Usual Dietary Intakes: SAS Macros for Running the NCI Method

NCI has developed two macros for estimating usual intakes. One -- MIXTRAN -- fits the model to obtain parameter estimates and allows for the evaluation of covariate effects. The other -- DISTRIB -- uses parameter estimates from MIXTRAN and a Monte Carlo method to estimate distributions of usual intake. These macros can be used jointly to estimate the distribution of usual intakes of both foods and nutrients. The MIXTRAN macro alone is sufficient for testing covariate effects on either food or nutrient intakes. The DISTRIB Macro is not intended to be used alone.

• MIXTRAN Macro
• DISTRIB Macro

Documentation for both macros is provided in the NCI Users' Guide for Analysis of Usual Intake Distributions (PDF).

To help analysts get started, NCI has developed several sample programs. These programs employ the MIXTRAN and DISTRIB macros in conjunction with a preliminary analytic dataset containing data from the National Health and Nutrition Examination Survey (NHANES). The preliminary dataset is based on NHANES 2001-04 data, but includes the addition of balanced repeated replication (BRR) weights, imputed values for some of the MyPyramid equivalents data, and differences in a few variable names.

• Analytic dataset
• Details of dataset contents (PDF)
• Sample programs:
• Example 1 - Program and output to illustrate model fitting and estimation of the distribution of usual intake for a dietary component that is consumed nearly daily,
• Example 2 - Program and output to illustrate model fitting and estimation of the distribution of usual intake for a dietary component that is consumed episodically,
• Example 3 - Program and output to illustrate time-saving programming techniques and macro features for a more complicated situation.