Premium
Screening for dementia in an Irish community sample using MMSE: a comparison of norm‐adjusted versus fixed cut‐points
Author(s) -
Cullen Breda,
Fahy Sabina,
Cunningham Conal J.,
Coen Robert F.,
Bruce Irene,
Greene Elaine,
Coakley Davis,
Walsh J. Bernard,
Lawlor Brian A.
Publication year - 2005
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.1291
Subject(s) - percentile , dementia , irish , norm (philosophy) , cut off , cut point , cognitive impairment , medicine , gerontology , sample (material) , psychology , cognition , statistics , mathematics , psychiatry , linguistics , philosophy , power (physics) , disease , physics , quantum mechanics , political science , law , chromatography , chemistry
Background The MMSE is a widely‐used instrument in screening for dementia in the community. The traditional cut‐point of <24/30 may be unsuitable for use in some settings, due to biases in age and education. Objectives This study aimed to investigate whether cut‐points derived from age‐ and education‐specific norms would improve the performance of the MMSE as a screening tool for dementia and cognitive impairment in an Irish community setting, potentially reducing the number of onward referrals. A secondary aim was to obtain score norms in an Irish sample. Methods One thousand one hundred and fifteen people aged over 65 (mean age 74.8±6.8 years; 68% female) were assessed in their own homes, using MMSE (world) and the AGECAT computerised diagnostic system for mental illness in the elderly. The performance of the MMSE in identifying case‐ and sub‐case‐level dementia was investigated using cut‐points of <24 and <23, as well as adjusted cut‐points based on published norms and norms derived from the Irish sample. Results Published norms did not prove useful in improving screening accuracy. Cut‐points based on 10th percentile Irish norms and the <23 cut‐point performed comparably well, both yielding 93% specificity, with overall accuracy of 90% and 91% respectively (as sensitivity was higher for the <23 cut‐point). Conclusions Locally‐derived norms yielded better screening accuracy than did published norms or the traditional <24 cut‐point. The importance of selecting an appropriate percentile cut‐off when using norms is discussed. The best results were obtained with a simple <23 cut‐point, and this may be optimal when screening for dementia in an Irish community setting. Copyright © 2005 John Wiley & Sons, Ltd.