Validade relativa de métodos de avaliação de consumo alimentar de escolares

Abstract

Introduction: To evaluate of the food consumption of schoolchildren is relevantfor the public health field considering the changes in dietary patterns in recent decades and deleterious consequences to health in the short, medium and long term. However, such evaluation is challengingdue to the scarcity of specific instruments for this life cycle. Objective: To investigate the relative validity of two methods of food consumption evaluation for schoolchildren. Methods: Cross-sectional study with second-year students from the municipal school system of Belo Horizonte/MG. Socio-demographic and anthropometric data were collected. The Body Mass Index (weight/height2) was classified according to parameters of the World Health Organization. Food consumption was investigated by direct observation of three school meals (reference method) and application on two different days of two test methods: 24-hour dietary recall (24hDR) and Previous Day Food Questionnaire version 3 (PDFQ-3). Descriptive analysis, Shapiro Wilk test, Spearman correlation, agreement Kappa (nutrient or food adequacy) and sensitivity, specificity, omissions and intrusions analyzes were performed for each food item of the PDFQ-3. Results: A total of 41 schoolchildren answered the 24hDR, of which 59% were boys, with a median age of 8.35 (8.17-8.58), and 22% presented overweight. Median of energy, macro and micronutrient (calcium, iron and vitamin A) were similar betweenthe methods, correct by the nutritional status. Strong correlations were obtained for energy and macronutrients (r>0.67) and moderate to strong for micronutrients (r=0.52-0.73). Stronger correlations were identified among overweight schoolchildren. The Kappa coefficient resulted in perfect agreement for energy, good agreement for protein and iron, and moderate agreement for carbohydrate and lipid. For the PDFQ-3 validation, a total of 43 schoolchildren were evaluated, of which 62.8% were boys, with a median age of 7.9 years (7.70-8.15), and 23.2% presented overweight. Ten of the 21 PDFQ-3foods were consumed on the day of observation. Sensitivity values were higher than 84%, except for sweet potato (52.63%). The specificity varied from 15% (cookies) to 91% (sweet potato). There were low omission rates of 13.3% (3.33-14.64) and intrusion 19.05% (15.38-38.46). It was obtained a Kappa coefficient varied from very low (cookies and beans), moderate (sweet potato) to good (coffee with 14 milk, carrot, pasta, sausage and banana. Conclusion: The two test methods presented validity to evaluate the food consumption of schoolchildren. To improve the results of future investigations, it is suggested, for 24hDR, to increase the amount of applications to detect the consumption of micronutrients with greater intrapersonal variability, such as vitamin A; and for PDFQ-3the improvement of graphic design is recommended to promote better identification of the foods.