Linear programming to design low cost, local RUTF

Summary of research¹

Ready to Use Therapeutic Food (RUTF) is not always available where needed. In India, where the need is enormous, it has not been possible to legally import RUTF from Europe since 2009. In this and other countries, the relatively high cost of Western brands and local policies have prevented the widespread importation of RUTF, boosting the demand for regionally appropriate solutions. The current RUTF formulation is based on results from a limited number of studies, in a few settings, showing rapid weight gain. Consequently, in other settings, with different underlying nutrient deficiencies and infectious disease profiles, similar weight gains would perhaps occur with nutrient levels different from those in use. Furthermore, the current commercial formulation of RUTF is not acceptable to all the patients in need of therapeutic nutrition in developing countries, e.g. HIV positive wasted adults.

Evidence based nutrition research ideally relies on costly randomised clinical trials. Therefore a robust method is needed to design the trial RUTF before such studies. At present, there is no internationally endorsed protocol to design products of this kind. Linear Programming (LP) is a suitable decision tool for designing novel food-based formulations. The method helps by identifying the cheapest possible combination of food ingredients that meet a set of nutritional requirements, avoiding a ‘trial and error’ approach.

The objective of a recent study was to test a LP-based method for designing the cheapest formulation of a readyto- use food (RUF) that fulfils predefined macronutrients requirements. It used region-specific foods that are culturally acceptable and can be processed with locally available technologies. The LP objective function and decision variables consisted of the lowest formulation price and the weights of the chosen commodities (soy, sorghum, maize, oil and sugar) respectively. The LP constraints were based on current United Nations (UN) recommendations for the macronutrient content of therapeutic feeds and included palatability, texture and maximum food ingredient weight criteria. Non linear constraints for nutrient ratios were converted to linear equations to allow their use in LP. The only software needed is MS Excel, including a freely available add-in called ‘Solver’ (see Figure 1).

The method was used to successfully design a prototype RUTF for the rehabilitation of HIV/TB-wasted adults and children under five years of age with severe acute malnutrition (SAM) in East Africa. The safety and acceptability of the prototype RUTF was subsequently confirmed in a trial². Laboratory analysis confirmed that the energy, protein and lipid values of the prototype formulation were within the pre-established cut-offs.

Some constraints were highlighted in applying LP to design food formulations of this kind. These were mainly to do with the accuracy of the food composition data in relation to local food ingredients, e.g. the oil descriptor used was palm oil but the actual commodity used was palm olein oil (an industrially prepared fraction of palm oil). The authors concluded that the LP method used was widely applicable for the rational design of therapeutic food products at minimum cost. The study provided a prototype formulation which met almost all the pre-defined requirements (one had to be relaxed by 0.2%). One lesson learnt is the need for improved methods to determine the ingredient prices to use in the model that takes into account seasonal/regional/ national fluctuations. The RUTF cost (based only on food ingredients) was approximately 4 to 5 times cheaper than the current standard product (food ingredients and premix) - hence even with the addition of micronutrient mix, still substantially cheaper. Using the methods described in the paper, public health nutritionists and food technologists could apply these steps to design other RUF formulations, such as ready-to-use supplementary or complementary foods. The authors do, however, caution that the macronutrient contents of LP prototypes always need to be confirmed by food composition analysis and the finalised products trialled under field conditions before they can be recommended for general use.

For more information, contact: Filippo Dibari,
email: filippo@validinternational.org

¹Dibari F et al (2012). Low-Cost, Ready-to-Use Therapeutic Foods Can Be Designed Using Locally Available Commodities with the Aid of Linear Programming. The Journal of Nutrition. First published ahead of print March 28th, 2012 as doi: 10.3945/jn.111.156943

²Owino, V. O., Irena, A. H., Dibari, F. and Collins, S. (2012), Development and acceptability of a novel milk-free soybean-maize-sorghum ready-to-use therapeutic food (SMS-RUTF) based on industrial extrusion cooking process. Maternal & Child Nutrition. http://onlinelibrary.wiley.com/doi/10.1111/j.1740-8709.2012.00400.x/abstract