IntroductionDelirium is a prevalent condition in patients admitted to intensive care units (ICU) associated with worse outcomes. The principal aim of the present study was compare the agreement between two tools for delirium assessment in medical and surgical patients admitted to the ICU.MethodsConsecutive adult surgical and medical patients admitted to the ICU for more than 24 hours between March 2009 and September 2010 were included. Delirium was evaluated twice a day using the Intensive Care Delirium Screening Checklist (ICDSC) and Confusion Assessment Method adapted to the Intensive Care Unit (CAM-ICU). The kappa (k) and AC1 coefficients were calculated as a measure of agreement between the CAM-ICU and ICDSC.ResultsA total of 595 patients were enrolled in the study. There were 69 (12%) emergency surgical, 207 (35%) elective surgical and 319 (54%) medical patients. Delirium incidence evaluated by the ICDSC, but not by the CAM-ICU, was similar among the three groups. Overall agreement between CAM-ICU and ICDSC was moderate (k = 0.5) to substantial (AC1 = 0.71). In medical patients the agreement between the two instruments was moderate (k = 0.53) to substantial (AC1 = 0.76). The agreement between the two tools in emergency surgical patients was also moderate (k = 0.53) to substantial (AC1 = 0.68). In elective surgical patients the agreement between the two instruments was low (k = 0.42) to substantial (AC1 = 0.74).Agreement rates seemed to be influenced by disease severity. The agreement rate in the general ICU population with APACHE II = <14 was k = 0.57 and AC1 = 0.81, compared to k = 0.44 and AC1 = 0.59, in patients with more severe disease. This was even more different when the need for mechanical ventilation was used as a surrogate of disease severity.ConclusionsThe agreement rates between CAM-ICU and ICDSC may vary between different groups of ICU patients and seems to be affected by disease severity.