A Genetic Algorithm Approach to Optimize Planning of Food Fortification Programs

• Activity Number: 302
• Type: Contributed
• Date/Time: Tuesday, August 2, 2011 : 8:30 AM to 10:20 AM
• Sponsor: Section on Health Policy Statistics
• Author(s): Dave Osthus*+ and Alicia Carriquiry and Todd Campbell
• Companies: Iowa State University and Iowa State University and Iowa State University
• Keywords: Genetic algorithm ; optimization ; nutrition ; food fortification ; measurement error

Abstract

Methods for reliably estimating the distributions of usual (long-run average) daily nutrient intakes have been proposed (National Research Council, 1986; Nusser et. al. 1996). These estimates are then utilized for evaluating the adequacy of nutrient intake in sub-populations and for the subsequent development of programs to combat those inadequacies. One potential population-level intervention to reduce the prevalence of inadequacy is food fortification - where specific amounts of a nutrient is added to specific food vehicles. The goal of food fortification is to reduce the proportion of the population with inadequate nutrient consumption, at a reasonable cost. But how are food vehicles and nutrient amounts selected? We propose a method to optimize the process to plan food fortification. The approach we propose relies on the methodology proposed by Nusser et al., and uses a genetic algorithm to minimize the analytically untractable optimization function. The goal of the methodology is not to deliver a single, universal "best" food fortification plan, but rather a "best" plan under a variety of constraints. We illustrate the method by planning vitamin A intakes of Ugandan children.