Two bedside tests are fairly accurate in screening for delirium in older patients

Reference

Quispel-Aggenbach DWP, Holtman GA, Zwartjes H, Zuidema SU, Luijendijk HJ. Attention, arousal and other rapid bedside screening instruments for delirium in older patients: a systematic review of test accuracy studies. Age Ageing 2018;47(5):644-653.

Synopsis

These authors systematically searched multiple databases to identify quick and simple bedside tests to screen for delirium in persons older than 60 years. They also searched reference lists of included papers and other relevant studies, searched Google Scholar, and queried the authors of included studies. Two authors independently assessed papers for inclusion, extracted data, and assessed each study's risk of bias. Ultimately, they identified 27 studies published in 31 papers. These studies included between 14 and 500 patients with variable rates of delirium (between 4% and 57%). The included studies evaluated more than 40 separate tests, with each study assessing up to 20 separate screening tools, the most common of which was a backward recitation of the months of the year. The study settings were mainly geriatric, general medical, or surgical wards of hospitals and emergency departments. One study took place in a psychiatry liaison consultation service, 1 in a hospice, and 4 in nursing homes. Only one study was of low risk of bias; 14 were of high risk of bias. The authors don't describe how the diagnosis of delirium was established. The informed consent process in the included studies may have resulted in excluding patients with severe delirium or severe dementia. Most of the tests evaluated attentiveness or levels of arousal and varied between 7 seconds and 3 minutes to administer. The tests require less than 45 minutes of training and can be administered by staff with minimal clinical experience. Although they chose not to pool data to generate summary estimates, the authors report the sensitivity and specificity ranged from 17% to 100% and from 38% to 99%, respectively. Of the 22 screening tests that were more than 90% sensitive, 7 were also more than 80% specific. Of these, only 2 appeared to be reproducible: the Observational Scale of Level of Arousal and the Richmond Agitation and Sedation Scale.