Using Multiple Methods to Estimate Respiratory Syncytial Virus (RSV)-associated Hospitalization Rates in Children Aged < 5 Years—Hamilton County, Ohio, 2009–2017

Abstract

Background

Respiratory syncytial virus (RSV) is a leading cause of lower respiratory tract infection in children less than 5 years of age worldwide. In the United States, RSV commonly causes hospitalization in young children and is the leading cause of hospitalizations in infants. As new RSV immunizations become available, burden estimates are critical to guide the implementation of recommendations and quantify impact.

Methods

We estimated RSV-associated hospitalization rates at a large US pediatric medical center during an 8-year period using five approaches, namely, estimation directly from active and passive surveillance systems, both a crude and stratified capture–recapture analysis of data from both systems, and estimation based on discharge diagnosis codes. The stratified analysis was performed to ensure adherence with the capture–recapture methodology assumption that samples are independent and participants have an equal probability of being observed within each system.

Results

Overall, estimated RSV-associated hospitalization rates per 1000 children were 4.0 (2.5, 6.1) based on adjusted estimates from active surveillance, 1.7 (2.1, 4.4) from passive surveillance, 7.9 (5.7, 13.0) from crude capture–recapture analysis, 5.0 (3.8, 7.2) from the stratified capture–recapture, and 4.4 (4.0, 4.9) from discharge diagnosis codes.

Conclusions

Each method has limitations and inherent biases that may impact the estimation of the burden of RSV. Capture–recapture analysis may be a useful tool to estimate the burden of RSV, but needs to be adjusted to account for possible violation of the assumptions of independence and equal probability of capture to ensure accurate approximation of disease burden and avoid over estimation.