ADAPTATION AND TRANSCULTURAL VALIDATION OF THE BRAZILIAN VERSION OF THE BACK PAIN FUNCTIONAL SCALE

Objective: The Back Pain Functional Scale (BPFS) was designed to evaluate the functional state of individuals with low back pain. The scale consists of twelve items, covering functional aspects of the daily life activities of these individuals. The final score is calculated by summing the responses to each item, the values of which range from 0 to 5, obtaining a total result of 0 to 60 points. Methods: The validation process was developed in accordance with the World Health Organization (WHO) protocol, covering translation, back translation, semantic equivalence, evaluation by specialists from previous stages, pre-test of the tool, and final version. Subsequently, the final version was applied in a sample of 90 individuals and the data obtained were subjected to descriptive statistical analysis, factorial analysis, evaluation of internal consistency, and correlation with other validated tools. Results: The tool was adapted to Brazilian Portuguese, making use of terms to approximate the language of everyday expressions. The final version presented results similar to those from the original version, as demonstrated by the factorial analysis, the internal consistency (Cronbach’s alpha: 0.990), and the strong correlation with tools validated for the Portuguese language. Conclusion: The Brazilian version of BPFS proved to be easy to apply and understand, and presented high internal consistency and construct validity similar to that of the original instrument. Level of evidence 1B; Study of adaptation of a valid score. validación fue desarrollado de acuerdo con el protocolo de la Organización Mundial de la Salud (OMS), abarcando traducción, retrotraducción, equivalencia semántica, evaluación de especialistas de las etapas anteriores, test previo del instrumento y versión final. A continuación, la versión final fue aplicada en una muestra de 90 individuos y los datos obtenidos se sometieron a análisis estadístico descriptivo, análisis factorial, evaluación de la consistencia interna y correlación con otros instrumentos validados. Resultados: Se realizó la adecuación del instrumento para el portugués utilizado en Brasil haciendo uso de términos para aproximar el lenguaje a las expresiones de la vida cotidiana. La versión final presentó resultados similares a la versión original, demostrados por el análisis factorial, por la consistencia interna (alfa de Cronbach: 0,990) y correlación fuerte con instrumentos validados para el idioma portugués. Conclusiones: La versión brasileña de la EFDL mostró tener fácil aplicación y comprensión, presentó alta consistencia interna y similar validez de constructo al instrumento original. Nivel de evidencia 1B; Estudio de adaptación de una puntuación válida. Descriptores: Dolor de la Región Lumbar; Estudio de Validación; Funcionalidad; Encuestas y Cuestionarios; Escalas.


INTRODUCTION
Low back pain or lumbalgia is a symptom, not a disease, and can result from various known or unknown abnormalities or diseases. It is defined by the location of the pain, typically between the lower margins of the ribs and the gluteal folds. The condition is often accompanied by pain in one or both lower limbs and some people have associated neurological symptoms. 1 It is a painful condition that can affect up to 65% of people each year and it is estimated that up to 84% of people will have low back pain at some point in their life. 2 It is more prevalent among women and in individuals between 40 and 80 years of age. 3 Recent data have shown that low back pain has generated an increase in the number of years that sufferers live with disability. 4 In Brazil the incidence of low back pain is known to be high (>50%), but the fragility of Brazilian studies may mean that the real numbers for this condition are underestimated. 5 The costs of managing low back pain are increasing, mainly in low-and medium-income countries that are overburdening health and social systems that are already overburdened. 1 Measures to prevent the onset of pain and the persistence of disability associated with low back pain demand an awareness that the disability is intimately tied to the social and economic context of those who suffer from it, as well as to personal and cultural beliefs. 6 In an effort to understand and study low back pain, questionnaires have been widely used as a tool in epidemiological, monitoring, and treatment efficacy analysis studies. 2 The first studies to use self-administered questionnaires to analyze low back pain were launched in the 1980s 7 and they have since been adapted for use in Brazilian Portuguese. [8][9][10] The Back Pain Functional Scale (BPFS) was conceived by Stratford et al. to be used to measure functional status and a tool to assist decision making, both individually and for groups of patients. The scale showed excellent results in relation to its descriptive analysis, internal consistency, reliability, validity, and sensitivity. 11 Administration and interpretation of the BPFS are extremely practical. It consists of 12 questions with 6 possible responses, each associated with a point value (Likert scale), with 0 being unable to perform the activity, 1 being able to perform the activity with extreme difficulty, 2 with quite a bit of difficulty, 3 with moderate difficulty, 4 with a little bit of difficulty, and 5 with no difficulty. The total score can range from 0 to 60 and the higher the score, the greater the functional capacity of the person being assessed. A score of 0 indicates that the individual is extremely unable to perform any activity and a score of 60 signifies the absence of difficulty in performing any activity.
The objective of this study was the adaptation and validation of the BPFS for Brazilian Portuguese.

METHODS
To initiate the study, the BPFS authors were asked to authorize its use and adaptation to Brazilian Portuguese, after which it was approved by the Institutional Review Board of the Universidade Paranaense as IRF Consolidated Opinion number 1.698.082.
The translation and back-translation protocol recommended by the World Health Organization (WHO) was used, which aims to make translation to languages other than English possible such that the translated versions retain conceptual equivalence in each of the countries and cultures. The tool should be natural, acceptable, and applicable in practically the same form as the original. 12 In the first stage, two health professionals and a language professor, whose native language is English and who are fluent in Brazilian Portuguese, independently performed three translations from English to Brazilian Portuguese. The translators were considered blind from each other. These professionals were selected for their familiarity with the constructs used by the tool. The professionals were asked to make an effort to maintain the semantic consistency of the terms and constructions used.
In stage 2, three bilingual Brazilian health professionals, independently and without knowledge of the original scale, performed back-translations of the first translations to English.
In stage 3, the authors of this study conducted an evaluation of semantic equivalence and the preliminary Brazilian Portuguese version was revised by consensus based on the translations and backtranslations, resulting in the version in Appendix 1. All the translators agreed with the conversion of one mile to 1600 meters (question 9) to facilitate understanding of the unit of distance.
In stage 4, this version was presented to 15 individuals (8 females and 7 males) ranging in age from 20 to 50 years old, 4 of whom had an elementary school education, 8 with a high school education, and 3 with a college education. In all cases, the interviewees were questioned about their comprehension of the questions and the clarity of the alternatives. All participants stated and demonstrated that they understood all the questions and that they were simple, objective, and easy to understand. The final version of the BPFS for Brazilian Portuguese was prepared from these considerations (Appendix 1).
After this process, the final version was used for 90 patients at the Toledo unit of the Clínica Escola de Fisioterapia da Universidade Paranaense. The selected participants were duly informed about the objective of the study and voluntarily signed the Informed Consent Form, agreeing to participate in it. The inclusion criteria were individuals of both sexes from 18 to 60 years of age who had been experiencing chronic specific or non-specific low back pain for more than 3 months.
Other studies using the BPFS 11,13 were consulted for sample calculation via factor analysis. The original study that validated the BPFS used a sample of 77 individuals (47 females and 30 males). 11 Samples of approximately 100 individuals are sufficient for conducting an exploratory factor analysis. 14 It is also noteworthy that the consensus in the literature is that the stability levels in exploratory analysis depend on the quality of the psychometric tool and not to the sample size. 15 The characteristics of the study sample are presented descriptively in Table 1.
Dimensionality was evaluated using exploratory factor analysis by the principal component analysis technique with varimax rotation. The number of factors was determined by examining the slope on the sedimentation graph. Reliability was analyzed using Cronbach's alpha coefficient. Because of the lack of a gold standard tool, two tools were used to evaluate construct validity between these and the BPFS by means of the Spearman's correlation coefficient. One of them was the Roland Morris Questionnaire, which consists of 24 questions related to common everyday situations in which patients may have difficulty performing due to the presence of low back pain. For each answer marked as true a point is added, such that the sum of the points can vary from 0 to 24. Thus, the greater the number of questions marked as true, the greater the disability of the patient. The validity and reproducibility of the translated, adapted Portuguese version are well-established in the literature. 9 The other tool used in this study was the Quebec Back Pain Disability Scale, 16 developed to measure functional disability resulting from low back pain and composed of 20 questions that describe the difficulty performing activities of low intensity. It is based on a six-point scale (0-5), where 0 is the absence of difficulty and 5 is the maximum inability to perform the activity. The final score ranges from 0 to 100 points and the higher the score, the worse the clinical condition. The tool is validated for Brazil. 17

RESULTS
The final version of the BPFS was applied to 90 individuals of both sexes (75 females and 15 males) with a mean age of 45.8±11.8 years and mean body mass index of 29.2±5.9 kg/m 2 , who volunteered to participate in the study. The overall BPFS score for the sample was 33.4 ±15 points. The male subjects had statistically higher scores than the females at 44.4±9 and 31.2±15 points, respectively (p>0.05, Mann-Whitney U test).
The mean values for each individual component are shown in Table 2.
In the simultaneous factor analysis of all the variables, the Kaiser-Meyer-Olkin (KMO) test yielded a value of 0.925, indicating that this analysis is perfectly adequate for the treatment of the data. Bartlett's sphericity test, based on the Chi-squared statistical distribution, showed a result of 2441.009 for a significance level less than 0.001, which allowed us to reject the null hypothesis that there was no correlation among the data analyzed. When analyzing the communalities, it was observed that all the questions of the tool had values greater than 0.50 by the extraction method. Therefore, it was not necessary to extract any component.
Initially the eigenvalues that represented the variability of each component and the percentage of variance explained by means of each one were identified, as can be seen in Table 3.
As a criterion for deciding the number of factors to be selected to represent the latent structure of the data, the latent root criterion, which selects only eigenvalues greater than 1, was initially considered. When considering this criterion, it was observed that a single factor corresponded to 91.39% of the total variability. This result is also satisfactory for the percentage of variance criterion, which suggests that a minimum explanation of 60% variability is sufficient. Also in Table 3, the factor loadings that represent the contribution of each variable to the formation of component 1, are presented. Varimax rotation was performed with Kaiser factor normalization to facilitate visualization of the representative factor loadings in each factor.
Cronbach's alpha test was used for internal BPFS consistency. All the correlations between the items were positive and significantly different from zero, which indicates that a scale can be built with these items, since they measure the same attribute. In the present study, the Cronbach's alpha value obtained was 0.990.
Finally, the validity of the construct was tested by means of Spearman's correlation coefficient between the BPFS and the Quebec Back Pain Disability Scale (r=-0.739) and the Roland Morris Questionnaire (r=-0.867), revealing in both cases a strong correlation between the tools (p<0.001).

DISCUSSION
The prevalence of low back pain has increased proportionally as the world population ages. 3 Globally, the scientific community is dedicated to the early and accurate identification of this dysfunction mainly in low-and middle-income countries. 18,1 In Brazil, the territorial expanse, the heterogeneity of the population, and the non-uniformity of studies make precise low back pain prevalence and incidence data analysis in the population difficult, 5 however, it is clear that this dysfunction has significantly impacted the healthcare and pension systems of the country. 19 In this regard, in recent years the costs of healthcare and lost productivity due to low back pain have been substantial with social/social security costs reaching US$2.2 billion. 20 Despite various initiatives to address low back pain as a public health issue, 21 specific strategies related to cost effectiveness and the management of this condition and its consequences still need to be identified. 1 Related specifically to low back pain, we highlight the use in Brazil of the Quebec Low Back Disability scale and the Roland Morris Questionnaire, both validated for the Portuguese language. 22 The present study offers the alternative of a tool with fewer questions that maintains the intended sensitivity and specificity. 11,23 The present study succeeded in developing a Brazilian Portuguese version of the BPFS, following the parameters for transcultural tool adaptation recommended by the WHO. 12 The BPFS proved to be a tool that is easy for individuals with low back pain at different educational levels to understand, useful for this purpose, and with a reduced application time.
The mean BPFS score of the study sample was 33.4 ±15 points. Similar results can be found in the original tool validation studies, which reported an initial evaluation value of 34.1±13 points, 11 as well as in a similar validation study where a value of 33.0±9.9 points was reported. 13 The adequacy of the factor analysis was confirmed by means of KMO and Bartlett's sphericity tests and made it possible to verify the suitability of the data to the application of factor analysis. In the present study, factor analysis proved to be adequate for the treatment of data with two factors, explaining the approximately 91.3% variance found. No other factor presented an eigenvalue greater than 1. The KMO test, which verifies the degree of intercorrelation among the variables, generated a value of 0.925, suggesting good adaptation of the data to factor analysis. The values obtained in this test varied from 0 to 1. The closer to 1, the more suitable the sample is to the application of factor analysis. 24 The internal consistency among the twelve BPFS questions was tested using Cronbach's alpha test, which generated a value equal to 0.99 indicating the high reliability of the responses attributed to the questions. Cronbach's alpha coefficient is one of the most used For each question the responder has the following alternatives on a scale from 0 to 5 points: Unable to perform the activity (0), Extreme difficulty (1); Quite a bit of difficulty (2); Moderate difficulty (3); A little bit of difficulty (4); No difficulty (5).