Identifying clusters on multiple long-term conditions for adults with learning disabilities

Cluster analysis has been applied in several clinical studies, leading to improved management and allocation of healthcare. However, there is still limited application of cluster analysis to group common multiple long-term conditions (MLTCs) for patients with learning disabilities. Performing such cluster analysis on people with learning disabilities could provide critical insights into the prevalent conditions across individual groups and possibly common trajectories of these conditions among the respective groups. Identification of clusters of MLTCs, alongside associated risk factors, may reveal pathways to prevent certain outcomes such as disease progression and early mortality, which are common among this group. Cluster analysis may also enable the development of specialised clinical systems to provide personalised care to these patients. This paper compares six clustering algorithms based on their ability to effectively create separable MLTC clusters. The algorithms were independently applied to datasets of male and female adults with learning disabilities from Wales. This analysis is part of an ongoing research effort to identify major MLTC clusters for people with learning disabilities.