Falls are a common, debilitating problem among older adults and a significant source of morbidity. The objective of the present proposal is to investigate a novel approach to the evaluation of fall risk. During the past three decades, tremendous advances have been made in the understanding of the factors that contribute to falls, however, assessment tools are not yet optimal. In the clinic, observational gait analysis may be performed to characterize balance, walking abilities and fall risk. For further insight, balance and gait may be challenged, e.g., using an obstacle course. However, if these approaches prove insufficient and the geriatrician or neurologist needs to understand what occurs as a patient carries out routine activities of daily living, self-report will be relied upon. In fact, despite its subjective nature and the known problems concerning recall, self-report is currently, to a large degree, the gold standard of fall frequency and risk. We suggest that much can be gained by offering the clinician who wishes to evaluate falls an approach similar to that of the cardiologist's Holter monitor: ambulatory monitoring of near falls (NF). A near fall refers to a misstep, trip, stumble or loss of balance in which recovery mechanisms are activated to prevent a fall. Automatic identification of NF should, a priori, provide a sensitive and objective marker of fall risk, perhaps over a shorter study period. We recently began to investigate the potential of using an ambulatory monitor to identify NF. In preliminary work in laboratory testing, mostly in young adults, we were able to achieve detection rates better than 85% sensitivity and 85% specificity. The present proposal is designed as a bridge to larger scale validity studies. To this end, we have defined the following specific aims: 1. To develop algorithms to automatically identify NF in older adults under laboratory conditions. 2. To develop and evaluate algorithms for the detection of NF (using accelerometers and gyroscopes and determining which configuration and combinations are ideal) in real-world conditions. 3. A) To establish an annotated data base of NF, as recorded by movement sensors, as study participants carry out their routine activities of daily living. B) To make this archive available via the open-access NIH-funded PhysioNet Resource website. To achieve these objectives, 30 older adults with a history of multiple falls, 30 age-matched controls and 10 patients with Parkinson's disease with a history of recurrent falls will be tested in the lab using previously established tests of balance, gait and fall risk. Subsequently, they will be asked to wear an ambulatory monitor for 3 consecutive days. We will assess which methods optimally detect NF and the association between NF and laboratory-based measures. The results of this exploratory study should help to promote a third approach to the study of falls and falls risk, one based on ambulatory monitoring that may, ultimately, lead to more comprehensive fall risk assessment options.
Falls in older adult are a major cause of morbidity and mortality, with significant associated healthcare costs. We suggest that much can be gained by developing a new approach to fall risk assessment: ambulatory monitoring of near falls. Automatic identification of near falls (e.g., trips, missteps, stumbles), as subjects carry out their routine activities of daily living, may provide a sensitive and objective marker of fall risk based on actual performance that may, ultimately, lead to improved assessment and treatment options.
