Development of the Brazilian Portuguese version of the Achilles Tendon Total Rupture Score (ATRS BrP): a cross-cultural adaptation with reliability and construct validity evaluation

This is a cross sectional study approved by the Ethics and Research Committee of our institution (Hospital Mater Dei, Belo Horizonte, Minas Gerais, Brazil). Patients from three tertiary hospitals were included into this study. All were examined for the purposes of the present study in our centre (Orthopedic Department, Hospital Mater Dei, Belo Horizonte, Minas Gerais, Brazil). All patients gave written informed consent to participate in this study.

The ATRS is a self-reported questionnaire composed of ten items that reflect the opinion of patients regarding their symptoms, limitations in daily activities and physical activities after a complete rupture of the Achilles tendon. At the end of each question, the patient is given a scale from 0 to 10, where 0 corresponds to more symptoms and greater limitations of physical activity, and 10 indicates no symptoms and limitations. The final score is obtained by the sum of the responses for each item. The ATRS total score ranges from 0 to 100, with higher scores equal to less symptoms and limitations [10] (Additional file 1).

The cross-cultural adaptation of the ATRS was divided into 5 phases: initial translation (two bilingual Brazilian translators), synthesis of translation, back-translation (two native English language translators), consensus version and evaluation (expert committee), and testing phase [4] (Fig. 1). The first phase was the initial translation of the ATRS into the Brazilian-Portuguese, by two Brazilian bilingual translators, who produced two distinct versions, T1 version and T2. Only one of the translators had prior knowledge of the content of questionnaire [18].

The second stage was the synthesis of the translation, where the T1 and T2 versions were discussed with both translators, thus producing a single translation, the T1-2 version.

In the third phase, two bilingual translators, whose native language was English, translated separately the T1-2 version back into English, which results in the BT1 and BT2 versions. These two translators had no previous contact with the original questionnaire. This new translation into English allowed us to identify possible translation errors and grammatical inconsistencies, and was compared with the original version [19, 20].

The fourth phase consisted of the evaluation of all reports by an expert committee, which was composed by the authors of this work. The purpose of this phase was to review all translations to obtain a single version. For that purpose, the semantic equivalence, with regard to the meanings of words, with attention to idioms and colloquialisms, experimental equivalence, meaning to match activities of daily living in different countries and cultures and conceptual equivalence, ensuring the words have the same definition, were evaluated.

The final phase was the testing phase, where the final version of the Brazilian-Portuguese version of the ATRS was administered to 46 patients who suffered a total rupture of the Achilles tendon between the years 2007 and 2012 in 3 private hospitals in MG, Brazil. After answering the questionnaire, patients were asked about the difficulty of interpreting the questions or any other limitation in understanding the questionnaire.

This test-retest reproducibility of the Brazilian Portuguese version of the ATRS was conducted with a minimum interval of one week and a maximum interval of two months for each patient. The coefficient of Intra-Class Correlation (ICC) type 2,1 was used to determine the reproducibility of ATRS. The strength of agreement was classified using the following benchmarks: poor reproducibility for values below 0.40, good reproducibility for values between 0.40 and 0.75, and excellent reproducibility for those values above 0.75 [21]. We also calculated the standard error of the measurement (SEM) and the minimal detectable change (MDC) as additional measures. The SEM and MDC were calculated as follows: SEM = s√1 − ICC (where s is the standard deviation of the baseline) [21] and MDC = 1.96 × √2 × SEM, respectively [14, 22]. The SEM reflects the error of the instrument; and the MDC reflects the smallest within-person change in a score that, with p < 0.05, can be interpreted as a “real” change, above the measurement error.

The construct validity, which is the ability of the instrument to measure an abstract concept, [21] in this case, functional limitation, was determined by the Pearson coefficient between the ATRS and Orthopedic American Foot and Ankle Society (AOFAS) questionnaire for the hindfoot. The AOFAS questionnaire, already translated into Portuguese, was developed to evaluate the different anatomic regions of the foot, including the ankle and hindfoot. The questionnaire consists of nine items, divided into three categories: pain (40 points), functional aspects (50 points) and alignment (10 points), totaling 100 points [23]. In the analysis of construct validity, the Pearson correlation coefficient was calculated and its 95 % confidence interval estimated using bootstrapping (1000 samples). Pearson correlation coeficient were interpreted as follow: values between 0.00 and 0.25 indicate little or no correlation, between 0.25 and 0.50 indicate low correlation, between 0.50 and 0.75 a moderate correlation, above 0.75 indicate excellent correlation [24].

To characterize the sample, the mean (range) for age, median (interquartile range) for time of assessment after surgery, and frequency for gender and limb affected were calculated. Mean and standard deviation as well as median and interquartile range were reported for ATRS data. Sample size calculation followed the proposed method by Walter et al. [25] for reliability studies. Based on an expected ICC of 0.85, two measures per participant on two separate days and accepting at least moderate reliability coefficient (ICC = 0.7), a total of 46 participants was required. Wilcoxon paired test was used to compare test-retest data since the score presents in ordinal data. Independent t test and Chi square test were used to compare groups for age, gender and ATRS variables, respectively. IBM SPSS (IBM Corporation, Somers, NY) software 18.0 version was used for all statistical analyses.

The authors declare that the study received ethics approval. The raw data are available as Additional file 3.