In 2005, the USDA issued a remarkable report called  “What We Eat in America, NHANES 2001-2002” (also available here for 2005-2006). It contains tables for some two-dozen essential nutrients (Vitamin A, C, and other vitamins; zinc, iron and other minerals) and components in food (e.g.. carbohydrates, protein).

Each table reports mean nutrient intakes in different portions of the population: Children ages 1-3; Males 9-13; Males 19-30; Females 19-30; Females 19-30; Females 71+; and, All persons 1 or older.

For each nutrient, the table compares mean intakes in different population cohorts compared to the amount a person in the cohort should ingest to assure no health problems triggered by insufficient intake of the given nutrient. For some nutrients, the USDA compares intakes to the upper-limit of safe intakes.

Four different measures of adequate or safe intakes are used in “What We Eat in America” tables –

• EAR – Estimated Average Intakes;
• DRI – Dietary Reference Intakes;
• AI – Adequate Intakes; and
• UL – Tolerable Upper Intakes Level.

Each table reports the number of people in a cohort for which dietary intake data were available; and, the mean level of intake across all cohorts and the Standard Error (SE) of the calculated mean level of intake.

Then, mean intake levels in a population subgroup are reported at different points along the distribution of intake levels – 5%, 10%, 25%, 50%, 75%, 90%, and the 95 percentile.

A few other data points follow –

• Percent of people above the target intake level (EAR/DRI/AI/UL);
• Standard Error (SE) of the percentile above the target intake level;
• Target intake level; and
• Percent of people above the target intake level.

            Use in Nutrient Profiling Systems

The data tables in “What We Eat in America” are vital in deploying a nutrient  profiling system (NPS).  Such systems quantify the nutritional contribution of a serving or known quantity of a given food.  The score for a serving of a given food is the sum of a set of fractions, each fraction representing the person’s intake of nutrient x in the serving, relative to the adequate intake for that nutrient in a day.

A simple sum of these fractions across some two-dozen nutrients produces a NPS score that reflects the assumption that each nutrient is equally important in promoting good health.

But this is clearly not the case. For example, a majority of woman ages 19-30 years have grossly inadequate intakes of iron (especially for pregnant women), but adequate, or nearly adequate levels of protein intake.

Accordingly, “What We Eat in America” data can be used to increase or reduce the weight placed on a given nutrient within a NPS, in order to better reflect the relative need across all nutrients of higher or lower intakes. See the Nutritional Quality Index section for one method used within a NPS to make such an adjustment.