RESUMEN
The aim of this research was to examine the validity of a brief screen for cognitive impairment (BSCI) consisting of three questions administered by telephone (delayed recall, frequency of help with planning trips for errands, and frequency of help remembering to take medications). The study design was an age and gender matched case-control study. Seventy managed care members, 35 with dementia (cases) and 35 without dementia (controls), were assessed using BSCI embedded within a longer health assessment questionnaire commonly used in Medicare-managed care. A number of measures were used to examine validity of BSCI, including comparisons of the differences between cases and controls in BSCI scores, comparisons of the correlations between patient scores on BSCI and the Mini Mental Status Exam (MMSE, a common screening test for dementia) and the Alzheimer's Disease Assessment Scale (ADAS, a common dementia assessment test), and comparisons of the areas under the receiver operating characteristic (ROC) curves for the three instruments. BSCI scores for cases and controls were significantly different, as were their scores for the MMSE and ADAS. Scores on BSCI were significantly correlated with scores for the MMSE and ADAS using both the Kendall's tau-b and Spearman rank-order correlation; correlations ranged from 0.654 between BSCI and ADAS to -0.83 for the correlation between BSCI and the MMSE (p < 0.001 for both). The areas under the ROC curves ranged from 0.94 to 0.96 for the three tests, meaning that they were equally accurate in discriminating between demented and nondemented patients. BSCI, a brief telephone screen for cognitive impairment due to dementia, discriminates between demented patients and normal controls as well as two standard tests of dementia, and may be considered a valid screen for dementia. Compared to existing screening tests, it has the additional advantages of extreme brevity, and ease of administration and scoring by lay interviewers via telephone. The use of brief screening instruments for dementia, such as the one validated here, will be increasingly important for the effective management of dementia and other chronic diseases where dementia is a coexisting condition.
