Screening for Delirium Using the Electronic Medical Record

Abstract

Background: Delirium is characterized by an acute disturbance in cognition and attention. It is common and disproportionately effects elderly hospitalized adults. It is associated with increased risk of dementia, increased mortality, and increased healthcare costs. Early detection and identification of those at highest risk for delirium is crucial to mitigate these adverse effects. The current study seeks to determine the validity of a novel screening tool: the Neurological Orientation, Verbal response Assessment of Delirium (NOVAD) to detect delirium utilizing Electronic Medical Record (EMR) data. Methods: This is a retrospective observational study of EMR data extracted from hospitalized patients at a large multi-state hospital system. Data for the NOVAD score was documented by nurses as part of routine assessments and extracted from the patient's EMR. This was then compared to the patient’s Confusion Assessment Method (CAM) score as recorded by nursing as well as whether a physician had documented the patient as having Delirium or Encephalopathy by International Classification of Disease version 10 (ICD-10) code. Sensitivity, specificity, positive predictive value, and negative predictive value were calculated by comparing NOVAD detection of delirium to CAM and delirium or encephalopathy ICD-10 code. Results: 464,395 participants were included in this study. The mean age of study participants was 56.18 years (SD 20.82) and the sample was 56.17% female (n = 260,856). The prevalence of participants in the study with a positive NOVAD score was 29.23% (n = 135,776; 95% CI: 0.2911, 0.2937). Nursing administered modified CAM score was documented positive in 2.41% (n = 11,183; 95% CI: 2.31%, 2.45%) of participants. NOVAD was found to be 86.63% sensitive (95% CI 86.53, 86.73) and 78.08% specific (95% CI: 77.96 ,78.20) based on ICD-10 code diagnosis of delirium or encephalopathy as being truly positive for delirium. Conclusion: The current study’s results suggest that NOVAD is a sensitive early warning screening tool for delirium in hospitalized adults. Given poor specificity the tool is best suited for identifying patients who would benefit most from specialist care teams.