Low-Cost Dengue Triage: Predicting Disease Severity using Machine Learning Without Laboratory Biomarkers

• Maximus B. Rosa UFRGS
• Gabriel C. Tavares UFRGS
• Mariana Recamonde-Mendoza UFRGS / HCPA

DOI: https://doi.org/10.5753/sbcas.2026.21367

Resumo

Dengue remains a major public health challenge in tropical regions, where early risk stratification is critical during outbreaks. We predict dengue severity using only demographic and clinical data available at patient triage, without laboratory tests or biomarkers. Using a nationwide Brazilian surveillance dataset with over 700,000 confirmed cases, we trained tree-based machine learning models to classify patients into the three World Health Organization severity categories: low risk, dengue with warning signs, and severe dengue. The best-performing model, a Random Forest enriched with urban hierarchy information, achieved a macro F1-score of 0.61 (± 0.001) in validation and 0.63 on an independent test set. SHAP analysis revealed clinically plausible predictors of severity, highlighting the potential of surveillance data to support early dengue triage in resource-constrained settings.

Referências

Akhtar, R., Gupta, P., and Srivastava, A. (2016). Urbanization, Urban Heat Island Effects and Dengue Outbreak in Delhi, pages 99–111.

Akiba, T., Sano, S., Yanase, T., Ohta, T., and Koyama, M. (2019). Optuna: A next-generation hyperparameter optimization framework. In Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, pages 2623–2631. ACM.

Bhatt, S., Gething, P. W., Brady, O. J., et al. (2013). The global distribution and burden of dengue. Nature, 496(7446), 504–507.

da Silva Neto, S. R., Tabosa Oliveira, T., Teixeira, I. V., Aguiar de Oliveira, S. B., Souza Sampaio, V., Lynn, T., and Endo, P. T. (2022). Machine learning and deep learning techniques to support clinical diagnosis of arboviral diseases: A systematic review. PLoS Neglected Tropical Diseases, 16(1):e0010061.

Frank, E. and Kramer, S. (2004). Ensembles of nested dichotomies for multi-class problems. In Proceedings of the Twenty-First International Conference on Machine Learning, ICML ’04, page 39, New York, NY, USA. Association for Computing Machinery.

Grinsztajn, L., Oyallon, E., and Varoquaux, G. (2022). Why do tree-based models still outperform deep learning on typical tabular data? In Proceedings of the 36th International Conference on Neural Information Processing Systems, NIPS ’22.

Gubler, D. J. (2011). Dengue, urbanization and globalization: The unholy trinity of the 21st century. Tropical Medicine and Health, 39:3 – 11.

Huang, S.-W., Tsai, H.-P., Hung, S.-J., Ko, W.-C., and Wang, J.-R. (2020). Assessing the risk of dengue severity using demographic information and laboratory test results with machine learning. PLoS Neglected Tropical Diseases, 14(12):e0008960.

IBGE (2020). Regiões de influência das cidades: 2018. IBGE, Rio de Janeiro. Coordenação de Geografia.

Jean Pierre, A. R., Green, S. R., Anandaraj, L., Sivaprakasam, M., Kasirajan, A., Devaraju, P., Anumulapuri, S., Mutheneni, S. R., and Balakrishna Pillai, A. (2024). Severity prediction markers in dengue: a prospective cohort study using machine learning approach. Biomarkers, 29(8):557–564.

Leung, X. Y., Islam, R. M., Adhami, M., Ilic, D., McDonald, L., Palawaththa, S., Diug, B., Munshi, S. U., and Karim, M. N. (2023). A systematic review of dengue outbreak prediction models: Current scenario and future directions. PLOS Neglected Tropical Diseases, 17(2):1–21.

Lim, A., Shearer, F. M., Sewalk, K., Pigott, D. M., Clarke, J., Ghouse, A., Judge, C., Kang, H., Messina, J. P., Kraemer, M. U., et al. (2025). The overlapping global distribution of dengue, chikungunya, zika and yellow fever. Nature Communications, 16(1):3418.

Lundberg, S. and Lee, S.-I. (2017). A unified approach to interpreting model predictions.

Ministério da Saúde (2024a). Casos suspeitos de dengue. [link]. Acesso em: 05 fev. 2026.

Ministério da Saúde (2024b). Dengue. [link]. Acesso em: 05 fev. 2026.

Salazar Flórez, J., Marín Velásquez, K., Giraldo, L., Segura Cardona, A., Jaramillo, B., and Arboleda, M. (2025). Dengue severity prediction in a hyperendemic region in Colombia. Viruses, 17:740.

Sivasubramanian, K., Bharath R, R., Vajravelu, L. K., Kumar D, M., and Banerjee, A. (2025). Key laboratory markers for early detection of severe dengue. Viruses, 17(5):661.

Villabona-Arenas, C. J., de Oliveira, J. L., Capra, C. d. S., Balarini, K., Loureiro, M., Fonseca, C. R. T. P., Passos, S. D., and Zanotto, P. M. d. A. (2014). Detection of four dengue serotypes suggests rise in hyperendemicity in urban centers of brazil. PLoS Neglected Tropical Diseases, 8(2):e2620. eCollection 2014 Feb.

World Health Organization (2009). Dengue: guidelines for diagnosis, treatment, prevention and control. New Edition. Geneva: World Health Organization.

World Health Organization (2025). Dengue: global situation, surveillance and progress – 2024 update. Weekly Epidemiological Record, No 52, 100, 665–678. Published 26 December 2025.