Search

Abstract Background Excessive and patterned ambulation is associated with falls, urinary tract infections, co-occurring delirium and other acute events among long-term care residents with cognitive impairment/dementia. This study will test a predictive longitudinal data model that may lead to the preservation of function of this vulnerable population. Methods/design This is a single group, longitudinal study with natural observations. Data from a real-time locating system (RTLS) will be used to objectively and continuously measure ambulation activity for up to 2 years. These data will be combined with longitudinal acute event and functional status data to capture patterns of change in health status over time. Theory-driven multilevel models will be used to test the trajectories of falls and other acute conditions as a function of the ambulation activity and demographic, functional status, gait quality and balance ability including potential mediation and/or moderation effects. Data-driven machine learning algorithms will be applied to run screening of the high dimensional RTLS data together with other variables to discover new and robust predictors of acute events. Discussion The findings from this study will lead to the early identification of older adults at risk for falls and the onset of acute medical conditions and interventions for individualized care.