Validity and reproducibility of an adolescent web-based food frequency questionnaire

OBJECTIVE: To assess the validity and reproducibility of a newly developed, cost-effective, and easy to administer, Web-based food frequency questionnaire (FFQ) for the dietary assessment of adolescents. DESIGN: Cross-sectional validation of a Web-based FFQ, which contained 15 food groups (water; soft drinks, including fruit juice and diet soft drinks; alcoholic beverages; soup; milk and milk products; fruit; sweet and savory snacks/fillings; sauces and fat spreads; breakfast cereals; bread; cheese, fish/eggs/meat; pasta/rice; vegetables; and potatoes). Reference data for validation were 3-day estimated food records. Analyses were done for all participants and for consumers only (consumer in both Web-based FFQ and food record). SUBJECTS/SETTING: A selected group of adolescents (12 to 18 years of age, n=104) from participating secondary schools in Ghent, Belgium. RESULTS: For all participants, the measured intakes of water, fruit, breakfast cereals, fish/eggs/meat, pasta/rice, and potatoes were not significantly different between the two methods. The Spearman correlation for all foods was on average 0.38 (ranging from 0.20 for pasta/rice to 0.64 for breakfast cereals). When analyzed for consumers only, the Web-based FFQ showed significantly lower estimates for soft drinks, sweet and savory snacks/fillings, sauces and fat spreads, cheese, pasta/rice, and vegetables. The average Spearman correlation for all foods collectively was 0.30. When reproducibility of the Web-based FFQ was assessed, the average Spearman correlation coefficient for all foods collectively was 0.62. CONCLUSIONS: Most questions on the 15 food groups had acceptable reproducibility, whereas validation only showed reliable intakes for water, fruit, bread, and fish/eggs/meat. This Web-based FFQ could be a start for development of dietary assessment methods in public health nutrition contexts.