Sugar snack consumption in Ugandan schoolchildren: Validity and reliability of a food frequency questionnaire.

OBJECTIVES This study assessed the reproducibility and relative validity of an eight-item self-administered food frequency questionnaire (FFQ) on intake of sugared snacks in Ugandan schoolchildren. A 5-day precoded food behaviour checklist (FBC) was used as validation criteria. Sociodemographic correlates of a sum frequency sugar score were explored. METHODS The study was conducted in Kampala, Uganda, in 2004. Six hundred and fourteen schoolchildren (mean age 12.4 years) completed the FFQ on cakes/biscuits, chocolate, ice sticks, soft drinks, coffee, tea, sugared desserts and sweets/candies at school. They were examined clinically for dental caries. Forty students completed the FFQ twice, 1 week apart and 325 students completed the 5 day FBC at school. RESULTS The mean decayed, missing and filled tooth index score was 0.98 (SD 1.6, range 0-15). Reproducibility scores (Cohen's kappa) for the sugar items ranged from 0.17 (ice sticks) to 0.55 (biscuits). No differences were seen between the average intakes at test and retest. Higher intake was reported in FFQ than in FBC across all sugar items. Crude agreement between students reporting intake at least 3-5 times a week/less than three times a week ranged from 50% to 55% (e.g. biscuits, chocolate) to 87% (tea). Spearman's correlation coefficients ranged from 0.14 (desserts) to 0.27 (sweets). anova revealed significant increase (P = 0.001) in the mean FBC sum scores by increasing quartiles of the FFQ sum scores. The average sum FFQ sugar scores were higher in girls than in boys and higher in older than in younger students. CONCLUSION Fair reproducibility was established for the FFQ sugar items. The FFQ was acceptable in classifying individuals into broad categories of low and high sugar consumption.