To understand the impact of hospital-acquired infections on mortality and medical costs in modern medical care systems in different healthcare settings, we performed a case-control study at a medical centre and two community hospitals. A total of 144 and 129 adult case-control pairs who received care in a 2000-bed tertiary referral medical centre and two 800-bed community hospitals, respectively, between October 2002 and December 2002 were enrolled. Prolongation of hospital stay, extra costs and complications associated with hospital-acquired infections were analysed. Patients in the medical centre had more severe underlying disease status (P < 0.001), more malignancies (P < 0.001), more multiple episodes of hospital-acquired infection (p = 0.03), and more infections with multidrug-resistant bacteria (P < 0.001) than patients in community hospitals. The additional length of hospital stay and extra costs were similar for patients with hospital-acquired infections in the community hospitals and the medical centre (mean 19.2 days vs. 20.1 days, P = 0.79; mean 5335 US dollars vs. 5058 US dollars, P = 0.83; respectively). The additional length of hospital stay and extra costs in both the medical centre and the community hospitals were not related to the sites of infection or the bacterial pathogens causing hospital-acquired infections, although medical costs attributable to hospital-acquired fungal infections due to Candida spp. were much higher for patients in the medical centre. Prevalence of hospital-acquired-infection-related complications, such as adult respiratory distress syndrome, disseminated intravascular coagulation, organ failure or shock, was similar between the two groups, but patients in the medical centre had a higher mortality rate because of their underlying co-morbidities.