Motion Sensor-Based Monitoring of Physical Activity in Older Adults: Current Approaches and Clinical Applicability

This thesis addresses the use of motion sensor-based approaches to assess and monitor physical activity as input into interventions for older adults, specifically in geriatric health care settings. The work, as part of this thesis, pursues the following aims: (1) to compare the outcomes from different physical activity assessment approaches and the implications for their clinical applicability; (2) to investigate the associations of self-reported chronotypes with the circadian aspects of mobility; (3) to explore how circadian aspects of mobility-related behavior and clinical characteristics in dementia relate; and (4) to give an overview how on motion- sensor based monitoring of physical activity is currently used to guide and personalize physical activity interventions in geriatric health care.
A key finding is, that usually applied, intensity-based assessment approaches (e.g., actigraphy) are not recommendable in older adults, as they do not address the core aspects of physical activity in old age (i.e., performance
. The comparison of different physical activity assessment approaches within this thesis revealed that a mobility-related assessment of physical activity, however, can provide clinically applicable outcomes of overall physical activity and sedentary behavior in older adults. In a proof-of-concept study, this thesis presented a novel approach using mobility-related physical activity outcomes to analyze circadian aspects of physical activity and compared it to the self-reported individual diurnal preference (known as chronotype). A temporal correspondence between the daytime of peak gait activity and the individual chronotype in relatively healthy older adults were found. The timing of peak mobility-related behavior may constitute an innovative approach to monitor changes in physical activity behavior and personalize physical activity interventions. Building on this, this thesis explored how the diurnal timing of peak gait activity as well as nocturnal gait activity relate to clinical characteristics of patients in acute dementia care. The analyses revealed a large variability and differences in the timing of peak gait activity between patients and a high incidence of gait activity in acute dementia care. Nocturnal gait activity was associated to higher caregiver burden and more pronounced neuropsychiatric symptoms (i.e., symptoms of sleep and nighttime behaviors disorders). Analyzing the nocturnal gait activity as well as the diurnal timing of peak gait activity seem to provide clinical applicable information on the circadian aspects of mobility-related and may be used to personalize (non)- pharmacological treatment in acute dementia care. Subsequently, a scoping review was conducted to investigate how motion sensor-based monitoring is currently applied to guide and personalize physical activity interventions in geriatric health care. The results show a large heterogeneity in existing motion-sensor monitoring-based physical activity interventions. Furthermore, it seems that the potential of monitors with regard to the use of physical activity data to personalize interventions goals and feedback has not been exploited to the fullest yet. Future studies should focus on determining which intervention components are most effective in improving physical activity and especially sedentary behavior, investigate the effects of personalized physical activity goals and how physical activity monitor based interventions can be applied over the continuum of geriatric health care. Furthermore, the clinical validity of the proposed approach to analyze circadian aspects of physical activity based on mobility-related physical activity outcomes need to be tested. The findings of this thesis reveal important information on clinical applicable approaches using motion sensor-based monitoring to assess and intervene on physical activity in older adults and specifically in geriatric health care
settings.