Deterministic and stochastic mathematical models were formulated to investigate the roles that environmental contamination and the presence of volunteers played in the dynamics of hospital infections in China. Semi-stochastic simulation was used to estimate some of the parameters by fitting the observed data and investigating the impacts of interventions such as cleaning, hand hygiene and isolation of admitted MRSA (Methicillin-resistant Staphylococcus aureus) patients on mean prevalence of infection. The basic reproduction number was estimated to be 0.9753. Numerical simulations show that environmental contamination is a threat to hospital infection and free-living bacteria in the environment can promote transmission and initiate infection even if an infection has died out among HCWs (health-care workers) and patients. Sensitivity analysis indicates that a contaminated environment and volunteers contribute substantially to MRSA transmission in hospital infections, and hence effective control measures should be targeted. Hand hygiene of volunteers and cleaning are more effective in reducing the mean prevalence of colonized patients than isolation of newly admitted MRSA-positive patients and hand hygiene of HCWs. Hence volunteers, a cadre of semi-professional nurses, are beneficial to both disease control and supplementary treatment of HCWs if they are well trained. However, isolation of newly admitted MRSA-positive patients could be influential and dominant in reducing the prevalence of infection when the environment within a ward is sufficiently clean. (C) 2011 Elsevier Ltd. All rights reserved.