Background: Machine learning techniques, specifically classification algorithms, may be effective to assist in understanding key health, nutritional and environmental factors associated with cognitive function in ageing populations. Objective: The objective of this study was to use classification techniques to identify the key patient predictors considered most important in the classification of poorer cognitive performance, which is an early risk factor for dementia. Methods: Data was utilised from the Trinity-Ulster and Department of Agriculture (TUDA) study, which included detailed information on sociodemographic, clinical, biochemical, nutritional, and lifestyle factors on 5,186 older adults recruited from the Republic of Ireland and Northern Ireland, a proportion of whom (20%) were followed up 5-7 years later for reassessment. Cognitive function at both time points was assessed using a battery of tests, including the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS) with a score <70 classed as poorer cognitive performance. This paper trained three classifiers; decision trees, Naïve Bayes and random forests, to classify the RBANS score and to identify key health, nutritional and environmental predictors of cognitive performance and cognitive decline over the follow-up period, and assessed their performance, taking note of the variables these optimised classifiers deemed as of key importance for their computational diagnostics. Results: In the classification of a ‘low’ RBANS score (<70), our classification models performed well (range F1-score 0.73 to 0.93), all highlighting the individual’s score from the Timed Up and Go (TUG) test, the age the participant left education and whether or not the participant’s family reported memory concerns as of key importance. The classification models performed well in classifying a greater rate of decline in the RBANS score (range F1-score 0.66-0.85), also indicating the TUG score as of key importance, followed by blood indicators: plasma homocysteine (tHcy), vitamin B6 biomarker (plasma pyridoxal-5-phosphate; PLP) and glycated haemoglobin (HbA1c). Conclusions:The results presented here would suggest that it may be possible for a healthcare professional to make an initial evaluation, with a high level of confidence, of the potential for cognitive dysfunction using only a few short, non-invasive questions, thus providing a quick, efficient and non-invasive way to help them decide whether or not a patient requires a full cognitive evaluation. This approach has the potential benefits of making time and cost savings for health service providers and avoiding stress created through unnecessary cognitive assessment in low risk patients.
|Journal||JMIR Medical Informatics|
|Publication status||Accepted/In press - 24 Jul 2020|
- supervised machine learning
- timed up and go
Rankin, D., Black, M., Flanagan, B., Hughes, C., Moore, A., Hoey, L., Wallace, J., Gill, C., Carlin, P., Molloy, A., Cunningham, C., & McNulty, H. (Accepted/In press). Identifying Key Predictors of Cognitive Dysfunction in Older People Using Supervised Machine Learning Techniques. JMIR Medical Informatics.