Mini-Mental State Examination: new normative values on subjects in Southern Italy

Objective The aims of the present study were: (a) to obtain new normative data of the Italian version of the Mini-Mental Examination State (MMSE) (Measso et al. in Dev Neuropsychol 9:77–85, 1993 ) by administering the tool to a sample of normal Italian individuals more representative of the current Italian population; (b) to compare the sensitivity of this tool in detecting patients suffering from Alzheimer’s disease (AD according to NIA-AA), as compared to the those reported in previous normative Italian studies. Methods MMSE was administered to 314 normal subjects recruited among individuals (and/or their relatives) attending the Offices of General Practitioners (GP) or Memory Clinics in Campania (Italy) by convenience sampling. A group of 47 patients with AD were included into the study. The effect of demographic variables on the raw MMSE scores of normal subjects was checked by multiple linear regression assuming MMSE scores as dependent variable and age, gender and education as the independent one(s). Therefore, a simultaneous regression model was constructed to correct the raw scores according the sensitive variables. Correction grid and equivalent scores were devised to classify subject’s performance. Results The mean raw MMSE score was 27.78 (SD = 1.80) (range 22–30/30). There was no significant difference between scores achieved by men or women ( p = 0.688). Multiple linear regression analysis showed a significant effect of age and years of school attendance on the MMSE raw score, whereas gender did not show any significant effect. The cutoff score, distinguishing between pathological and normal performances, was fixed at the fifth centile corresponding to 24.9/30, higher than the current score of 23.8/30. The new cutoff value was able to identify 44/47 patients with AD, in contrast to 38/47 subjects detected by currently used norms. Conclusions (1) A more updated and representative population sample; (2) a new cutoff threshold able to distinguish between normal and pathological performances; (3) a correction grid that reduces the risk of false-positive and false-negative values due to the influence of the main demographic factors; (4) greater sensitivity, compared to previous Italian normative studies in identifying people with dementia.