Background Successful blood glucose control is associated with improved outcomes of critically ill patients. However, insulin treatment can cause hypoglycemia, an important patient safety concern. The Joint Commission has recommended that all episodes of hypoglycemia be evaluated with a root cause analysis. Objective To reduce episodes of hypoglycemia through the analysis of data related to each episode. Methods The interdisciplinary team of the 16-bed critical care unit of a university-affilited teaching hospital developed a process to analyze, in real time, each episode of hypoglycemia (blood glucose level <60 mg/dL), including evaluation of patient risk factors and nursing interventions. The nursing staff integrated the root cause analysis into daily practice. The preimplementation period encompassed 2429 consecutive admissions, and the implementation period encompassed 2608 consecutive admissions. Results The percentage of patients with hypoglycemia decreased substantially during the implementation period among those without (from 6.15% to 3.78%; P = .001) and with diabetes (from 13.14% to 7.23%; P = .002). Mean blood glucose level decreased during the implementation period among patients without diabetes (P < .001), and did not change significantly among patients with diabetes (P = .23). The coefficient of variation, reflecting glucose variability, decreased during the implementation period among patients without and with diabetes (P < .001 for each). Conclusion The nurse-driven root cause analysis was associated with a substantial reduction in hypoglycemia, with concomitant decreases in mean blood glucose level among patients without diabetes and glucose variability in patients without and with diabetes.