Premium
Validation of a short cognitive tool for the screening of dementia in elderly people with low educational level
Author(s) -
de Yébenes María Jesús García,
Otero Angel,
Zunzunegui María Victoria,
RodríguezLaso Angel,
SánchezSánchez Fernando,
Del Ser Teodoro
Publication year - 2003
Publication title -
international journal of geriatric psychiatry
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.28
H-Index - 129
eISSN - 1099-1166
pISSN - 0885-6230
DOI - 10.1002/gps.947
Subject(s) - dementia , confidence interval , intraclass correlation , logistic regression , cognition , medicine , confounding , population , psychology , likelihood ratios in diagnostic testing , gerontology , psychometrics , psychiatry , clinical psychology , disease , environmental health
Aim To validate the ‘Prueba Cognitiva de Leganés’ (PCL) as a screening tool for cognitive impairment in elderly people with little formal education. Methods The PCL is a simple cognitive test with 32 items that includes two scores of orientation and memory and a global score of 0–32 points. It was applied to a population sample of 527 elderly people over 70 with low educational level, who were independently diagnosed by consensus between two neurologists as having normal cognitive function, age associated cognitive decline (AACD, IPA‐OMS criteria) or dementia (DSM‐IV criteria). Individuals with severe visual or hearing defects and those who rejected the exam were excluded from the study. The PCL was validated in a sample of 375 individuals: 300 normal, 42 with AACD and 33 with dementia. The sensitivity, specificity, accuracy and likelihood ratios, as well as the ROC curves for dementia and for AACD‐dementia, were calculated. The confounding effect of sociodemographic variables was assessed by logistic regression analysis and convergent validity by partial correlations of the PCL with other cognitive tests. Inter‐rater reliability was evaluated with the intraclass correlation coefficient. Results The PCL identified dementia (cut‐off ≤22) and AACD‐dementia (cut‐off ≤26), with the following diagnostic parameters, respectively: sensitivity 93.9%–80%, specificity 94.7%–84.3%, positive likelihood ratio 17.8–5.1, negative likelihood ratio 0.06–0.24, and accuracy 94.6%–83.4%. The areas under the ROC curve were 0.985 (95% Confidence Intervals (CI) 0.967–0.995) and 0.904 (95% CI: 0.870–0.932) respectively. The intraclass correlation coefficient was 0.79 (0.74–0.83). Conclusion The PCL is a simple instrument, which is both valid and reliable, for the screening of dementia in population samples of individuals with low educational level. This instrument could be useful in primary health care. Copyright © 2003 John Wiley & Sons, Ltd.