Development, Validation, and Reproducibility of a Food Frequency Questionnaire (FFQ) for Adult Population in Brunei Darussalam

Introduction: The dietary intake of adults in Brunei Darussalam has not been thoroughly investigated via the food frequency questionnaire (FFQ). This study aimed to establish, validate and test the reproducibility of an FFQ to determine and evaluate the adult dietary intake in Brunei Darussalam. Methods: Participants were divided into; (1) development phase (50 adults); (2) validation phase (100 adults); (3) reproducibility phase (subsample of 40 adults). 3-day dietary record (3DR) was used in development phase to gather food items consumed. 2-day dietary record(2DR) was used as a reference method in validation. Nutrients analyze were energy, macronutrients, calcium, and vitamins A and C. Validation analysis were spearman correlation, mean difference, Bland-Altman plot, cohen’s kappa and cross-classification analyses. Reliability analyses were intra-class correlation coefficient (ICC), Cronbach’s alpha and cross classification to compare repeat FFQ administered with 1 to 3 months interval. Results: The FFQ consisted 138 food items from 12 food groups. In validation, mean difference between FFQ and 2DR exceeded 10%, spearman correlation were significantly strong (r=0.994 to r=0.107). Cross-classification demonstrated > 50% of nutrients were correctly and adjacently classified, Bland-Altman plots were relatively consistent for the two methods. Cohen’s kappa ranged from 0.081 to 0.040, demonstrating weak agreement. In reproducibility, ICC outcomes for FFQ1(0.463) and FFQ2(0.978), Cronbach’s alpha was > 0.8, and > 95% of the nutrients were correctly or adjacently classified. Conclusion: The FFQ designed in this study was valid due to the consistency with the dietary records; hence, a reliable tool to predict the dietary intake of Brunei Darussalam adults.