A New Approach to Individualize Physical Activity Interventions for Individuals With Dementia: Cluster Analysis Based on Physical and Cognitive Performance.

Abstract

Physical activity (PA) can have a beneficial effect on cognitive and physical performance in individuals with dementia (IWD), including those residing in nursing homes. However, PA interventions in nursing homes are usually delivered using a group setting, which may limit the effectiveness of the intervention due to the heterogenous nature of IWD. Therefore, the purpose of this study was to identify clusters based on cognitive and physical performance values, which could be used to improve individualization of PA interventions. Based on the cognitive and physical performance variables of 230 IWD, a cluster analysis was conducted. Global cognition (Mini-Mental State Examination), mobility (6-Meter Walking Test), balance (Frailty and Injuries: Cooperative Studies of Intervention Techniques-subtest-4), and strength and function of lower extremities (30-Second Chair-Stand Test) were assessed, and values were used to perform a hierarchical cluster analysis with Ward's method. Differences in physical and cognitive performance as well as other secondary outcomes (age, sex, body mass index, use of walking aids, diagnosis and etiology of dementia, number of medications, and Cumulative Illness Rating Scale) were tested using 1-factorial analyses of variance. Out of 230 data sets, 3-cluster solutions were identified with similar cluster sizes of 73 to 79. The silhouette coefficients for all calculated clusters ranged between 0.15 and 0.34. The cluster solutions were discussed in the context of cognitive and physical functions as well as training modalities and opportunities. The 4-cluster solution appears to be best suited for providing or developing an individualized PA intervention. The identified clusters of the 4-cluster solution may be used in future research to improve individualization of dementia-specific PA interventions. By assigning IWD to these clusters, more homogenous groups with regard to cognitive and physical performance can be formed. This allows for more individualized PA interventions and may result in a higher effectiveness, particularly in nursing homes. Our findings are relevant for therapists and nursing staff who design or deliver PA interventions in nursing homes or similar settings.