Team: 28
School: Los Alamos High
Area of Science: Translational Medical Sciences
Proposal: Percentages of acute malnutrition continue to be unsettlingly high in developing countries, while coverage of treatment remains unsatisfactory. Treatment and prevention of acute malnutrition typically relies on ready-to-use food (RUF) distributed by international agencies. Common RUF is effective but costly with a limited global supply chain. The method to optimize the treatment of malnutrition can be broken into three sections: Optimization of locally suitable ready-to-use food (RUF), optimizing local production of RUF, and forecasting of malnutrition.
Optimizing RUF:
- Optimization of RUF formula to improve cost-effectiveness using environmentally suitable local ingredients and local production can reduce cost, improve treatment acceptance (efficacy), and enhance food security.
- Using a novel optimization tool, RUF formulae for sub-Saharan Africa were optimized for low cost while meeting macro/micronutrient, product composition, and protein quality goals.
- Compared to current RUF, ingredient cost was reduced by over 2/3rds, and water footprint by over 3/4ths.
Optimizing local production of RUF:
- National survey data adjusted for incidence served to estimate the yearly RUF demand per country.
- Transport cost was estimated using satellite road maps and trucking fees.
- The fixed and operating cost of local production was estimated based on comparable manufacturing cost in each country.
- The spatial distribution of RUF manufacturing sites was optimized for maximum coverage at lowest cost.
Predicting prevalence of acute malnutrition:
- Methods were also created to predict the spatial prevalence of malnutrition throughout sub-saharan Africa.
The training data can be broken into two sections: real-time and static variables.
- Real-time indicators include variables that could help analyze crop production and crop availability on the markets, such as crop yield predictions from daily MODIS satellite imagery, indicators of political stability (e.g. conflicts), news media reports, and food price indices.
- Static variables include poverty indicators from the DHS, roads, urban agglomerations, night lights, major waterways, market potentials, and land cover.
- When all of these static and real-time predictors are joined using a machine learning algorithm, they are able to predict future malnutrition with reasonable accuracy.
Using the estimated prevalence of acute malnutrition, the production of RUF can be adjusted to meet demand for the following year, and the net cost may be compared to current models.
Team Members:
Sponsoring Teacher: NA