We have addressed knowledge gaps in the epidemiology of Clostridium difficile using a programme of epidemiological research involving prospective molecular epidemiological surveys, mathematical modelling, economic studies, systematic reviews and meta-analyses. We demonstrated a very similar profile of PCR ribotypes of C. difficile in hospitals and the community. Modelling studies demonstrated that transmission of infections in the hospital is sustained by introduction of infections from the community, without which, infection in hospitals would die out. The current two-day cut-off defining hospital versus community acquisition after date of admission underestimates the proportion of infections that are community-acquired (the cut-off should be five or six days). Screening of admissions is unlikely to effectively reduce importation of infections from the community because many exposed individuals will be missed by current screening approaches. Infections in hospital are still epidemiologically important because hospitals have a high concentration of susceptible individuals (who have multiple comorbidities and a high exposure to antibiotics and other drugs), and a higher proportion of individuals are discharged from hospitals carrying C. difficile than are admitted. Improving hygiene and sanitation, and reducing average length of stay (as opposed to antibiotic stewardship) are the most cost-effective means of reducing infection rates in the hospital, particularly when delivered in combination. In the community, infants and animals likely contribute significantly to transmission. Targeted, risk-based surveillance (both within hospitals and at the national level) are supported by the evidence. Such a risk-based approach would consider history of antibiotic and other pharmaceutical drug exposure, comorbidities, time of year and international travel history, and would use evidence-based surveillance definitions that attribute a much higher percentage of hospital-identified cases as being community-acquired than is currently the case. Overall, there needs to be a shift towards a more public health oriented approach to C. difficile and other healthcare-acquired infections that has a greater focus on community prevention and away from the current, dominant biomedical approach that focusses on therapeutic solutions.