Reliability and Validity of Child/Adolescent Food Frequency Questionnaires That Assess Foods and/or Food Groups

Objective: Summarize the validity and reliability of child/adolescent food frequency questionnaires (FFQs) that assess food and/or food groups. Methods: We performed a systematic review of child/adolescent (618 years) FFQ studies published between January 2001 and December 2010 using MEDLINE, Cochrane Library, PsycINFO, and Google Scholar. Main inclusion criteria were peer reviewed, written in English, and reported reliability or validity of questionnaires that assessed intake of food/food groups. Studies were excluded that focused on diseased people or used a combined dietary assessment method. Two authors independently selected the articles and extracted questionnaire characteristics such as number of items, portion size information, time span, category intake frequencies, and method of administration. Validity and reliability coefficients were extracted and reported for food categories and averaged across food categories for each study. Results: Twenty-one studies were selected from 873, 18 included validity data, and 14 included test-retest reliability data. Publications were from the United States, Europe, Africa, Brazil, and the south Pacific. Validity correlations ranged from 0.01 to 0.80, and reliability correlations ranged from 0.05 to 0.88. The highest average validity correlations were obtained when the questionnaire did not assess portion size, measured a shorter time span (ie, previous day/week), was of medium length (ie, similar to 20-60 items), and was not administered to the child's parents. Conclusions: There are design and administration features of child/adolescent FFQs that should be considered to obtain reliable and valid estimates of dietary intake in this population.