Identification and reproducibility of dietary patterns assessed with a FFQ among women planning pregnancy.

Objective: To identify a posteriori dietary patterns among women planning pregnancy and assess the reproducibility of these patterns in a subsample using two dietary assessment methods.
Design: A semi-quantitative FFQ was administered to women enrolled in the Singapore PREconception Study of long-Term maternal and child Outcomes study. Dietary patterns from the FFQ were identified using exploratory factor analysis (EFA). In a subsample of women (n 289), 3-d food diaries (3DFD) were also completed and analysed. Reproducibility of the identified patterns was assessed using confirmatory factor analysis (CFA) in the subsample, and goodness of fit of the CFA models was examined using several fit indices. Subsequently, EFA was conducted in the subsample and dietary patterns of the FFQ and the 3DFD were compared.
Setting: Singapore.
Participants: 1007 women planning pregnancy (18-45 years).
Results: Three dietary patterns were identified from the FFQ: the 'Fish, Poultry/Meat and Noodles' pattern was characterised by higher intakes of fish, poultry/meat and noodles in soup; 'Fast Food and Sweetened Beverages' pattern was characterised by higher intakes of fast food, sweetened beverages and fried snacks; 'Bread, Legumes and Dairy' pattern was characterised by higher intakes of buns/ethnic breads, nuts/legumes and dairy products. The comparative fit indices from the CFA models were 0·79 and 0·34 for the FFQ and 3DFD of the subsample, respectively. In the subsample, three similar patterns were identified in the FFQ while only two for the 3DFD.
Conclusions: Dietary patterns from the FFQ are reproducible within this cohort, providing a basis for future investigations on diet and health outcomes.