Development and validation of an interpretable measure of dietary intake of whole plant foods

We aimed to develop an interpretable measure of dietary intake of whole plant foods (WPF‐ whole grains, vegetables, fruit, legumes, nuts, seeds). Two variables were calculated as the number of cup/ounce equivalents of WPF per 1000 kcal (WPF/kcal) and per kg (WPF/kg) using NHANES 1999–2002 data for persons 2–80 y. We examined the proportion of variance explained by WPF density measures in truncated (scoring standards for min/max levels) and non‐truncated (no min/max) Healthy Eating Index‐2005 scores (HEI‐2005‐T, HEI‐2005‐NT), separately for children (< 20 y) and adults (≥ 20 y). Correlations of diet quality measures and log serum carotenoids (SC) were explored using NHANES ‘01‐’02 data. Mean ± SD WPF/kcal was 1.2 ± 0.9 for children and 1.8 ± 1.4 for adults. Mean WPF/kg was 1.4 ± 1.1 for children and 1.6 ± 1.4 for adults. WPF/kcal explained 18% and 39% of the variance of child HEI‐2005‐T and HEI‐2005‐NT, respectively, and 25% and 51% of adult HEI‐2005‐T and HEI‐2005‐NT. The proportion of variance in SC explained for adults was 10 % for HEI‐2005‐T, 15% for HEI‐2005‐NT and 14% for WPF/kcal; values for children were 1%, 3% and 4%. Results for WPF/kg were similar to those for WPF/kcal, but smaller in magnitude. All relations with SC were significant (p < 0.01) after controlling for covariates. WPF/kg and WPF/kcal are easily interpretable measures that are strongly related to HEI‐2005 and SC, and may be of public health utility.