Evaluation of Two Intensive Care Delirium Screening Tools for Non-Critically Ill Hospitalized Patients

Background Delirium is a common, yet frequently under-recognized medical/psychiatric complication for hospitalized patients, associated with substantial morbidity and mortality. While easy-to-use delirium screening tools exist for ventilated patients, their use in non-critically ill, hospitalized patients has not been validated. Methods This prospective 4-week comparison of daily delirium status, using screening tools (Confusion Assessment Method for the Intensive Care Unit [CAM-ICU] and Intensive Care Delirium Screening Checklist [ICDSC]) vs. a daily neuropsychiatric examination as a reference standard, was conducted in 139 inpatients in two medical oncology units at a large teaching hospital during July, 2009. Results Based on neuropsychiatric examination, 36 (26%) patients had at least 1 day of delirium during their hospital admission. For 21 (15%) patients, delirium was present at the initial assessment, while 15 (11%) developed delirium at a median (IQR) of three (2–7) subsequent assessments. Delirium was present on 20% of all patient-days. For the initial evaluation, the CAM-ICU had a sensitivity of 18% (95% confidence interval [CI], 5%–44%), and a sensitivity of 18% (9% –32%) when using all assessments, adjusting for repeated measures on each patient. The ICDSC had sensitivities of 47% (24%–72%) and 62% (49%–74%). The specificity of both instruments was very high (≥98%). Conclusions This study suggests that in non-critically ill hospitalized patients, the CAM-ICU and ICDSC intensive care delirium screening tools are not adequately sensitive for use in routine clinical practice. Further work is needed to develop more sensitive, efficient tools in this population.