Development and validation of a quantitative snack and beverage food frequency questionnaire for adolescents

• Published in: Journal of human nutrition and dietetics 2017-04, Vol.30 (2), p.141-150
• Main Authors: De Cock, N., Van Camp, J., Kolsteren, P., Lachat, C., Huybregts, L., Maes, L., Deforche, B., Verstraeten, R., Vangeel, J., Beullens, K., Eggermont, S., Van Lippevelde, W.
• Format: Article
• Language: English
• Subjects: Adolescent; adolescents; Beverages; Cross-Sectional Studies; Diet Records; Evaluation Studies as Topic; Female; FFQ; Humans; Male; Mental Recall; Nutrition research; Questionnaires; Reproducibility of Results; Snack foods; Snacks; Surveys and Questionnaires; Teenagers; validation; Validity
• ISSN: 0952-3871; 1365-277X
• DOI: 10.1111/jhn.12411

Background A short, reliable and valid tool to measure snack and beverage consumption in adolescents, taking into account the correct definitions, would benefit both epidemiological and intervention research. The present study aimed to develop a short quantitative beverage and snack food frequency questionnaire (FFQ) and to assess the reliability and validity of this FFQ against three 24‐h recalls. Methods Reliability was assessed by comparing estimates of the FFQ administered 14 days apart (FFQ1 and FFQ2) in a convenience sample of 179 adolescents [60.3% male; mean (SD) 14.7 (0.9) years]. Validity was assessed by comparing FFQ1 with three telephone‐administered 24‐h recalls in a convenience sample of 99 adolescents [52.5% male, mean (SD) 14.8 (0.9) years]. Reliability and validity were assessed using Bland–Altman plots, classification agreements and correlation coefficients for the amount and frequency of consumption of unhealthy snacks, healthy snacks, unhealthy beverages, healthy beverages, and for the healthy snack and beverage ratios. Results Small mean differences (FFQ1 versus FFQ2) were observed for reliability, ranking ability ranged from fair to substantial, and Spearman coefficients fell within normal ranges. For the validity, mean differences (FFQ1 versus recalls) were small for beverage intake but large for snack intake, except for the healthy snack ratio. Ranking ability ranged from slightly to moderate, and Spearman coefficients fell within normal ranges. Conclusions Reliability and validity of the FFQ for all outcomes were found to be acceptable at a group level for epidemiological purposes, whereas for intervention purposes only the healthy snack and beverage ratios were found to be acceptable at a group level.