OBJECTIVES: To estimate the total economic burden attributed to drug-resistant (DR) nosocomial gram-negative (GN) bacteraemia amongst intensive care unit (ICU) patients in Singapore, from a hospital perspective. METHODS: A decision-tree model was developed to estimate the hospital costs of ICU patients with nosocomial GN bacteraemia. Upon development of a GN bacteraemia, patients were initially treated by an empirical antibiotic, E1, and a culture test was taken. If the empirical antibiotic was not appropriate, patients were subsequently treated by a definitive antibiotic, D1, based on the culture test results. If D1 was also not appropriate, a second definitive antibiotic, D2, was administered. Finally, if the patient remained infected after treatment by D2, the patient was assumed to die. Each stage of therapy was associated with a probability of success, leading to discharge; or of failure, leading to prolonged hospital stay and subsequent lines of therapy, or death. Singapore data from a structured literature review were used when possible, supplemented by Singapore data from non-peer reviewed sources and conservative assumptions. Guidance on model structure and assumptions were provided by local clinical experts. RESULTS: The model provides a conceptual framework for assessing the impact of drug-resistance on hospital costs amongst patients with GN bacteraemia. With the current assumptions, DR nosocomial GN bacteraemia amongst ICU patients in Singapore were estimated to be associated with excess hospital costs of 6.96 million SGD over one year, with length of hospital stay as a major cost driver. CONCLUSIONS: Drug-resistance in GN bacteria is a major public health concern associated with excess hospital costs and deaths. Our conservative estimations suggest that antibiotic surveillance programmes and other measures to reduce the prevalence of DR infections may contribute to cost savings in Singapore hospitals.