Determining variable contagiousness of MRSA by setting

Abstract

Objective and Hypothesis Methicillin resistant Staphylococcus aureus (MRSA) is currently a major cause of skin and soft tissue infections (SSTI) in the United States. In order to characterize the spread of MRSA in the pediatric population we built a probabilistic, discrete-event, individual-based simulation. Specifically, our model looked at the spread of MRSA in households and at schools to determine if there was a difference in communicability between the two settings. Methods We developed a probabilistic, discrete-event, individual-based model. This model was validated using insurance billing data for skin and soft tissue infections. The first validation trained the model for two years of data, and validated it with the next two years of data. The second method trained the model in one region and validated it in another. Following the validation, the Poisson-bootstrap resampling method was used to find specific values for a contagiousness factor(CF) in households and schools. Results Both methods of validation supported the model with no statistically significant difference. The bootstrap resulted in a CFhousehold of 30.69 (95% CI [29.09, 32.29]) and a CFschool of 0.55 (95% CI [0.46 to 0.64]). Effective reproduction number for the school setting was found to be 0.0015 and 0.06 to 3.04 for households of different size. Conclusion In this study we characterize a marked difference in communicability in the household and at school, which has not previously been shown. The identification of colonization clusters in households can be used to design strategies reduce the disease burden. The model can be used to simulate and predict responses to different interventions.