There is increasing evidence that tighter control of glucose levels in critically ill patients improves patient outcomes. OBJECTIVES: To determine if education and implementation of a continuous insulin infusion protocol during 2002 improved glucose control in the intensive care setting. METHODS: A retrospective chart review was conducted on ICU patients during a four-month period in one of three consecutive years. Patients were included if they were in the ICU for a minimum of three days, were ventilated, received care from an intensivist, and had a medical DRG at discharge. Additional information was gathered from electronic charge data. Alpha for all analyses was set at 0.05. RESULTS: The following number of charts was reviewed for each time period: 22 in 2001, 51 in 2002, and 31 in 2003. There were no significant differences in patient age, gender, marital status, race, diabetes as a secondary diagnosis, or hospital or ICU length of stay. Regression analysis demonstrated that median glucose values per patient day were highest in 2001 (164.1 mg/dL), compared to 2002 (136.1 mg/dL) and 2003 (127.7 mg/dL). The number of documented glucose values was lowest in 2001 [2.4 per pt day (sd 2.5)], and highest in 2002 [5.8 per pt day (sd 5.4)], compared to 2003 patients [4.3 per pt day (sd 5.2)]. CONCLUSION: Efforts to manage glucose in the critically ill appear effective as evidenced by a higher number of documented values/day and a decrease in median glucose values per patient day.