Bedside tests of attention and organized thinking were performed in patients with cognitive impairment or dementia but without delirium, to provide estimates of false positive rates for detecting delirium superimposed on dementia (DSD).

This cross-sectional study was conducted in outpatients and institutionalized patients without delirium representing a wide spectrum of severity of cognitive impairments.

Patients with dementia or a cognitive disorder according to DSM IV criteria, after exclusion of (suspected) delirium according to DSM IV criteria.

Tests for inattention and disorganized thinking from the CAM-ICU were assessed.

The sample included 163 patients (mean age 83 years (SD 6; 64% women)), with Alzheimer's disease as most prevalent (45%) diagnosis and a mean MMSE-score of 16.8 (SD 7.5). False positive rates of the test of attention varied from 0.04 in patients with normal to borderline cognitive function to 0.8 in those with severe dementia. The false positive rate of the test of disorganized thinking was zero in the normal to borderline group, increasing to 0.67 in patients with severe dementia. When combining test results false positive rates decreased to 0.03 in patients with MMSE scores above 9.

Use of simple bedside tests of attention and organized thinking for the clinical diagnosis of DSD will result in high rates of false positive observations if used regardless of the severity of dementia. However, if test results are combined they may be useful to exclude DSD in patients with minimal to moderate degrees of dementia, but not in the severe group.