Inference and control of the nosocomial transmission of methicillin-resistant Staphylococcus aureus

Pei, Sen; Morone, Flaviano; Liljeros, Fredrik; Makse, Hernán; Shaman, Jeffrey L

Model patients are classified into three categories: susceptible individuals who are free of MRSA (S), colonized individuals who carry the bacteria asymptomatically (C), and confirmed positive patients (P). The model simulates two connected dynamics: nosocomial transmission and importation from the community. Here, the community is broadly defined as all locations outside the study hospitals, and may include households and healthcare facilities not covered in the study. Within hospitals, transitions between states (S, C, P) are governed by parameters that help define either interaction dynamics or the progression of infection. Specifically, a susceptible individual staying in a ward with a colonized person can become MRSA colonized with transmission probability b per day. In our model, we assume that patients within a ward have the same rate of contact with each other, presumably mediated by the shared healthcare workers in a ward. The transmission process is densitydependent, as the force of infection in a ward increases with the number of colonized patients within the ward (Begon et al., 2002) . Upon colonization, asymptomatic persons can return to the susceptible state at a spontaneous decolonization rate a, or they can test positive with an infection progression rate p. We assume infected patients will receive treatment, no longer spread bacteria, and return to state S with a recovery rate . Treatment is assumed to continue until infected patients are clear of MRSA. Given the exponential decay of infection probability, the characteristic treatment period is 1= days. Note that colonization only occurs between individuals connected by a link in the contact network, whereas decolonization, infection and recovery progress spontaneously, independent of the contact network. Outside the study hospitals, the transmission process is not explicitly simulated; instead, two additional parameters are introduced to represent transmission intensity. For Sources for parameter rangesa: (Cooper et al., 2004a; Bootsma et al., 2006; Eveillard et al., 2006; Wang et al., 2013; Macal et al., 2014; Jarynowski and Liljeros, 2015) ; p: (Kajita et al., 2007; Jarynowski and Liljeros, 2015) ; : (D' Agata et al., 2009; Wang et al., 2013) ; b: Prior; I 0 : Prior; C 0 : Prior, (Hidron et al., 2005; Eveillard et al., 2006; Jarvis et al., 2012) . For each individual, the infection progress rate p is drawn after a is specified. patients who appear for the first time in hospital, we assume they belong to states C and P with probability C 0 and I 0 , respectively. As importation rates of colonized and infected patients depend on the time-varying MRSA prevalence outside hospitals, we assume the parameters C 0 and I 0 are time-dependent. Once patients appear in the contact network, the evolution of their states follows the dynamics as defined above. After discharge, we continue tracking the progression of colonized individuals; however, transmission outside the study hospitals is not represented. The flow of individuals between categories is illustrated schematically in Figure 1E . For a realistic scenario, disease-related model parameters may differ from person to person. To account for this variability during implementation, parameters, for example a, p and , for each individual are randomly drawn from uniform ranges obtained from prior literature ( Table 1) . The parameter ranges are enlarged slightly to cover the values reported in these works. Our main objective is