Application of linear programming in the development of complementary feeding recommendations: A systematic review

The use of linear programming (LP) to develop complementary food recommendations (CFRs) is gaining interest due to its ability to produce low-cost population-specific food-based recommendations (FBRs). This review aimed to identify the components of LP models commonly used in developing CFRs and summarize the evidence on the use of LP-developed CFRs as an intervention strategy. The databases PubMed, Science Direct, Scopus, Cochrane Library, Web of Science, and Google Scholar were searched for relevant articles. LP was used in twenty-six studies to develop CFRs for young children and in three studies LP-developed CFRs were applied to improving complementary feeding. The objective function varied across studies, such as to maximize nutrient content, minimize cost, minimize deviation between observed and modeled diets, and minimize multiple nutrient deficiencies. All studies applied nutritional and acceptability constraints. Individual intervention studies showed that LP-developed CFRs can improve children's nutrient intake and feeding practices, as well as mother's nutrition knowledge. Various applications of LP have been used to develop optimal infant diets. However, LP-developed CFRs as a nutrition intervention strategy have been used in only a few studies. Further robust research is needed to test LP-developed CFRs.