MMSE or ACE-R in cognitive assessment in primary care?

Abstract

Background: The growing prevalence of neurocognitive disorders in aging populations highlights the need for reliable cognitive screening tools in primary care. The Mini-Mental State Examination (MMSE) and the Addenbrooke’s Cognitive Examination Revised (ACE-R) are widely used instruments, yet their comparative diagnostic performance and susceptibility to sociodemographic influences remain clinically relevant issues. Objective: To compare the diagnostic performance of MMSE and ACE-R in primary care, evaluating total and domain-specific scores, the influence of age and education, group discrimination capacity, and sensitivity and specificity parameters. Methods: This observational, descriptive, and analytical study included 246 individuals aged 50–93 years assessed in a primary care setting. Participants underwent MMSE and ACE-R administration. Statistical analyses included descriptive measures, Pearson’s correlation, Kruskal–Wallis test with Bonferroni-adjusted post hoc comparisons, and receiver operating characteristic (ROC) curve analyses to determine sensitivity and specificity adjusted for age and education (SPSS v29.0; p<0.05). Results: Mean age was 67.02±9.28 years and mean education was 7.39±5.16 years. The mean total score was 68.20±17.60 for ACE-R and 24.52±4.18 for MMSE. No significant correlation was observed between total scores (r=0.075; p=0.244), although significant correlations were found between equivalent domains: attention (r=0.403), memory (r=0.595), language (r=0.628), and visuospatial abilities (r=0.474), all p<0.001. ACE-R total scores were not correlated with age (p=0.929) and showed a weak positive correlation with education (r=0.145; p=0.023). In contrast, MMSE demonstrated a moderate negative correlation with age (r=−0.338; p<0.001) and a moderate positive correlation with education (r=0.476; p<0.001). ACE-R significantly differentiated groups in attention (p=0.029) and memory (p=0.030), whereas MMSE showed highly significant discrimination in total scores across diagnostic categories (p<0.001). MMSE exhibited high sensitivity in illiterate individuals (97.7%) and high specificity in individuals with >11 years of education (97%). ACE-R demonstrated higher specificity in highly educated individuals (79.8%) but lower sensitivity (21.4%). Conclusion: MMSE demonstrated greater global sensitivity for cognitive screening in primary care, particularly in advanced stages, although significantly influenced by age and education. ACE-R showed superior domain-level discrimination, particularly in attention and memory, and higher specificity. These findings suggest complementary roles for both instruments, and their combined use with education-adjusted cutoffs may optimize diagnostic accuracy in primary care settings.