Detecting Delirium in Patients With Acute Stroke: A Systematic Review of Test Accuracy

BMC Neurology

Mansutti, I., Saiani, L., et al. (2019).

BMC Neurology, 19(1), 310.

<div>This systematic review examines the accuracy of screening tools in detecting delirium in patients with acute stroke. Of note, although speech-language pathologists (SLPs) do not diagnose delirium, SLPs may administer delirium screening tools as part of a multidisciplinary effort to prevent and manage delirium.</div>

No funding received

From database inception to September 2018

<div>Diagnostic test accuracy studies</div>

4

<div>The 4-Assessment Test for delirium (4AT) demonstrated the highest diagnostic accuracy for detecting delirium in individuals with acute stroke. When compared to the 4AT, the Confusion Assessment Method-Intensive Care Unit (CAM-ICU) demonstrated lower sensitivity and higher specificity. These tools demonstrated the following psychometrics: </div> <ul> <li><strong>The 4AT</strong>: sensitivity = 90.2%-100%, specificity = 64.5%-86%, positive predictive value (PPV) = 43%-86%, negative predictive value (NPV) = 85.6%-100%, internal consistency = 0.80, Area Under the Curve = 0.82-0.89;</li> <li><strong>The CAM-ICU</strong>: sensitivity = 76%, specificity = 98%, PPV = 91%, NPV = 94%,accuracy = 94%, interrater reliability = 0.94, likelihood ratio = 0.47.</li> </ul> <div>Researchers conducted all studies in acute care hospital settings. Limitations to this review include potential bias in patient selection, heterogeneity of test administrators, and lack of longer-term screening results. Future research should include family members and nurses as test administrators. Additional research should investigate the accuracy of tools in the longer-term (e.g., until delirium resolves).</div>

<div>Clinicians should use the following tools with caution when screening for delirium in hospitalized adult stroke patients:&nbsp;</div> <ul> <li><strong>The Abbreviated Mental Test-4 (AMT-4):</strong> sensitivity = 83%, specificity = 61%, PPV = 23%, NPV = 96%;</li> <li><strong>The Abbreviated Mental Test-10 (AMT-10)</strong>: sensitivity = 75%, specificity = 61%, PPV = 21%, NPV = 95%;</li> <li><strong>The Clock Drawing Test (CDT)</strong>: sensitivity = 67%, specificity = 38%, PPV = 7%, NPV = 95%;</li> <li><strong>The COG4</strong>: sensitivity = 70%, specificity = 44%, PPV = 13%, NPV = 92%;&nbsp;</li> <li><strong>The Glasgow Coma Scale</strong>: sensitivity = 17%, specificity = 81%, PPV = 11%, NVP = 88%;</li> <li><strong>Asking "Does this patient have cognitive issues?"</strong>: sensitivity = 58%, specificity = 85%, PPV = 35%, NPV = 93%.</li> </ul> <div>Researchers conducted all studies in acute care hospital settings. This review has some limitations, including potential bias in patient selection, heterogeneity of test administrators, and lack of longer-term screening results. Future research should include family members and nurses as test administrators. Additional research should investigate the accuracy of tools in the longer-term (e.g., until delirium resolves).</div>