Icu Bedside Assessments of Delirium: Sustainability and Reliability

Background:

Delirium affects 50% of intensive care unit (ICU) patients and is associated with increased mortality and long‐term cognitive impairment. Instruments can accurately measure delirium; however, the sustainability and reliability of delirium measurement tools are unknown in usual practice. We therefore examined delirium assessments by ICU bedside nurses during routine clinical care compared with concurrent measurements made by research personnel.

Methods:

This prospective cohort study included patients admitted to medical or surgical ICUs between 2007 and 2010 at a tertiary‐care teaching hospital. We excluded patients with states preventing delirium assessment (e.g., severe baseline cognitive impairment, inability to understand English). Delirium was independently measured by bedside nurses and reference‐rater research nurses using the Confusion Assessment Method for the Intensive Care Unit (CAM‐ICU) and the Richmond Agitation‐Sedation Scale (RASS), a component of the CAM‐ICU. CAM‐ICU and RASS bedside nurse assessment frequency was calculated as the mean number of assessments for each patient‐day. For each research nurse CAM‐ICU and RASS assessment, we selected the closest bedside nurse assessment that was no more than 4 hours apart. We assessed agreement with weighted kappa statistics. To describe the nature of disagreement, we calculated sensitivity and specificity of bedside nurse assessments of delirium compared with the research nurse assessments.

Results:

Five hundred and ten patients were assessed. Their median age was 58 (IQR, 47‐67), 12% had cognitive impairment prior to admission, and the median APACHE II score at enrollment was 27, indicating a high severity of illness. Bedside nurses assessed delirium (CAM‐ICU) and sedation (RASS) an average of 7.4 (SD, 2.9) and 7.8 (SD, 3) times per patient‐day, respectively. There were 7156 CAM‐ICU and RASS paired assessments within 4 hours; 98% of these occurred < 2 hours apart. Across the entire population, there was substantial agreement between bedside and research nurses (CAM‐ICU weighted kappa, 0.67; 95% CI, 0.66‐0.70; RASS weighted kappa, 0.66; 95% CI, 0.64‐0.68). Agreement was stable regardless of severity of illness, preexisting cognitive impairment, and study year. The sensitivity and specificity of delirium nurse assessments was 0.81 (95% CI, 0.78‐0.83) and 0.81 (95% CI, 0.78‐0.85), respectively, corresponding to a positive and negative likelihood ratio of 4.3 (95% CI, 3.6‐5.2) and 0.2 (95% CI, 0.2‐0.3), respectively.

Conclusions:

We have demonstrated that delirium and sedation measurements performed by bedside nurses are sustainable and a reliable source of information. These measures can be used for clinical decision making, quality improvement, and quality measurement activities.

Disclosures:

E. Vasilevskis ‐ none; A. Morandi ‐ none; L. Boehm ‐ Hospira Inc., honoraria; P. Pandharipande ‐ Hospira Inc. and GlaxoSmithKline, honoraria; T. Girard ‐Hospira Inc., honoraria; J. Jackson ‐ none; J. Thompson ‐ none; A. Shintani ‐none; B. Pun ‐ Hospira Inc., honoraria; E. Ely ‐ Eli LIlly, Pfizer, Hospira Inc., Aspect Medical Systems, GlaxoSmithKline, grant support and honoraria