OBJECTIVES: To describe the components of an emergency and urgent care system within one health authority and to investigate ways in which patient flows and system capacity could be improved. METHODS: Using a qualitative system dynamics (SD) approach, data from interviews were used to build a conceptual map of the system illustrating patient pathways from entry to discharge. The map was used to construct a quantitative SD model populated with demographic and activity data to simulate patterns of demand, activity, contingencies, and system bottlenecks. Using simulation experiments, a range of scenarios were tested to determine their likely effectiveness in meeting future objectives and targets. RESULTS: Emergency hospital admissions grew at a faster annual rate than the national average for 1998-2001. Without intervention, and assuming this trend continued, acute hospitals were likely to have difficulty sustaining levels of elective work, in reaching elective admission targets and in achieving bed occupancy targets. General practice admissions exerted the greatest influence on occupancy rates. Prevention of emergency admissions for older people (3%-6% each year) reduced bed occupancy in both hospitals by 1% per annum over five years. Prevention of emergency admissions for patients with chronic respiratory disease affected occupancy less noticeably, but because of the seasonal pattern of admissions, had an effect on peak winter occupancy. CONCLUSIONS: Modelling showed the potential consequences of continued growth in demand for emergency care, but also considerable scope to intervene to ameliorate the worst case scenarios, in particular by increasing the care management options available in the community.