DIGEST: Dietary Intake General Estimation and Statistical Toolkit

DIGEST is a free, open-source R toolkit under development at the UC Davis Institute for Global Nutrition, in collaboration with the International Food Policy Research Institute (IFPRI), that makes advanced analyses of usual dietary intake accessible to researchers, program implementers, and policy makers worldwide. Reliable estimates of usual (habitual) intake are essential for describing population diets, identifying groups at risk of nutrient inadequacy, and evaluating the impact of dietary interventions. Obtaining these estimates from 24-hour recall data requires removing day-to-day variation using the National Cancer Institute (NCI) method—a powerful approach that was, until recently, available only through costly SAS software and largely out of reach for any application beyond simple distributions without advanced statistical expertise.

Overhead view of five people brainstorming around a wooden table with laptops and sticky notes

Building on the NCI Biometry Group's 2025 release of the ncimultivar R package, DIGEST translates best-practice NCI methodology into a streamlined, three-function workflow. It enables users to (1) jointly model multiple dietary variables to derive composite outcomes such as Mean Probability of Adequacy and nutrient density; (2) estimate the prevalence of inadequate intake using cut-point, full-probability, and iron full-probability methods; (3) assess how non-dietary factors (e.g., sex, region, season, socioeconomic status) shape usual intake; and (4) rigorously evaluate intervention effects with ANCOVA-style pre-post comparisons. 

By removing technical and financial barriers, DIGEST opens the door to answering questions in nutrition research, policy, and advocacy that were previously out of reach for many teams. Development is underway, with vignettes covering descriptive distributions, composite outcomes, association testing, intervention effects, and regression calibration.

This work is part of the CGIAR Better Diets and Nutrition Science Program. We would like to thank all funders who supported this research through their contributions to the CGIAR Trust Fund.