Adaptação transcultral e qualidades psicométricas do recordatório de três dias de atividade física (three day physical activity recall) em escolares de 10 a 12 anos

Abstract

Introduction: The use of questionnaires to measure the level of physical activity and daily energy expenditure in children and adolescents is easy to apply and has a low cost in population studies. However, most of the instruments used in Brazil are not cross-culturally adapted and, therefore, their psychometric qualities have not been tested for the target population. Objective: To adapt, validate and test the reproducibility of the Three-Day Physical Activity Recall (3DPAR) in children and adolescents of municipal public schools in the city of São João Nepomuceno, MG, Brazil. Materials and methodology: Cross-sectional study with 148 randomly selected 10-12 year-old schoolchildren of both sexes from the municipal schools of São João Nepomuceno (MG, Brazil). The study was divided into two phases following the cultural adaptation proposed by Reichenheim Moraes (2007): Phase 1 - Stage of conceptual equivalence and items, as well as semantic equivalence; Phase 2 - Stage of the measurement of socioeconomic and anthropometric variables, and equivalence measurement. In the first phase the selection of the tool, search in the experts database and the target population, translation and back translation done by sworn and bilingual translators, as well as the cross-cultural adaptation of the 3DPAR were performed. In the second phase, the economic status, height, body mass, waist, hip and abdomen, triceps skin fold thickness, subscapular and medial calf circumferences were collected. Upon completion of the evaluation, electrodes and the Actiheart® were placed in the thorax. For the calculation of daily energy expenditure the compendium for children and adolescents proposed by Ridely was used. The daily energy expenditure estimated by Actiheart® was added to the basal metabolic rate calculated by Schoefield's equation (1995). The study was approved by the Department of Pediatrics, Ethics Committee of UFMG, and all of the patients signed an informed consent form. To verify the normality of the data the Kolmogorov-Smirnov test was used. For the statistical analysis the descriptive statistics was used in order to describe the level of physical activity and energy expenditure predicted by the 3DPAR and Actiheart. To check the internal consistency of 3DPAR Cronbach's Alfa was used. As for the analysis of the validation criterion Spearman's correlation was used between the 3DPAR scores and the estimated values of Actiheart®, and for the analysis of residual scores, Bland & Altman's test was utilized. Reproducibility was checked by the intraclass coefficient between the scores generated in the first application (Week 1) and the second application (Week 2) of the 3DPAR. In order to analyze the data the software Statistica® for Windows® version 10.0, Statasoft® was used. For all treatments the significance level of 5% was adopted. Results: In the adaptation process, no problems of semantic equivalence were found. There were some changes in the list of activities of the original instrument, which were reduced from 72 to 61 activities, and details were added (lying, sitting, standing) for some sedentary activities which could influence the calculation of energy expenditure. Reproducibility: The Alpha Cronbach of the first and second application were = .73 and = 0.83, considering the energy expenditure per hour (72 items). Considering the daily energy expenditure of the three days, the alpha was 0.89 for the first application and 0.83 for the second one. The intraclass coefficient (ICC) between the average energy expenditure of the three days of application and replication was 0.74 (single measures ICC 95% CI - 0.64 to 0.81, p = 0.000). The ICC for the energy expenditure of 3DPAR (application and replication) for Thursday (kcal/day) was 0.55 (single measures ICC 95% CI - .40 to 0.66, p = 0.000), Friday was 0.82 (single measure ICC 95% CI - 0.75 to 0.87, p = 0.000) and Saturday from 0.74 (single measures ICC 95% CI - 0.64 to 0.82, p = 0.000). Validity: The criterion validity was obtained by Spearman's correlation, with r = 0.57, r = 0.60 and r = 0.63 for Thursday, Friday and Saturday, respectively. The Bland-Altman plot indicates a mean error of approximately 330 kcal (95% CI ± 900 kcal). Construct validity for all anthropometric variables showed significant positive correlations with estimates of daily energy expenditure of 3DPAR for the three days. Conclusions: The 3DPAR was culturally adapted with changes in the activity list, but there was the inclusion of subcategories for some sedentary activities. Regarding the psychometric qualities, both reproducibility and validity were satisfactory, presenting good reproducibility for replication after a week and with moderate correlations with the criterion measure.