Development and validation of the nutrient-rich foods index: a tool to measure nutritional quality of foods

J Nutr. 2009 Aug;139(8):1549-54. doi: 10.3945/jn.108.101360. Epub 2009 Jun 23.


Ranking and/or classifying foods based on their nutrient composition is known as nutrient profiling. Nutrition quality indices need to be tested and validated against quality of the total diet. A family of nutrient-rich foods (NRF) indices were validated against the Healthy Eating Index (HEI), an accepted measure of diet quality. All foods consumed by participants in NHANES 1999-2002 studies were scored using NRFn.3 (where n = 6-15) indices based on unweighted sums, means, and ratios of percent daily values (DV) for nutrients to encourage (n) and for nutrients to limit (LIM) (3). Individual food scores were calculated based on 100 kcal (418 kJ) and FDA serving sizes [reference amounts customarily consumed (RACC)]. Energy-weighted food-based scores per person were then regressed against HEI, adjusting for gender, age, and ethnicity. The measure of index performance was the percentage of variation in HEI (R2) explained by each NRF score. NRF indices based on both nutrients to encourage and LIM performed better than indices based on LIM only. Maximum variance in HEI was explained using 6 or 9 nutrients to encourage; index performance actually declined with the inclusion of additional vitamins and minerals. NRF indices based on 100 kcal (418 kJ) performed similarly to indices based on RACC. Algorithms based on sums or means of nutrient DV performed better than ratio-based scores. The NRF9.3 index, based on 9 nutrients to encourage and 3 LIM per RACC and per 100 kcal, explained the highest percentage of variation from HEI and could be readily expected to rank foods based on nutrient density.

Publication types

  • Research Support, Non-U.S. Gov't
  • Validation Study

MeSH terms

  • Algorithms
  • Diet / classification
  • Diet / standards*
  • Food / classification*
  • Health Behavior
  • Humans
  • Nutrition Surveys
  • Nutritive Value*
  • Reproducibility of Results