Exploring the utility of ultra-brief delirium assessments in a nonintensive care geriatric population: the GEM study.

OBJECTIVE To determine how an ultra-brief structured tool that would require usually less than a minute for delirium assessment compares with a clinical assessment based on Diagnostic and Statistical Manual-IV (DSM-IV) in a geriatric postacute care (PAC) rehabilitation unit. DESIGN Prospective observational cohort study. SETTING Postacute geriatric hospital ward of a Veteran's Affairs hospital. PARTICIPANTS Consecutively admitted patients between 50 and 100 years old for inpatient postacute medical care. MEASUREMENTS Two teams, blinded to one another's evaluations, performed daily delirium assessments using either the Confusion Assessment Method for the intensive care unit (CAM-ICU) or clinical assessment based on DSM-IV. RESULTS There were 61 patients enrolled (median 73 years old, range: 52-94), who underwent 521 paired observations. Delirium was detected in 18 patients (29.5%) by one of the two screening methods over the course of the study, most of whom (14 patients, 23%) were delirious on the first day of enrollment. Delirium was identified by the CAM-ICU on 12.6% of the observations and by the clinical assessment on 6% of the observations (κ = 0.25, 95% confidence interval [CI]: 0.09, 0.40). Examination of disagreement between the 2 evaluations revealed that patients with dementia (κ = 0.11, 95% CI: -0.14, 0.27) had 10.7 times higher odds (95% CI: [3.1, 36.8], p value < .001) of having discordance than patients without dementia. CONCLUSIONS Different delirium assessments may disagree depending on the study population. Dementia patients are especially challenging to evaluate for delirium.