[Concordance among Mini-Examen Cognoscitivo and Mini-Mental State Examination in cognitive impairment screening]

To compare the concordance between two cognitive impairment (CI) screening tests: the 30-point Folstein's Mini Mental State Examination (MMSE) and their validated and slightly modified spanish translation, the 35-point Lobo's Mini-Examen Cognoscitivo (MEC).

Analytical cross-sectional multicenter study.

Primary care.

Randomized sample of subjects aged 65 years or more attended in 54 primary care centers in Catalonia. Inclusion of 3,167 subjects. Institutionalized patients were excluded.

After a training period, primary care doctors and nurses administered MMSE and MEC simultaneously to their own patients. Standardization of variables.

Age 74 (6.1) years-old, 1,611 (50.9%) women, 954 (30.1%) illiterate or without primary education. Prevalence of CI according to MMSE was 16.52% (n = 449) and according to MEC 5,49% (n = 165) (P<.001). Although intraclass correlation coefficient was 0,864 (95% CI, 0.855-0.873), the kappa index at score 24 for both tests was 0.468, but it increased up to 0.788 taking 20 and 23 scores for MMSE and MEC, respectively. At multivariate analysis, low educational level, and age more than 80 years-old predict a bad concordance among both tests.

In spite of good correlation between MMSE and MEC, both detect different CI prevalences with the cut-off point at 23/24. Their agreement is only moderated in practice, because we interpret tests in a dichotomic way (CI versus no-CI). In our experience, we cannot use them indistinctly with the cut-off point at 23/24, especially in subjects with low educational level or aged 80 or more years.