Adaptação transcultural: equivalência de mensuração da versão brasileira (Rio de Janeiro) do Child Feeding Questionnaire

Abstract

The parental feeding practices have been identified as important in the etiology of overweight and obesity. Although there are different ways of measuring parental feeding practices, the Child Feeding Questionnaire (CFQ) developed by Birch et al. (2001) is one of the most used, with a good psychometric history and several cross-cultural adaptations, being used in different countries. The CFQ has seven scales: Perceived responsibility (three items); Perceived parent weight (four items); Perceived child weight (six items); Concern about child weight (three items); Restriction (eight items); Pressure to eat (four items); Monitoring (three items); totaling 31 items. The objective of this dissertation was to evaluate the measurement equivalence step of the Portuguese (Brazil) version of the CFQ prepared by Ferreira et al. (2018). In the present study, reliability was evaluated using internal consistency (α’s Cronbach), inter-observer reliability was calculated using quadratic weighted kappa estimated separately for each item, with 95% confidence intervals and construct validity using the evaluation of the association between parental feeding practices and selected variables described in the literature as associated with such practices: birth weight, maternal age, maternal schooling and anthropometric variables such as weight/age, height/age and BMI/age. Associations were evaluated based on logistic regression analyzes (dichotomous outcomes) and linear regressions (continuous outcomes). Among the main characteristics of the sample studied, according to the Body Mass Index (BMI) 18.23% were obese. Regarding the internal consistency of the Brazilian version of the CFQ, most scales presented values above 0.6 (perceived responsibility, perception of the child's weight, concern about the child's weight, pressure to eat and monitoring). Regarding inter-observer reliability, weighted kappa values were above 0.859, being considered high. Regarding the construct validity, birth weight was inversely associated with the child's weight perception scale (r -0.79; p<0.05); maternal age was associated with a lower chance of pressure to eat (OR: 0.67; p< 0.05); the scales parental weight perception, child weight perception were associated with the BMI/age index (OR 2.27;15.86; p-value < 0.05). And the pressure to eat scale was associated with lower chances of being overweight (OR 0.44; p-value < 0.05). Our findings were satisfactory in relation to the original questionnaire, contributing to the process of cross-cultural adaptation of the questionnaire on parenting practices of infant feeding, for the identification and measurement of this aspect so important for the development of childhood overweight.