Predicting age of respiratory syncytial virus infection from birth timing

Respiratory syncytial virus (RSV) infects nearly all children by age 2 to 3 years, and early-life infection-defined using active and passive surveillance with quantitative polymerase chain reaction- and serology-identified infection-has been implicated as a causal factor in childhood asthma. As such, identifying infants that are likely to be infected with RSV during this critical susceptibility window has important implications for identifying individuals at risk for chronic respiratory sequelae. However, determining the age of RSV infection in large populations is challenging because many infections are asymptomatic, making accurate detection dependent on intensive and costly surveillance. To address this, we developed a probability model for age of first RSV infection. It uses an infant's birthdate, demographic covariates, and publicly available RSV circulation data to determine the probability they were first infected at any age from birth to one year. Our model is interpretable, accounts for nearly 37% of the variance in age at first infection, and generalizes across four independent datasets collected from participants in the United States, where we use it to accurately predict age of first infection in two independent cohorts. Our work facilitates reliable estimation of the age of infant RSV infection during the first year of life without the need for active surveillance.