Using machine learning in sarcopenia research

We’re pleased to announce the publication of a new paper in BMJ Open.  Members of the AGE Research Group aimed to build on previous work showing that living with multiple long-term conditions is associated with increased risk of sarcopenia by establishing whether this is driven by specific combinations of long-term conditions.

To do this we utilised machine learning.  This encompasses a number of different techniques for exploring large data sets. In this research we ‘grew’ decision trees which ‘decide’ whether a person is at increased risk of sarcopenia if they have any one of a wide range of different combinations of conditions.  The decision trees need information about a lot of people with and without many different conditions and with and without sarcopenia, and we used the UK Biobank data set for this.

We found a number of combinations of conditions where risk of sarcopenia was higher than would be expected if we had simply added together the effects of each separate condition.  The knowledge of combinations of long-term conditions that are associated with increased sarcopenia risk that we have generated could aid the identification of individuals for targeted interventions, recruitment of participants to sarcopenia studies and contribute to the understanding of the aetiology of sarcopenia. To find out more please see the full paper: https://bmjopen.bmj.com/content/14/9/e085204.full

Last modified: Mon, 30 Sep 2024 14:46:56 BST