A comparison of methods to assess changes in dietary patterns from pregnancy to 4 years post-partum obtained using principal components analysis

Abstract

Few studies have examined the longitudinal nature of dietary patterns obtained using principal components analysis (PCA); the methods used are inconsistent. This paper investigates the methodologies used to assess stability and changes in such patterns. Pregnant women recorded frequency of consumption of various food items as part of regular self-completed questionnaires in the Avon Longitudinal Study of Parents and Children. This was repeated when their children were 4 years of age; 8953 women provided data at both times. Dietary patterns were identified using PCA and component scores were calculated at each time point. Additional ‘applied’ scores were created at 4 years using the loadings obtained from the PCA on the pregnancy data. Correlations were similar for each component across the time points, though slightly larger using the applied method. The applied scores were considerably lower on average than those obtained from separate PCA at 4 years. Women's scores decreased on ‘health conscious’ and ‘confectionery’ components while ‘processed’ and ‘vegetarian’ scores both increased over the 4-year period. In contrast, applied scores were systematically lower for all components. When split into quintiles, weighted κ was slightly higher between pregnancy and applied 4-year scores compared to the separate scores. In this cohort it was felt that the ‘applied’ method to obtain scores at the second time point was inappropriate, primarily due to the differences in FFQ between the two time points. We recommend that future studies using such ‘applied’ scores compare them with cross-sectional scores and consider the implications of any differences.