ABC | Volume 112, Nº1, January 2019

Original Article Santos et al Validation of CADE-Q II in portuguese Arq Bras Cardiol. 2019; 112(1):78-84 and continuous variables not normally distributed were expressed as median and interquartile range. Absolute and relative frequencies were used for categorical variables. The chi-square test was used to assess associations between categorical variables. Overall knowledge of patients was expressed as the median of the total CADE-QII score. Median scores obtained in each domain were also described. Translation, cultural adaptation and pretest Initial translation of CADE-QII was made by three independent translators, aware of the objectives and underlying concepts of the study. They were asked to detect ambiguities and unexpected meanings in the Brazilian version compared with the original one. Back translation was conducted by four translators unaware of the initial objectives of the study as well as of the original version of the instrument. A commission composed of five bilingual specialists reviewed all the versions and made necessary changes according to Brazilian culture. A final version was generated, and the clarity of the questions tested in 23 coronary patients. During translation and cultural adaptation processes, question 4 of the physical activity domain (“Three things that one can do to exercise safely outdoors in the winter are”) was adapted to the Brazilian cultural context; the change was related to the weather, the expression “hot and dry weather”, referring to summer season, were substituted for “winter” (“What one can do to exercise safely outdoors in hot and dry weather are”), which better reflects the reality of Brazil, a tropical country. No further changes were required. Mean time required to complete the CADE-QII among participants was 22.5 ± 3.5 minutes. Mean rates for clarity of the instrument was 7.0 ± 1.77. Regarding content validity, following the administration of the instrument and discussion between patients and researchers, it was concluded that the CADE-QII clearly describes the aim of the measurements, the target population, the concepts measured and the selection of the items. Psychometric validation Three hundred and seven patients that participated in CR programs completed the CADE-QII. Sociodemographic and clinical characteristics of the patients are described in Table 1. Of these patients, 228 were participants of CR in Florianopolis, and 77 in Belo Horizonte. Most patients were men (n = 200, 65.1%) and had low educational attainment (incomplete elementary school, n = 188, 61.2%). Mean age was 63.3 ± 10.4 years (minimum = 31 years old; maximum = 88 years old). For test retest reliability analysis, 49 patients were selected by convenience and asked to complete the questionnaire again, with an interval of 15 days between the evaluations. Among these patients, 24 participated in a private CR program, and 25 in a public one. The test retest reliability was assessed by the ICC of each item, and the results are described in Table 2. The following items did not meet the minimum standards – question 4 (“A heart attack occurs”) of the medical condition domain, question 4 of factor risk domain (“The first step towards controlling a risk factor, such as blood pressure or cholesterol, is”, question 7 of nutrition domain (“How many servings of fruits and vegetables should adults consume?”) and question 5 of psychosocial risk domain (“ Chronic stress is defined as”). These items were excluded from the Brazilian version of the CADE-QII. Thus, from the 31 items of the original version, 27 items composed the Brazilian CADE-QII in Portuguese, with a maximum score of 81 points. The internal consistency of the 27-item instrument was tested, with a Cronbach alpha coefficient of 0.78. Regarding criterion validity, as described in Table 1, patients with higher educational level (p < 0.001) and higher family income (p < 0.001) showed higher level of knowledge about the disease as compared with the other patients. Medians and interquartile ranges of the items and domains are described in Table 2. The median total score was 53 (14) points, corresponding to 65.4% of the possible total score. The highest scores were obtained for the items: “What is the best source of omega 3 fats in food?” (“What one can do to exercise safely outdoors in hot and dry the winter are” and “Which of the following describes your best option for reducing your risk from depression”. The lowest scores were observed for the items “The first step towards controlling a risk factor (such as blood pressure or cholesterol) is”, “How many servings of fruits and vegetables should adults consume?” and “The statin medications have a beneficial effect in the body by”. Domains with the highest and the lowest scores were “Exercise” and “Psychosocial Risk” domains, respectively. Discussion This study aimed to validate and adapt the CADE-QII in Brazilian Portuguese. During these processes, we followed strict standards, since adaptation of an instrument to be used in a country other than that in which it was developed may require more than simply semantic and idiomatic analyses. 18 The psychometric properties – content validity, test retest reliability, internal consistency and construct validity – were established, confirming the validity of the CADE-QII for the Brazilian population. Our results were consistent with those reported in the original validation, 17 particularly with respect to internal consistency (Cronbach alpha of 0.91 vs. 0.78), indicating an adequate correlation between the questionnaires’ items, both in the original and in the adapted version. 17 Nevertheless, the fact that the CADE-QII was validated in a multicentric study may have affected alpha’s value (not as high as in the original version). Another difference between the original and the adapted version was in the way the questionnaire was administered; while the adapted CADE-QII was administered by a questionnaire, in the version in English was self-administered. With respect to criterion validity, there was a positive association of the level of knowledge about the disease with educational attainment and family income, suggesting that socioeconomic factors may be determinants to knowledge in health, which is consistent with previous studies. 14,17,20 This is corroborated by the fact that, in the present study, there was a positive association between enrollment in a 80