Dementia Overdiagnosis in Younger, Higher Educated Individuals Based on MMSE Alone: Analysis Using Deep Learning...

Area of Science:

• Neuroscience
• Gerontology
• Psychometrics

Background:

• Dementia diagnosis requires accurate cognitive assessment for effective management.
• The Mini-Mental State Examination (MMSE) is a common but debated tool for assessing cognitive impairment.
• This study aimed to improve dementia diagnostic accuracy by comparing MMSE alone with combined cognitive assessments.

Purpose of the Study:

• To evaluate the diagnostic effectiveness of the MMSE compared to a combination of MMSE and additional cognitive tests.
• To refine dementia diagnosis by assessing the impact of supplementary cognitive assessments.
• To identify demographic factors influencing diagnostic reclassification.

Main Methods:

• Developed two random forest models: one using MMSE scores only, and another incorporating additional cognitive test data.
• Evaluated model performance using accuracy, precision, recall, F1-score, and AUC on a large cohort (n=2,863) with subjective cognitive complaints.
• Utilized a 70:30 training-to-testing data split for model validation.

Main Results:

• The MMSE-alone model achieved 86% accuracy and an AUC of 0.872.
• The expanded model integrating additional tests showed improved accuracy (88%) and AUC (0.934).
• A significant 17.46% of cases were reclassified from dementia to non-dementia with the expanded model, associated with higher education and younger age.

Conclusions:

• The MMSE is a valuable initial screening tool but insufficient for definitive dementia severity assessment.
• Integrating diverse cognitive assessments substantially enhances diagnostic precision, particularly for younger and more educated populations.
• Future dementia diagnostic protocols should incorporate multifaceted cognitive evaluations to capture the disorder's complexity.