pre-test with 30 health professionals. Such misunderstanding may have two possible causes: semantic/syntax difficulty or the professionals did not consider these two items relevant for evaluation. The second reason could result in the exclusion of items for the Brazilian context, given that the "Main Message" and "Risk" domains would perform differently than the original instrument. As the problems detected were of syntax, our corrections allowed the two domains to follow the original instrument.
Following the evaluation by the committee, the pre-test version of question 1 remained the same as in the synthesis of the translations, but the pre-test with the 30 professionals showed they were confused about what the "Main Message" would be. Modifications to the order of words made the question clearer. In question 18, the word "damage," contained in the explanation of the question, was replaced by "risk," repeating the term already used in the question to reinforce the meaning in the original version of the CDC-CCI. We corrected the misunderstanding on what the instrument calls the "Main Message" and "Risk" during the adaptation process. These findings reinforce those found in other studies, highlighting the importance of the work of the committee of experts and pre-test in cross-cultural adaptations 11,12,14,15 .
This study also observed the importance of the participation of translators compatible with the criteria that the literature advocates. The independent translations and back-translations allowed us to locate the errors and discrepancies in ambiguous or unmatched items between the two languages. Translation by both a translator with health training and one without this training made it possible to detect a greater range of difficulties in understanding the instrument [11][12][13] . The author of the instrument observed additional discrepancies or misalignments not detected during translation, synthesis, or back-translations, giving greater credibility and fidelity to the initial proposal.
A future study will evaluate some psychometric properties of the BR-CDC-CCI, after we complete this cross-cultural adaptation. The adapted instrument still needs to undergo a process of evaluation in larger groups of professionals and materials. Despite the fact that the Brazilian National Health System has the same principles -such as comprehensiveness, universality, equitability-for the whole country, some cultural differences between regions and professional groups are likely to occur. Testing the BR-CDC-CCI with a larger number of health professionals located in different Brazilian regions and professionals with different training and experiences is necessary. These evaluation methods could allow the assessment of reliability and validity. We also will do a qualitative assessment to compare this instrument with others 6,17,18 . Other instruments that evaluate the quality of written clinical treatment choices could also be used when we evaluate and create health education materials 24,25 . In conclusion, the process of cross-cultural adaptation of the Clear Communication Index provided an adapted instrument to the Brazilian Portuguese language, which this is the first step in a longer process of testing and refining the BR-CDC-CCI for broad use among health professionals.