Aeroallergen Exposure as Short-Term Predictor of Respiratory Viral Infections in Two Health Regions (Caguas and San Juan) in Puerto Rico: A Seasonal and Machine Learning Approach

In this late-breaking abstract, we investigated whether seasonal and short-term variations in airborne fungal spores and pollen are associated with, and predictive of, influenza and COVID-19 incidence in the San Juan and Caguas health regions of Puerto Rico from 2022 to 2024. Using correlation analyses, lag modeling, logistic regression, and machine learning approaches, the study found that fungal spore concentrations—but not pollen—were consistently associated with and predictive of high-incidence influenza and COVID-19 days, particularly during the fall season. Peak associations occurred at short lags of 2–4 days, and random forest models demonstrated strong predictive performance for both influenza and COVID-19 outbreaks. These findings highlight fungal spores as a key seasonal environmental factor influencing respiratory virus transmission and support their integration into public health surveillance and outbreak forecasting systems.