Responsible AI for Public Health: A Methodological Illustration with a Forecasting Model applied to Respiratory Hospitalizations on SUS Data

• Ramon G. Pereira UFMG
• Luís Eduardo Limas Brito UFMG
• Italo Avelar UFMG
• Matheus Carvalho UFMG
• Marisa Vasconcelos UFMG
• Michele A. Brandão UFMG
• Wagner Meira Jr UFMG

DOI: https://doi.org/10.5753/sbcas.2026.20414

Abstract

This paper investigates how Responsible AI principles can be systematically integrated into public health predictive modeling using data from the Brazilian Unified Health System. We operationalize a four-layer architecture embedding governance, leakage-controlled temporal validation, fairness auditing, explainability, and structured documentation into the modeling lifecycle. In predicting monthly respiratory hospitalizations, LightGBM achieved an RMSE of 13.81 but exhibited a 44.8 percent sMAPE disparity in the highest disparity state. A resampling adjustment reduced this gap by 8.51 percentage points. The results demonstrate how structured Responsible AI controls reshape evaluation beyond aggregate predictive accuracy.

References

Albuquerque, M., Viana, A., Lima, L., Ferreira, M., Fusaro, E., and Iozzi, F. (2017). Regional Health Inequalities: Changes Observed in Brazil from 2000–2016. Ciência & Saúde Coletiva, 22(4):1055–1064.

Barocas, S., Hardt, M., and Narayanan, A. (2023). Fairness and Machine Learning: Limitations and Opportunities. MIT Press, Cambridge, MA, USA.

Bergmeir, C. and Benítez, J. M. (2012). On the use of cross-validation for time series predictor evaluation. Information Sciences, 191(1):192–213.

Birck, M. G., Ferreira, R., Curi, M., et al. (2023). Real-world treatment patterns of rheumatoid arthritis in brazil: analysis of datasus national administrative claims data for pharmacoepidemiology studies (2010–2020). Scientific Reports, 13(1):17739.

Chin, M. H. et al. (2023). Guiding principles to address the impact of algorithm bias on racial and ethnic disparities in health and health care. JAMA Network Open, 6(12):e2345050.

EU (2024). Regulation (EU) 2024/1689 of the European Parliament and of the Council (Artificial Intelligence Act). Official Journal of the European Union. In: [link]. Acesso em: 11/01/2026.

Ferreira, A. P. L. d. A., Redaelli, E., Maldaner, L. F., and Rigo, S. J. (2025). Inovação e transformação no ecossistema de saúde auditiva: Inteligência artificial e interoperabilidade. In Simpósio Brasileiro de Computação Aplicada à Saúde, pages 1–12, Porto Alegre, Brasil.

Justo, N. et al. (2019). Real-world evidence in healthcare decision making: Global trends and case studies from latin america. Value in Health, 22(6):739–749.

Ke, G., Meng, Q., Finley, T., Wang, T., Chen, W., Ma, W., Ye, Q., and Liu, T.-Y. (2017). LightGBM: A Highly Efficient Gradient Boosting Decision Tree. In Advances in Neural Information Processing Systems, pages 3146–3154, Long Beach, USA.

Landmann-Szwarcwald, C. and Macinko, J. (2016). A panorama of health inequalities in brazil. International Journal for Equity in Health, 15(1):174.

Lekadir, K. et al. (2025). FUTURE-AI: International Consensus Guideline for Trust-worthy and Deployable Artificial Intelligence in Healthcare. British Medical Journal, 388(1):e078499.

Lemaître, G., Nogueira, F., and Aridas, C. K. (2017). Imbalanced-learn: A python toolbox to tackle the curse of imbalanced datasets in machine learning. Journal of Machine Learning Research, 18(17):1–5.

Lemos, D., Fonseca, L., Florêncio, R., de Almeida, J., Lima, I., and Gualdi, L. (2024). Hospitalisations and Fatality due to Respiratory Diseases according to a National Database in Brazil: a Longitudinal Study. BMJ Open Respiratory Research, 11(1):e002103.

Lundberg, S. and Lee, S.-I. (2017). A unified approach to interpreting model predictions. In Advances in Neural Information Processing Systems, NIPS’17, pages 4765–4774, Long Beach, USA. NeurIPS, Curran Associates, Inc.

Mitchell, M. et al. (2019). Model Cards for Model Reporting. In ACM Conference on Fairness, Accountability, and Transparency, FAT* 2019, pages 220–229, Atlanta, USA. ACM.

Mittelstadt, B. (2019). Principles alone cannot guarantee ethical ai. Nature Machine Intelligence, 1(11):501–507.

NAM (2024). Health Care Artificial Intelligence Code of Conduct. National Academy of Medicine. In: [link].

NIST (2023). Artificial Intelligence Risk Management Framework (AI RMF 1.0) — NIST AI 100-1. National Institute of Standards and Technology. In: [link].

OECD (2024). Recommendation of the Council on Artificial Intelligence. OECD Legal Instruments. In: [link].

Paim, J., Travassos, C., Almeida, C., Bahia, L., and Macinko, J. (2011). The Brazilian health system: History, advances, and challenges. The Lancet, 377(9779):1778–1797.

Pushkarna, M., Zaldivar, A., and Kjartansson, O. (2022). Data Cards: Purposeful and Transparent Dataset Documentation for Responsible AI. In ACM Conference on Fairness, Accountability, and Transparency, pages 1936–1949, Seoul, South Korea.

Scaramussa, L. M. and Pacheco, A. G. C. (2025). Anotação de imagens médicas assistida por ia: um estudo sobre segmentação de lesões de pele por não especialistas. In Simpósio Brasileiro de Computação Aplicada à Saúde, pages 152–163, Porto Alegre, Brasil.

Soyiri, I. N. and Reidpath, D. D. (2013). An overview of health forecasting. Environmental Health and Preventive Medicine, 18(1):1–9.

UNESCO (2021). Recommendation on the ethics of artificial intelligence. United Nations Educational, Scientific and Cultural Organization. In: [link].

WHO (2021). Ethics and governance of artificial intelligence for health. World Health Organization. Guidance document. In: [link].