Enterococcus faecium is a leading cause of hospital-acquired infection, disproportionally affecting immunocompromised and critically ill patients. Despite infection control measures, rates of vancomycin-resistant E. faecium (VREfm) bacteraemias have failed to decline in the United Kingdom, and Cambridge University Hospitals (CUH) report the highest numbers nationally. The aims of my PhD were to use epidemiological and genomic surveillance data to establish risk factors for acquisition and infection with E. faecium in patients at CUH, and to use a One Health approach to consider possible sources for hospital patients by relating bloodstream-associated isolates with those cultured from livestock and the environment in the same geographic region. A retrospective matched nested case control study was performed to determine risk factors for VRE bacteraemia relating to antibiotic exposure. 235 cases were matched to 220 controls for length of admission, year, specialty and ward type. Multivariable analysis demonstrated that duration of exposure to parenteral vancomycin, fluoroquinolones and meropenem were independently associated with VRE bacteraemia. This provides evidence for the importance of antimicrobial stewardship targeting high-risk antibiotics in patients at risk of VRE bacteraemia. VREfm bacteraemia may be complicated by disease recurrence. Whole genome sequencing was used to distinguish between relapse and reinfection in 14 episodes of recurrent VREfm bacteraemia. This demonstrated that 10 (71%) episodes were due to reinfection with a new strain, with reinfection being more likely with increasing time between two positive cultures. This study also evaluated 9 patients with blood cultures positive for both VREfm and vancomycin-susceptible E. faecium (VSEfm), the majority (78%) of which were found to be unrelated strains. More than half of all study isolates from these two patient groups were closely related to another isolate causing bacteraemia at CUH, suggesting that hospital acquisition of VREfm is a driver for infection and recurrence. A cross-sectional study of E. faecium in raw and treated wastewater from 20 municipal water treatment plants across the East of England revealed widespread dissemination of healthcare-associated lineages of VREfm in all sampled locations including rural areas, and environmental release in treated wastewater in 17/20 locations. Wastewater isolates were genetically intermixed with isolates causing bacteraemia at CUH, including highly related isolates indicating recent transmission between the two reservoirs. These findings are consistent with widespread distribution of healthcare-associated VREfm in community populations. A One Health approach incorporating sampling from livestock (10 pork, 10 cattle, 9 poultry farms) detected no VREfm in animals whilst 2 independent meat surveys demonstrated VREfm in 1-2% of uncooked products. Genomic comparison of >1400 E. faecium isolates from livestock, meat, wastewater and almost 800 people with bloodstream infection demonstrated that livestock and human isolates were genetically distinct. Analysis of the accessory genome added further evidence for distinct gene content associated with niche adaptation. An analysis of mobile genes encoding antibiotic resistance revealed limited evidence of sharing between human and animal populations. A prospective longitudinal study in haematology patients at CUH over 6 months revealed high rates of VREfm carriage (63% of cases) and environmental contamination (49% of samples). Genomic analysis elucidated complex colonisation dynamics with frequent loss and acquisition of subtypes, including unsuspected acquisition of new VREfm subtypes in patients already colonised with VREfm, and multiple transmission chains involving patients and the environment, including some leading to bacteraemia. These findings highlight the shortcomings of infection control and environmental cleaning and provide the basis for revised interventions.