Forecasting Seasonal Peaks of Pediatric Respiratory Infections Using an Alert-Based Model Combining SIR Dynamics and Historical Trends in Santiago, Chile
Gloria Henríquez, Jhoan Báez, Víctor Riquelme, Pedro Gajardo, Michel Royer, Héctor Ramírez
TL;DR
The study tackles forecasting seasonal peaks of pediatric ARI hospitalizations in Santiago by coupling a seasonal SIR model with a history-based predictor within an alert-driven framework. The method blends mechanistic and data-driven signals through a time-varying weight $\omega(t)$ and leverages data smoothing, a three-point epidemic alert ($t_0$, $t_{\min}$, $t_1$), and a penalized parameter calibration to forecast the peak date $\hat{t}$ and peak magnitude $\hat{h}$ with interpretable uncertainty. Retrospective and real-world validation across 2023–2024 at four major hospitals show peak-date forecasts preceding the event by about one month with high accuracy two weeks ahead, while peak-magnitude forecasts become informative roughly ten days before the peak and stabilize about a week prior. The framework demonstrates robust, operationally useful predictions under noisy hospital data and variable epidemic dynamics, offering a practical tool for hospital capacity planning and potentially generalizable deployment in similar settings. Future work aims to incorporate exogenous covariates (climate, co-circulating viruses, public health interventions) to enhance robustness and transferability.
Abstract
Acute respiratory infections (ARI) are a major cause of pediatric hospitalization in Chile, producing marked winter increases in demand that challenge hospital planning. This study presents an alert-based forecasting model to predict the timing and magnitude of ARI hospitalization peaks in Santiago. The approach integrates a seasonal SIR model with a historical mobile predictor, activated by a derivative-based alert system that detects early epidemic growth. Daily hospitalization data from DEIS were smoothed using a 15-day moving average and Savitzky-Golay filtering, and parameters were estimated using a penalized loss function to reduce sensitivity to noise. Retrospective evaluation and real-world implementation in major Santiago pediatric hospitals during 2023 and 2024 show that peak date can be anticipated about one month before the event and predicted with high accuracy two weeks in advance. Peak magnitude becomes informative roughly ten days before the peak and stabilizes one week prior. The model provides a practical and interpretable tool for hospital preparedness.