Designing Optimal Breakfast for the United States Using Linear Programming and the NHANES 2011–2014 Database: A Study from the International Breakfast Research Initiative (IBRI)

Abstract

The quality of dietary patterns can be optimized using a mathematical technique known as linear programming (LP). LP methods have rarely been applied to individual meals. The present LP models optimized the breakfast meal for those participants in the nationally representative National Health and Nutrition Examination Survey 2011–2014 who ate breakfast (n = 11,565). The Nutrient Rich Food Index (NRF9.3) was a measure of diet quality. Breakfasts in the bottom tertile of NRF9.3 scores (T1) were LP-modeled to meet nutrient requirements without deviating too much from current eating habits. Separate LP models were run for children and for adults. The LP-modeled breakfasts resembled the existing ones in the top tertile of NRF9.3 scores (T3), but were more nutrient-rich. Favoring fruit, cereals, and dairy, the LP-modeled breakfasts had less meat, added sugars and fats, but more whole fruit and 100% juices, more whole grains, and more milk and yogurt. LP modeling methods can build on existing dietary patterns to construct food-based dietary guidelines and identify individual meals and/or snacks that need improvement.

Coffee Consumption among Adults in the United States by Demographic Variables and Purchase Location: Analyses of NHANES 2011–2016 Data

“…The present estimates of dietary nutrient density among coffee consumers and non-consumers provide some parallels as well as contrasts with tea consumption. Coffee consumption among adults in the 2011–2016 NHANES sample was associated with greater compliance with the 2015–2020 Dietary Guidelines for Americans, as assessed by higher HEI-2015 scores [ 25 ]. Each 3% reduction in total dietary energy from empty calories leads to a 2-point improvement in HEI-2015.…”

“…The second measure of diet quality was provided by the Nutrient-Rich Foods (NRF9.3) index [ 25 ]. The NRF9.3 score is based on two subscores: the positive nutrient rich (NR) subscore and the negative nutrients to limit (LIM) subscore.…”

SCHOOLTHY: Automatic Menu Planner for Healthy and Balanced School Meals

“…• Linear Programming and Mixed Integer Linear Programming: [19]- [22]. • Case-based and Rule-based Reasoning: [23]- [27].…”

“…Furthermore, the different optimisation methods applied to solve each formulation were also described in detail. As a particular example, Linear Programming was applied to generate breakfasts for children and adults in [22]. In this case, breakfasts need to meet energy and nutrient requirements at low cost, while minimising deviation between the current breakfasts of a given population and the breakfasts provided by Linear Programming.…”

“…The issue of what aspects to consider as problem objectives and what to consider as constraints may differ among authors; however, there is a clear tendency to consider the cost of the menu as an objective to be minimised [15]- [17], [30]. Other authors, who do not consider the cost of the menu, usually include user preferences for certain foods, the level of adequacy, the level of acceptance or the deviation between the current practice and the solutions attained, as the objective to be optimised [11], [13], [22], [35], [36].…”