A cross-sectional comparison of the functionality of the short-form FFQ to a 3-day food intake record completed early in the second trimester of pregnancy.

Abstract

Using a behavioural intervention to target nutrition during pregnancy may be key in meeting recommendations for healthy eating. The aim was to assess the use of a short-term dietary intake measurement tool (3-day food intake record) to infer long-term habitual dietary intake during pregnancy (using a short-form FFQ). A convenience sample (n 90) between 12- and 18-weeks' gestation was recruited from a larger randomised controlled trial for cross-sectional analysis. Participants completed a forty-four-item FFQ and 3-day food intake record. Using the participant food intake record, the investigator blindly completed a second frequency questionnaire. The frequency questionnaires were scored using dietary quality scores (DQS) and compared. Aggregate data were evaluated using a Wilcoxon signed rank test, and individual-level data were evaluated using a Bland-Altman plot. No significant difference was observed in the scores (Z = -1·88, P = 0·06), with small effect size (r= 0·19). The Bland-Altman plot showed that comparing the DQS derived from the two different dietary assessments underestimated scores by a mean difference of 0·4 points (95 % limits of agreement: -3·50 to 4·26). The data points were evenly spread suggesting no systematic variation for over- or underestimation of scores. Minimal difference was observed between the functionality of the two assessment instruments. However, the food intake record can be completed by pregnant individuals to estimate short-term nutrient intake and then scored by the investigator to estimate long-term dietary quality. Combining these two instruments may best capture the most accurate representation of dietary habits over time.