Prediction of Hospitalization Rates Due to Malnutrition in the Brazilian Unified Health System Using Machine Learning

• Vanessa P. Resmini UFRGS
• Mariana Recamonde-Mendoza UFRGS

DOI: https://doi.org/10.5753/sbcas.2026.21362

Abstract

According to the 2025 Hunger Map, 35 million people in Brazil face food insecurity, demonstrating the need to maintain and structure policies to reduce hunger and poverty. In this study, machine learning models were developed to predict the annual rate of hospitalizations due to malnutrition in the brazilian Sistema Único de Saúde (SUS), segmented by the country’s 439 health regions, based on patients place of residence. In the out-of-time dataset, R² values of 0.66 were observed for LGBM, 0.64 for Random Forest, and 0.65 for XGBoost, while MAPE values were 0.35, 0.42, and 0.37, respectively. The results also showed variations in model performance among the Brazilian macro-regions.

References

Almalki, A., Gokaraju, B., Mehta, N., and Doss, D. A. (2021). Geospatial and machine learning regression techniques for analyzing food access impact on health issues in sustainable communities. ISPRS International Journal of Geo-Information, 10(11).

Almeida, G., Brito Correia, F., Borges, A. R., and Bernardino, J. (2024). Hospital length-of-stay prediction using machine learning algorithms—a literature review. Applied Sciences, 14(22).

Awan, A. A. (2024). Introdução ao t-sne: Redução de dimensionalidade não linear e visualização de dados. Acessado em: 22 dez. 2025.

Barbosa, R. M. and Nelson, D. R. (2016). The use of support vector machine to analyze food security in a region of brazil. Applied Artificial Intelligence, 30(4):318–330.

Cao, G., Kornher, L., and Brandi, C. (2025). How robust are machine learning approaches for improving food security amid crises?-evidence from COVID-19 in Uganda. World Development, 196:107171.

da Silva, T. L. C., Furtado, L. S., Fernandes, G. S., de Macêdo, J. A. F., Cruz, L. A., Magalhães, R. P., and Gomes, L. G. A. (2025). Machine learning methods and models to predict food insecurity levels for families in Ceará, Brazil, based on employment, housing and other social indicators. Journal of the Brazilian Computer Society, 31(1):203–218.

FAO, IFAD, UNICEF, WFP, and WHO (2025). The state of food security and nutrition in the world 2025 – addressing high food price inflation for food security and nutrition. Technical report, Food and Agriculture Organization of the United Nations, Rome.

Jarray, N., Abbes, A. B., and Farah, I. R. (2023). Machine learning for food security: current status, challenges, and future perspectives. Artificial Intelligence Review, 56(Suppl 3):3853–3876.

Li, J. (2017). Assessing the accuracy of predictive models for numerical data: Not r nor r2, why not? then what? PLOS ONE, 12(8):1–16.

Simeon, Y. (2023). Alimentação escolar é a principal refeição para 56% dos estudantes do grande rio, revela pesquisa. Acessado em: 31 dez. 2025.

Valadares, A., Souza, M., and Esteves, M. (2023). Mapeamento da insegurança alimentar e nutricional com foco na desnutrição (mapa insan) a partir da análise do sistema nacional de vigilância alimentar e nutricional (sisvan). Technical report, Ministério do Desenvolvimento e da Assistência Social, Família e Combate à Fome Secretaria Extraordinária de Combate à Pobreza e à Fome.

Xiong, R., Peng, H., Chen, X., and Shuai, C. (2024). Machine learning-enhanced evaluation of food security across 169 economies. Environment, Development and Sustainability, 26(10):26971–27000.