Food Pattern Modeling as an Alternative Assessment Method to Multiday Dietary Recalls for Iron-Related Nutrients: A Proof-of-Concept Study.

Abstract

There are barriers to in-depth memory-based dietary assessment techniques in community-based research. Food pattern modeling may be an alternative method to traditional assessment techniques. The objective of this study was to pilot a comparison of food pattern modeling to 24 h diet recalls for predicting hematological outcomes of iron status. Data from 3–24 h dietary recalls in 27 women were analyzed by two methods: mean dietary intake estimates or food pattern modeling. Food pattern modeling was used to determine the total inventory of foods consumed with iron, phytate, or ascorbic acid or iron–phytate ratios. Each variable was analyzed for its relationship to hemoglobin, ferritin, and acute iron absorption from a meal challenge study by creating receiver operating characteristic (ROC) curves. There were no differences in ROC curves or diagnostic accuracies between food pattern modeling or mean dietary intake estimates for iron, vitamin C, phytate, or phytate–iron ratios for estimating hemoglobin or ferritin values (p > 0.05). Food pattern modeling was inferior to mean dietary estimates for acute iron absorption, suggesting that more detailed methods may be necessary for studies with sensitive or acute dietary measurement outcomes. Food pattern modeling for total iron, vitamin C, phytate, and phytate–iron ratios may be comparable to detailed memory-based recalls for larger studies assessing the impact of foods on iron status.