Send to

Choose Destination
Public Health Nutr. 2003 Sep;6(6):581-8.

A Belgian study on the reliability and relative validity of the Health Behaviour in School-Aged Children food-frequency questionnaire.

Author information

Department of Public Health, Ghent University, University Hospital, Bloc A, 2nd Floor, De Pintelaan 185, B-9000 Ghent, Belgium.



In the Health Behaviour in School-Aged Children (HBSC) survey, the frequency of consumption of a limited number of food items--focusing on fibre, calcium and less healthy items--is queried using a 15-item food-frequency questionnaire (FFQ). The present study was conducted to assess the reliability and relative validity of the HBSC FFQ among school pupils in Belgium.


To assess the reliability of the FFQ, 207 pupils aged 11-12 years and 560 pupils aged 13-14 years completed the questionnaire twice, with a test-retest interval of 6 to 15 days. To assess the relative validity of the FFQ, in a first study data were collected as part of the Flemish HBSC 2000 survey: 7072 pupils (11-18 years) completed the FFQ and a 24-hour food behaviour checklist (FBC). In a second study, 101 pupils (11-12 years) completed the FFQ and a 7-day food diary (FD).


Reliability--weighted kappa values between test and retest ranged from 0.43 to 0.70, percentage agreement from 37 to 87%, and Spearman correlations from 0.52 to 0.82. Relative validity--comparison of the FBC with the percentage of respondents who should have consumed the food items on a random day, computed from the FFQ, showed good agreement between the FFQ and the FBC for most items. Only for cereals, diet soft drinks and other milk products were considerably higher food frequencies than expected found from the FBC. Comparison of the FFQ with the FD showed overestimation for all but three food items (cheese, soft drinks and chips). Spearman correlations ranged between 0.10 for crisps and 0.65 for semi-skimmed milk.


The HBSC FFQ is a reliable questionnaire that can be used for ranking subjects for most food items, although one must consider the overestimation when the FFQ is used for estimating prevalences.

[Indexed for MEDLINE]

Supplemental Content

Loading ...
Support Center