Introduction: Respiratory syncytial virus (RSV) infection is a major cause of hospitalization in children. Meteorological factors are known to influence seasonal RSV epidemics, but the relationship between meteorological factors and RSV infection in children is not well understood. We aimed to explore the relationship between meteorological factors and RSV infections among hospitalized children, using different statistical models.
Methods: We conducted a retrospective review concerning children with RSV infections admitted to a tertiary pediatric hospital in Wenzhou, China, between January 2008 and December 2017. The relationship between meteorological factors (average daily temperatures, average daily relative humidity, rainfall, rainfall days, and wind speed) and the incidence of RSV in hospitalized children was analyzed using three time-series models, namely an autoregressive integrated moving average (ARIMA) model, a generalized additive model (GAM), and a least absolute shrinkage and selection operator (LASSO)-based model.
Results: In total, 15 858 (17.6%) children tested positive for RSV infection. The ARIMA model revealed a marked seasonal pattern in the RSV detection rate, which peaked in winter and spring. The model was a good predictor of RSV incidence (R2 : 83.5%). The GAM revealed that a lower temperature and higher wind speed preceded increases in RSV detection. The LASSO-based model revealed that temperature and relative humidity were negatively correlated with RSV detection.
Conclusions: Seasonality of RSV infection in hospitalized children correlated strongly with temperature. The LASSO-based model can be used to predict annual RSV epidemics using weather forecast data.
Keywords: meteorological factors; respiratory syncytial virus; seasonal variation; time-series analysis; weather.
© 2020 Wiley Periodicals, Inc.