Automatic Disability Categorisation based on ADLs among Older Adults in a Nationally Representative Population using Data Mining Methods.

ABSTRACT

The world's ageing population is rapidly increasing but people's healthspan is not being sustained. Activities of daily living and Montreal Cognitive Assessment scores from the first wave of a large nationally representative longitudinal study in ageing (TILDA) were analysed using multiple correspondence analysis, k-means clustering, network analysis and association rules mining, to find latent patterns in the data and categorise disability among older adults. It was observed that 6.2% of the population had a greater degree of frailty, specifically cognitive impairment. Additionally, the overall population showed difficulty in performing physically demanding activities. Thus, self-reported ADLs have a diagnostic importance as they indicate the level of cognitive and physical functional decline in the older population.