Routine use of the Confusion Assessment Method for the Intensive Care Unit (CAM-ICU) by bedside nurses may underdiagnose delirium

Background: The Confusion Assessment Method for the Intensive Care Unit (CAM-ICU) is emerging as the most frequently used tool for identifying delirium among critically ill patients. Objective: To determine whether the number of patients and nursing shifts in which delirium was diagnosed would increase after the introduction of the CAM-ICU in our unit. Design: Before-and-after study. In a 30-day Phase 1, we asked bedside nurses to assess their ICU patients for delirium each shift. We then conducted intensive education on the CAM-ICU for 30 days, including lectures, bedside tutorials, and supervised practice. In Phase 2, for 30 days we asked bedside nurses to record the results of their CAM-ICU assessments. Setting: 20-bed mixed medical and surgical ICU at the Austin Hospital, Melbourne. Participants: All patients admitted to the ICU during each phase. Main outcome measures: Diagnosis of delirium by bedside nurses using either the CAM-ICU or an unstructured clinical assessment, by patient and nursing shift. Results: Compared with unstructured assessments, the CAM-ICU identified a significantly lower proportion of patients (36.7% v 21.3%; P = 0.004) and a significantly lower proportion of shifts (14.7% v 6.4% of shifts, P = 0.002) with delirium. When adjusted for differences in age, sex, Acute Physiology and Chronic Health Evaluation III risk of death and total length of stay between the two periods, assessment type remained a significant predictor of the diagnosis of delirium. Conclusions: In our hospital, the CAM-ICU detected delirium less often than unstructured delirium assessments made by qualified intensive care nurses.