Delirium screening in the neurointensive care unit: Comparison of the confusion assessment method for the intensive care unit and intensive care delirium screening checklist

Background: Delirium is a common and serious potential complication of an ICU stay, associated with risk of mortality and worse outcomes. Screening tools including ICDSC and CAM-ICU have been validated in critical care populations, however, few studies have examined their utility in neurocritically ill patients. Research on delirium is highly segregated with varying terminology used among specialties, making research efforts difficult.

Objectives: Evaluate which screening tool is more feasible in neurocritically ill patients.

Methods: Single-center, retrospective analysis of a prospective QI initiative in a neurosciences ICU at a tertiary medical center in Chicago, IL. Patients admitted January 2019-2020 were screened for delirium with ICDSC, CAM-ICU and "gold-standard" DSM-5 throughout ICU stay. 206 patients were included who underwent 1442 assessments. Agreement of detection tools assessed using Fleiss' kappa.

Results: Prevalence of delirium using DSM-5 criteria was 1.6 %. Of the 1442 assessments, 116 were CAM-ICU-positive (8 %) and 343 were ICDSC-positive (23.8 %). The three tools had a kappa-agreement of 0.249. Of the 23 patients with confirmed delirium per DSM-5, 10 (4.8 %) were also positive for ICDSC but negative for CAM-ICU. 7 patients with a positive DSM-5 diagnosis were negative for both the CAM-ICU and ICDSC. CAM-ICU: sensitivity 26 %, specificity 94.5 %. ICDSC: sensitivity 69.5 %, specificity 87.6 %.

Conclusions: Current methods for detection of delirium in the neurocritically ill are imperfect, and terminology bias exists in the literature, limiting data comparison. While limited, the ICDSC appears a more sensitive tool for detection of delirium in the neurocritically ill compared to CAM-ICU. Larger validation studies are warranted.