Education-Adjusted Interpretation of MMSE for Diagnostic Validity of Cognitive Disorders

Abstract

Introduction: The Mini-Mental State Examination (MMSE) is widely utilized in clinical settings for cognitive screening, yet its diagnostic accuracy is often influenced by demographic factors such as educational attainment. This study investigates the educational gradient in MMSE performance and evaluates whether uniform cutoff scores adequately distinguish cognitively normal (CN), mild cognitive impairment (MCI), and Alzheimer&apos;s disease (AD) patients across different educational strata. Methods: A total of 300 older adults (CN = 100; MCI = 100; AD = 100) were retrospectively recruited from the Severance Hospital memory clinic, intentionally balanced to ensure statistical power and avoid class-imbalance bias across diagnostic groups. All participants completed the Korean version of MMSE and the Seoul Neuropsychological Screening Battery-II (SNSB-II) and underwent 3T brain MRI for hippocampal volumetry. Education level was categorized as low (<= 6 years), medium (7-12 years), and high (>= 13 years). MMSE diagnostic accuracy was evaluated using receiver operating characteristic (ROC) curve analyses stratified by education. Interaction effects were tested via multiple linear regression, and correlations with hippocampal volume were assessed. Results: MMSE scores showed a significant educational gradient, with higher education associated with higher performance (p < 0.001). MMSE scores demonstrated a pronounced educational gradient, with particularly reduced performance in individuals with low educational attainment, suggesting potential overestimation of cognitive impairment when uniform MMSE cutoffs are applied. ROC analyses revealed only moderate diagnostic accuracy of MMSE in the higher education groups (area under the curve [AUC] = 0.83 and 0.78). The AUC was 0.73 (95% CI: 0.58-0.88) in the low-education group; the AUC was 0.83 (95% CI: 0.75-0.91) in the middle-education group and 0.78 (95% CI: 0.70-0.87) in the high-education group, suggesting only moderate diagnostic accuracy of MMSE. Conversely, lower education groups showed underperformance potentially unrelated to pathology. Regression models confirmed that education and diagnosis had additive but noninteracting effects on MMSE scores. MMSE correlated strongly with hippocampal volume (r = 0.739, p < 0.001), validating its general neuroanatomical relevance. Conclusion: MMSE performance is substantially modulated by education, with uniform cutoffs yielding differential diagnostic validity across educational strata. We suggest education-adjusted interpretation of MMSE and emphasize the need for integrative diagnostic approaches combining cognitive testing with neuroimaging biomarkers.