|
Abstract:
|
Methicillin-resistant Staphylococcus aureus (MRSA) is a pathogen of public health importance. In 2005, the Centers for Disease Control and Prevention's Emerging Infections Program (EIP) started active, population-based surveillance for invasive MRSA, which represent about 19 million persons in more than 4,000 census tracts within 9 states in 2015. Over a decade, the national burden of MRSA was estimated based on the sample weighting stratified by age, gender, race and underline disease conditions. Recent study shown area-based socioeconomic factors are highly associated with invasive MRSA incidence. In this study, a Bayesian hierarchical geostatistical Poisson model with conditionally autoregressive (CAR) priors was fitted on the observed MRSA incidence data. Important underline disease conditions, demographic and socioeconomic factors were identified by applying Bayesian model averaging (BMA) within geostatistical model. Incidences for tracts not under surveillance are estimated automatically by the model for the selected states. We compare and contrast the new method with the traditional weighted sum based method and the results are evaluated.
|
