Relative Validity and Reproducibility of a Short FoodFrequency Questionnaire to Assess Nutrient Intakesof New Zealand Adults.

There is no recent validated short food frequency questionnaire (FFQ) for use in NewZealand (NZ) adults. This study aimed to evaluate the relative validity and reproducibility of a shortFFQ in free-living NZ adults aged 30-59 years. A 57-item, semi-quantitative FFQ was developedand pre-tested. During a 12-month study period the FFQ was administrated twice with a 9-monthinterval between administrations. Four two-day diet records were collected at months 0, 3, 6, and 9and a blood sample was taken at month 9. Spearman correlations were used to evaluate the validityof the FFQ with the eight-day diet records and selected biomarkers. Cross-classification analysisand the Bland-Altman method were used to assess the agreement between the FFQ and the dietrecord. Reproducibility over nine months was assessed using intra-class correlations. A total of 132males and females completed both FFQs, the eight-day diet record, and provided a blood sample.The highest energy-adjusted correlation coefficients were observed for alcohol (0.81), cholesterol(0.61), and carbohydrate (0.61), with the lowest for sodium (0.29), thiamin (0.33), and niacinequivalents (0.34). More than three quarters of the participants were correctly classified into thesame or adjacent quartile for most nutrients, with a low proportion of participants being grosslymisclassified (< 10%). For most nutrients, the limits of agreement from the Bland-Altman analyseswere between 50% and 250%. A positive correlation was observed between dietary intakes andplasma biomarkers for all selected nutrients. The FFQ showed moderate to good reproducibility,with almost all reliability coefficients ranging from 0.60 to 0.80. This short FFQ was shown to validlyand reliably rank individuals by their habitual intake of most major nutrients, indicating that theFFQ will offer a time-efficient way to assess the nutrient intake of NZ adults in future research.