JMIR Publications

ABSTRACT

Background:

As the global population ages, the prevalence of cognitive impairment increases significantly, highlighting the critical need for early and accurate diagnostic tools. The Mini-Mental State Examination (MMSE) is one of the most commonly used cognitive screening instruments to detect signs of cognitive decline. Prior research has demonstrated a notable correlation between deteriorated oral health and lower MMSE scores, suggesting that oral health could serve as a non-invasive, accessible indicator of cognitive health.

Objective:

This study aims to explore the potential of utilizing Machine Learning (ML) technologies using oral health and demographic examination data to predict the probability of having MMSE scores 30 or ≤26 in Swedish individuals over 60 years of age.

Methods:

Data from two longitudinal oral health and ongoing general health studies Support Monitoring and Reminder Technology for Mild Dementia (SMART4MD), and the Swedish National Study on Aging and Care (SNAC-B), involving individuals over 60 years of age were entered into ML models, including Random Forest (RF), Support Vector Machine (SVM), and CatBoost (CB) to classify MMSE scores as either 30 or ≤26, distinguishing between MMSE 30 and MMSE ≤26 scores groups. To ensure robust model performance and to reduce the risk of overfitting, a nested cross-validation (nCV) strategy with 2 to 10 inner and outer loop configurations was used during model training and evaluation. The best performance-giving model was further investigated for feature importance using Shaply Additive explanation (SHAP) summary plots to easily visualize the contribution of each feature and to determine the relative importance of each input feature in the prediction process. The sample consisted of 693 Individuals (350 females and 343 males).

Results:

: All three ML classifiers, CB, RF, and SVM achieved high classification accuracies in differentiating between MMSE score groups. However, CB exhibited superior performance with an average accuracy of 81.0% on the model employing 3X3 nCV and surpassed the performance of both RF and SVM models. SHAP summary plots provided valuable insights into the most influential predictors contributing to the model’s decisions. Key factors included age, plaque index (PI), periodontal pocket depth (PPD), self-reported sensation of dry mouth, educational level, and the use of dental hygiene tools for approximal cleaning. These variables collectively played a significant role in predicting MMSE classification and further emphasized the multidimensional relationship between oral health status and cognitive performance.

Conclusions:

The oral health parameters and demographic data used as inputs for ML classifiers contain sufficient information to differentiate between MMSE scores ≤26 and 30. This study suggests oral health parameters and ML techniques could offer a potential tool for screening MMSE scores for individuals aged 60 and older. Keywords: Classification, Machine Learning, Mini-Mental State Examination, Cognitive Impairment, Oral Health. Clinical Trial: NCT06611475