(P06-026-25) Using Linear Programming to Optimize a Locally-Produced Ready-to-Use Therapeutic Food for Childhood Acute Malnutrition in India and Pakistan

Objectives: Commercial ready-to-use therapeutic foods (RUTF) are efficacious but costly, especially for managing treatment of acutely malnourished children in poor areas. Our aim was to apply linear programming (LP) to develop a nutritionally adequate recipe for a RUTF from locally sourced foods available all year in India and Pakistan.

Methods: We accessed data sources for nutrient composition of South Asian foods, food prices in markets, as well as the literature on requirements for energy, macronutrients, vitamins, minerals, and essential fatty acids [EFA, alpha-linolenic acid (ALA) and linoleic acid (LA)] required for neurocognitive recovery from malnutrition. LP, an optimization technique, was applied to minimize costs of locally-sourced ingredients and determine selection and amounts for producing an RUTF that meets WHO therapeutic guidelines. The LP tool output is a predicted list of food ingredients with required amounts of each, and prices of an optimally formulated recipe. The selected RUTF recipe was produced locally in Lahore, Pakistan and samples were lab-tested for nutritional content.

Results: The optimized recipe comprised maize, milk powder, rice flour, soymeal, sugar, whey isolate, oil palm, and other oils, portioned to 100g, providing 552kcal from 12g protein, 33g fat, and 51g CHO, with essential vitamins, minerals and EFA that aligned with WHO recommendations.  The total energy predicted by the LP tool was 0.75% lower than the lab-measured value. Macronutrient predictions were within 10% of lab values.  Predicted LA and ALA levels were higher by 66% and 34%, respectively, than amounts measured in the food. Detected quantities of lipid, ALA, and moisture content in the lab fell within WHO recommended ranges, while the quantities of protein, LA, and total energy deviated from the suggested ranges by 13% lower, 19% lower, and .46% higher, respectively.

Conclusions: Linear programming is a tool that can assemble a low-cost, culturally appropriate RUTF meeting nutritional recommendations for managing acute childhood malnutrition, and predict with reasonable accuracy the food’s energy and nutrient composition when compared to laboratory-measured estimates.

Funding Sources: Sight and Life, Johns Hopkins University