Relative validity of a food frequency questionnaire supported by images/Validade relativa de questionario de frequencia alimentar com suporte de imagens

• Published in: Ciência & saude coletiva 2017-02, Vol.22 (2), p.457
• Main Authors: Brito, Alessandra Page, Araujo, Marina Campos, Guimaraes, Celso Pereira, Pereira, Rosangela Alves
• Format: Article
• Language: Portuguese
• ISSN: 1413-8123
• DOI: 10.1590/1413-81232017222.26392015

The scope of this study was to evaluate the relative validity of a food frequency questionnaire (FFQ) designed for adolescents. The study involved a sample of 106 adolescents (57% girls) aged between 11 and 18. The FFQ was administered in one-on-one interviews with the support of the 93 photographs of food items. Overall averages of energy and nutrient intake of three 24-hour recalls (24h-R) were used as benchmarks. They were then compared with those estimated from the FFQ using the Spearman correlation coefficient, weighted kappa for the agreement between classifications into quartiles and Bland-Altman limits of agreement. The correlation coefficients ranged from 0.27 to 0.49 (mean = 0.43), the weighted kappa ranged from 0.28 to 0.51 (mean = 0.40), and the estimated Bland-Altman mean limits of agreement ranged from 109 to 159% (mean = 130%). Approximately 30% of the estimates were classified into the same quartile while about 90% were categorized into the same or adjacent quartiles. Supporting the application of the FFQ with photographs provided more favorable results than those observed for the questionnaire alone. The FFQ was considered appropriate to categorize adolescents according to gradients of dietary intake. Key words Food consumption, Questionnaire, Photographs, Adolescents, Validity O objetivo do presente estudo foi avaliar a validade relativa de questionario de frequencia alimentar (QFA), desenvolvido para adolescentes. O estudo envolveu amostra de 106 adolescentes (57% meninas) de 11 a 18 anos de idade. O QFA foi aplicado em entrevistas pessoais com o suporte de fotografias dos 93 itens alimentares. As medias deatenuadas da ingestao de energia e nutrientes de tres recordatorios de 24 horas (R24h) foram consideradas como referencia e foram comparadas com aquelas estimadas pelo QFA com o uso do coeficiente de correlacao de Spearman, kappa ponderado para a concordancia da categorizacao em quartis, e limites de concordancia de Bland -Altman. Os coeficientes de correlacao variaram de 0,27 a 0,49 (media = 0,43), o kappa ponderado variou de 0,28 a 0,51 (media = 0,40), e a concordancia media estimada variou de 109 a 159% (media = 130%). Aproximadamente 30% das estimativas foram classificados de forma exata em um mesmo quartil e cerca de 90% em quartis iguais ou adjacentes. Resultados obtidos com a aplicacao do QFA com o suporte de fotografias foram mais favoraveis do que os observados para o questionario isoladamente. O QFA foi considerado apropriado