Selected Intakes as Ratios of Energy Intake, U.S. Population, 2001-04

The NCI method provides the capability, for the first time, to estimate the distribution of usual intake of a nutrient expressed as a ratio of usual energy intake in the US population. This greatly enhances our ability to monitor diets relative to those recommendations that are expressed as ratios of energy intake and to assess the scope of dietary deficiencies and excesses in these cases. An example is the ratio of usual saturated fat intake (expressed in terms of energy value) to usual energy intake.

We have applied the NCI method to data from two recent cycles of the National Health and Nutrition Examination Survey (NHANES), a nationally representative sample, to estimate means and percentiles of the distributions of the ratio of usual food or nutrient intake to usual energy intake for a range of sex-age groups in the U.S. population.


Dietary data were obtained from the 2001-2004 NHANES. The data were collected via two 24-hour recalls. For the first seven dietary constituents examined (see Results), the analysis included 17,889 persons 1 year of age and older. The analysis of the percentage of energy from solid fats and added sugars included 17,311 persons 2 years of age and older. Further information regarding the design and methodology used in the 2001-04 NHANESExternal Web Site Policy is available from the CDC.

Intakes of macronutrients are available in the NHANES datasets. In addition, intakes reported on the recalls were translated into quantities of solid fats, alcoholic beverages, and added sugars using the MyPyramid Equivalents Database (MPED) 1.0, which was developed for the 2001-02 survey, and MPED 2.0 which was developed for the 2003-04 survey. This analysis does not account for the differences in data collection and processing procedures during the 2001-2004 time period.

An extension of the NCI method of estimating usual dietary intake distributions was used. The model postulates a joint bivariate model for the nutrient/food in question and energy intake. The method deals with ratios of usual intake of nutrients or foods that are consumed almost every day by everyone in the population. That is, this is an extension of NCI's method of estimating usual intake of nutrients, and episodically consumed foods are not considered here.

Each reported nutrient intake was first mathematically transformed to approximate normality using a Box-Cox (power) transformation. Dietary recalls tend to be different depending on whether they are the first or second report from an individual (the "sequence") and whether the reported day was a weekday or weekend. For each nutrient, the joint distribution of the Box-Cox transformed reported nutrient and energy was modeled, correcting for sequence and weekend/weekday effects and based on sex/age group. Means and percentiles of the distribution of the ratio of usual nutrient intake to usual energy intake were then estimated. Analyses were conducted for the entire population and for a range of sex-age subgroups. The method uses the survey sampling weights to adjust estimates for the complex sampling design and nonresponse.

Return to Top


Results from the NCI Method using 2001-2004 NHANES data are presented in the tables below. There are separate tables for each of the following ratios, all of which are expressed as percentages: total fat (kcal) to energy (%); saturated fat (kcal) to energy (%); polyunsaturated fat (kcal) to energy (%); and monounsaturated fat (kcal) to energy (%); solid fats, alcoholic beverages, and added sugars (kcal) to energy (%); and solid fats and added sugars (kcal) to energy (%). These tables show the mean, standard error of the mean, and percentiles of the distribution of intake for each percentage, by sex/age group. Additionally, a second table for each ratio provides percentile values and their standard errors.

Return to Top

Selected Findings

  • The 2005 Dietary Guidelines for Americans recommend that saturated fat intakes be less than 10 percent of calories. Over half the population consumed in excess of that, as the usual intake at the 50th percentile was 11% of energy for the entire population among most sex-age groups.
  • The Dietary Guidelines for Americans recommend that total fat intake be kept between 20 to 35 percent of calories. Most Americans' dietary fat intake fell within this range. The usual intake at the 5th percentile was 25% of energy for the total population, and the value 35% of energy fell between the 50th and 75th percentiles.
  • There is no explicit recommendation for the percentage of calories coming from solid fats, alcohol and added sugars (SoFAAS). These sources of energy count against the discretionary calorie allowance, as do intakes in excess of the recommended amounts of any of the food groups. Discretionary calorie allowances in MyPyramid range from 8% to 20% of energy. Even at the 5th percentile, all sex/age subgroups had SoFAAS intakes at or above 20% of energy. Median intakes were about 32% of energy, and across nearly all subgroups, at the tail end of the distribution, intakes of SoFAAS exceeded half the energy intake!

Return to Top


The following individuals represent the team who developed the NCI Method and produced this analysis:

  • Laurence S. Freedman1
  • Susan M. Krebs-Smith2
  • Douglas Midthune2
  • Patricia M. Guenther3
  • Dennis W. Buckman4
  • Raymond J. Carroll5
  • Victor Kipnis2
  • Amy F. Subar2
  • Janet A. Tooze6
  • Kevin W. Dodd2

1 Gertner Institute for Epidemiology
2 National Cancer Institute
3 USDA Center for Nutrition Policy and Promotion
4 Information Management Services, Inc.
5 Texas A&M University
6 Wake Forest University School of Medicine

Return to Top

Suggested Citation

Suggested citation for information contained on this page:

Selected Intakes as Ratios of Energy Intake, U.S. Population, 2001-04. Epidemiology and Genomics Research Program website. National Cancer Institute. Updated May 22, 2015. Accessed January 23, 2017.

Return to Top