Attention, arousal and other rapid bedside screening instruments for delirium in older patients: a systematic review of test accuracy studies

Objective: delirium occurs frequently in frail patients but is easily missed. Screening with a rapid, easy-to-use and highly sensitive instrument might help improve recognition. The aim of this study was to review attention, arousal and other rapid bedside screening instruments for delirium in older patients. Methods: a literature search was performed in PubMed, PsycINFO and Embase. We scrutinized forward citations in Google Scholar, and references of included articles and prior reviews. We included studies among older patients that investigated the sensitivity and specificity of delirium screening instruments that could be administered in 3 min or less, and did not require surrogate information. We extracted study characteristics, risk of bias, sensitivity and specificity. Results: we identified 27 studies among 4,766 patients in hospitals and nursing homes. They tested many different single and several combined screening instruments. Prevalence of delirium varied between 4% and 57%. Only one study scored a low risk of bias on all domains. Sensitivity varied between 17% and 100%, and specificity between 38% and 99%. Of the 22 tests with sensitivity ≥90%, seven also had specificity ≥80% in older patients in general. These results were approximately reproduced for the Observational Scale of Level of Arousal (OSLA) and Richmond Agitation and Sedation Scale (RASS): sensitivity and specificity were >80%. Conclusion: two arousal tests-OSLA and RASS-had reproduced high sensitivity and specificity in older patients. Nurses can administer these tests during daily interaction with patients. Test accuracy studies about rapid screening tools for delirium superimposed on dementia were scarce