Validação do teste de avaliação do letramento em saúde (TALES), um questionário brasileiro de avaliação de letramento em saúde

Abstract

Health Literacy (HL) is the ability of the patient to read, obtain and process basic information about their health in order to make pertinent decisions about self-care and to empower themselves with their clinical condition. To do so, it must present good HL levels, since low levels are associated with negative clinical outcomes. In order to evaluate HL, it is essential that the aim of this work was to create and validate the first Brazilian Evaluation Instrument for Health Literacy - TALES - Health Assessment Test. Composed of 1400 individuals, among patients from the Brazilian Public Health System, students of Youth and Adult Education and graduation. A Primer was prepared, containing information about Chronic Kidney Disease using linguistic techniques of reading facilitation. Afterwards, a database of 63 items was elaborated, encompassing questions of numerical skills and reading comprehension distributed in three levels of literacy, based on the Reference Matrix of the National Indicator of Functional Literacy (INAF). The items were randomized, forming six instrument versions, each one containing 21 items. In the calibration, each version was answered by 200 individuals, totaling 400 answers for each item and 1200 respondents. We used, for construct and dimensionality evaluation, the Factorial Confirmatory Analysis and, for the psychometric properties of the items, the Rasch model of a parameter. Estimates were obtained by the Maximum Profile Likelihood Method to obtain unbiased estimates under conditions of completely random data loss. Once the optimal version, that is, the one with the best psychometric properties, the scores obtained by this version were compared with the Gold Standard, in this case, the INAF. In this stage, called criterion validation, the final version of TALES and INAF were applied in 200 patients. Cronbach's Alpha values and KMO indexes were always higher than 0.85, showing good internal consistency and suitability of the sample for all versions. Exploratory Factor Analysis and Confirmatory Factor Analysis corroborated the one-dimensionality of the instrument. A Rasch model of a parameter was adjusted, indicating that there are predominance of coefficients of negative difficulties, that is, items of low difficulty. The correlation between the INAF percentage and the TALES score was moderate (r = 0.72). However, when correlating the TALES score with the INAF proficiency, a strong correlation was observed if we evaluated individuals with incomplete elementary education (r = 0.79) and low income (r = 0.77), respectively. TALES proved to be a good HL screening tool, effectively identifying patients at potential risk for negative clinical outcomes.