Alzheimer's disease patient groups derived from a multivariate analysis of cognitive test outcomes in the coalition against major diseases dataset

Aim: The mini-mental state examination, commonly used to measure cognitive impairment of Alzheimer’s disease (AD) patients, consists of five test categories. The final score is calculated as their total sum, implying a loss of information. Materials & methods: In this study, we propose a new multivariate approach to address this issue. Results: We analyzed the current largest AD-related coalition against major diseases dataset comprising 3717 patients of interest. Our clustering approach revealed five groups of patients associated with distinct characteristics and prognosis. Interestingly, only three cognitive test categories significantly contribute to their determination: registration, attention and recall. Conclusion: The insight that only these categories are critical for AD group determination may help to resolve the patients’ educational background issue often discussed in relation to the mini-mental state examination assessment.