Use of a Validated Delirium Assessment Tool Improves theAbility of Physicians to Identify Delirium in Medical Intensive Care Unit Patients

John W. Devlin, Jeffrey J. Fong, Greg Schumaker, Heidi O’Connor, Robin Ruthazer, Erik Garpestad

Crit Care Med 2007 Vol. 35, No. 12

Objective: although medical intensive care unit nurses at our institution routinely use the Intensive Care Delirium Screening Checklist (ICDSC) to identify delirium, physicians rely on traditional diagnostic methods. We sought to measure the effect of physicians’ use of the ICDSC on their ability to detect delirium.

Design: before–after study.

Setting: academic medical center intensive care unit.

Patients and Participants: a total of 25 physicians with >1 month of clinical experience in the medical intensive care unit conducted 300 delirium assessments in 100 medical intensive care unit patients.

Measurements and Main Results: physicians sequentially evaluated  for delirium two patients using whatever diagnostic method preferred. Following standardized education regarding ICDSC use, each physician evaluated two different patients using the ICDSC. Each physician assessment was preceded by consecutive, independent, evaluations for delirium by the patient’s nurse and then a validated judge using the ICDSC. Before (PRE) physician ICDSC use, the validated judge identified delirium in five patients; physicians and nurses identified delirium in zero and four of these patients, respectively. The physicians incorrectly identified delirium in four additional patients. After (POST) physician ICDSC use, the validated judge identified delirium in 11 patients; the physicians and nurses identified delirium in eight and ten of these patients, respectively. The physicians incorrectly identified delirium in one patient. After physician ICDSC use, agreement improved between both the physicians and validated judge (PRE K = - 0.14 [95% confidence interval {CI} = - 0.27 to -0.02] to POST K = 0.67 [95% CI = 0.38 to 0.96]) and physicians and nurses (PRE K = - 0.15 [95% CI =0.29 to - 0.02] to POST K = 0.58 [95% CI = 0.25 to 0.91]). Nurses vs. validated judge agreement was strong in both periods (PRE K = 0.65 [95% CI = 0.29 to 1.00] and POST K = 0.92 [95% CI = 0.76 to 1.00]).

Conclusions: use of the ICDSC, along with education supporting its use, improves the ability of physicians to detect delirium in the medical intensive care unit.

fulltext