Estimating measles reintroduction risk in São Paulo using machine learning

Area of Science:

• Epidemiology
• Public Health
• Data Science

Background:

• Measles remains a global health concern, necessitating robust surveillance and risk prediction systems.
• Brazil, particularly São Paulo State, requires effective tools to monitor and manage potential measles outbreaks.

Purpose of the Study:

• To develop and validate a machine learning-based tool for estimating measles risk at the municipal level in São Paulo State.
• To identify key factors influencing measles risk to inform public health interventions.

Main Methods:

• Trained five machine learning models using demographic, mobility, socioeconomic, and vaccination data (2007-2023).
• Employed K-fold cross-validation and Receiver Operating Characteristic (ROC) curves for model evaluation and threshold determination.
• Utilized Shapley Additive Explanations (SHAP) to identify significant risk factors.

Main Results:

• Models demonstrated comparable performance in predicting measles occurrence.
• The São Paulo metropolitan region and Santos were identified as high-risk areas.
• International passenger arrivals, WHO-reported cases, population size, and density were key predictive features.

Conclusions:

• A novel measles risk estimation tool utilizing publicly available data has been developed for São Paulo State.
• The framework provides a scalable model for measles risk assessment in other Brazilian states or regions.