The first goal of this investigation was to identify individuals with delirium defined by the Confusion Assessment Method for the Intensive Care Unit (CAM-ICU) among medical patients with respiratory failure. Our second goal was to compare clinical interventions including use of continuous sedation infusions, the number of ventilator-free days, ICU length of stay, hospital mortality, and use of physical restraints in mechanically ventilated patients with and without delirium. A prospective, single-center, observational cohort study. The medical intensive care unit (19 beds) of an urban teaching hospital. Adult, intubated, and mechanically ventilated patients. Daily evaluation with the CAM-ICU, outcomes assessment, and prospective data collection. Among 93 patients evaluated using the CAM-ICU, 44 patients (47%) developed delirium (CAM-ICU+) for >/=1 day while in the intensive care unit. Twenty-two patients (24%) had no episodes of delirium recorded (CAM-ICU-), and 27 (29%) remained comatose until extubation or death. A statistically greater number of patients with delirium (CAM-ICU+) received continuous infusions of midazolam (59% vs. 32%, p < .05) or fentanyl (57% vs. 32%, p < .05) and physical soft-limb restraints (77% vs. 50%, p < .05) compared with patients without delirium (CAM-ICU-). The identification of delirium using the CAM-ICU was associated with greater use of continuous sedation infusions and physical restraints. Additional studies are required to determine how the use of these specific interventions influences the occurrence and the natural history of delirium among critically ill patients.