Understanding the transmission dynamics of antibiotic resistant organisms in hospitals is critical to decreasing their prevalence and slowing the emergence of further resistance. However, the transmission is often both obscure and complex, presenting a challenge to informed decision making in most settings. This challenge is particularly acute for the hospital pathogen vancomycin resistant Enterococcus faecium which has spread globally and remains endemic in many hospital systems, including the University of Michigan Health System. I will present a retrospective study of a single tertiary care hospital. A mathematical model of hospital transmission was developed using the parameters of interest: the transmission rate, the recovery rate, and the probability of having VRE at the time of admission. These parameters are then inferred from hospital screening data, and applied to an agent based simulation utilizing four years of detailed patient movement patterns. Finally, we use targeted whole genome sequencing approach to validate the prediction on the size of transmission chains in the hospital, utilizing a novel method to predicted within-hospital transmission based on phylogenetic distance between strains. The model identifies the parameter range in which the R0 is less then 1 yet the hospital still acts as an amplifier. Thus, it cannot maintain the VRE alone, but can contribute to the transmission problem as part of the hospital network. We estimate that VRE transmission in the UMHS exists in this parameter range, with an R0 less than 1, yet the hospital amplified VRE prevalence by a factor of 2. This amplification results in transmission chains that experience prolonged hospital exposure, increasing the risk of hospital acquired antibiotic resistance. Consistent with the model, the whole genome sequencing identified multiple introductions with sustain transmission. These method present opportunities to enhance studies of VRE transmission risk factors, including microbiome risk factors. These transmission models can also present null model of transmission against which host and pathogen specific risk factor can be tested.
Sponsored by the Host Microbiome Initiative