BACKGROUND: Review of resource use and patient outcomes of intensive care unit services over time provides insights into service delivery and safety. OBJECTIVE: To examine temporal trends in resource consumption and risk-adjusted mortality of adult ICU patients in Victoria. DESIGN, PARTICIPANTS AND SETTING: Retrospective cohort study of 214 619 adult ICU admissions recorded from 23 major hospitals over 12 years from 1 July 1999 to 30 June 2011. OUTCOMES: Primary outcomes were population rates of ICU admission and mechanical ventilation (MV), ICU and hospital length of stay, and hospital survival. Secondary outcomes included average ICU and MV bed numbers. Administrative data were derived from the Victorian Admitted Episodes Dataset and the Australian Bureau of Statistics. The Critical Care Outcome Prediction Equation informed estimates for risk-adjusted mortality. Temporal mortality trends were evaluated for outcome estimates and hierarchical logisticregression trends were evaluated for risk-adjusted mortality. RESULTS: Of ICU admissions, 104 103 (48.5%) were patients who received MV, and 87.6% ICU admissions were adults who survived to hospital discharge. There was a decline in the risk-adjusted mortality (odds ratio, 0.967 per year; 95% CI, 0.963-0.971; P<0.001). Similar results were found in 17 hospitals (74%) and in nine of 10 major diagnostic subgroups. There was an increase of 5.2 occupied ICU beds per year (range, ?4.2 ICU beds per year; P=0.002). Despite ICU admissions being a minority cohort (2.5% of public hospital admissions) this group used 8.6% of hospital bed-days and attracted 19.5% of funding. CONCLUSIONS: There was an increase in ICU resource availability and evidence of improvement in hospital survival, suggesting improved quality of care. These evaluation methods may be useful in monitoring statewide capacity, service delivery and patient safety.