Falling is not a normal part of the aging process and yet 1/3 to 1/2 of adults 65 years and older sustain at least one fall annually. Older adults are hospitalized for fall related injuries five times more often than from injuries from other causes contributing to a cost of $19 billion for nonfatal falls in the United States. Projected for the increasing aging population in the year 2020, it is expected that the costs related to falls will reach a staggering 54.9 billion dollars. Current research and clinical practice guidelines focus on multifactorial fall risk assessments as the critical deterrent to falls in the elderly. A primary factor within these assessments is activity of daily living performance of the individual elder. While current standardized clinical balance assessment tools have been proven effective for predicting fall risk, the tests are most commonly performed in the clinical environment and at isolated times during an individual's day. The goal of this application is to develop and validate a novel wearable device (Automatic Longitudinal Assessment Risk Monitor - ALARM) for longitudinal assessment of risk of falling. Such a device: - will allow early detection of risk of falling, when therapeutic interventions are most efficient - will provide real-time feedback about activity pattern - will provide feedback about compliance with interventions and effectiveness of interventions - will be incorporated into conventional footwear and require no extra effort to operate - can be used in research, clinical and potentially in consumer applications The development of the ALARM system will be addressed in three specific aims:
Specific Aims 1 : Develop a pattern recognition method that will improve recognition accuracy for activities of interest (such as walking and stepping up) by reducing the range of variation from current 76%- 100% to 9911%.
Specific Aim 2 : Collect data using the ALARM device on a group of elderly adults during clinical tests.
Specific Aim 3 : Develop algorithms for automatic assessment of risk of falling. In this Aim we will develop signal processing algorithms that automatically evaluate metrics indicative of the risk of falling in each activity of interest (e.g. duration of swing and stance phase during walking).
Specific Aim 4 : Validate the ALARM device in a double-blind unrestricted free living study. This set of Specific Aims will validate lead to creation of a unique wearable device capable of objective characterization of risk of falling.

Public Health Relevance

This application aims at development of a novel wearable device (Automatic Longitudinal Assessment Risk Monitor - ALARM) for longitudinal assessment of risk of falling. In our previous research we have shown that major activities and posture allocations such as standing, sitting, walking, etc. can be recognized with high degree of accuracy (76%-100%) by a wearable device incorporated into conventional footwear. We also have shown that sensor signals captured by the wearable shoe device during activities such as walking are well- correlated with the risk of falling (with numerical estimates of risk obtained through signal processing being directly proportional to the normalized scores from the clinical tests). The goal of this application is to develop and validate a novel wearable device (ALARM) for longitudinal assessment of risk of falling.

Agency
National Institute of Health (NIH)
Institute
National Institute of Biomedical Imaging and Bioengineering (NIBIB)
Type
Exploratory/Developmental Grants (R21)
Project #
5R21EB013183-02
Application #
8339885
Study Section
Social Psychology, Personality and Interpersonal Processes Study Section (SPIP)
Program Officer
Korte, Brenda
Project Start
2011-09-30
Project End
2014-08-31
Budget Start
2012-09-01
Budget End
2014-08-31
Support Year
2
Fiscal Year
2012
Total Cost
$185,300
Indirect Cost
$44,173
Name
University of Alabama in Tuscaloosa
Department
Engineering (All Types)
Type
Schools of Engineering
DUNS #
045632635
City
Tuscaloosa
State
AL
Country
United States
Zip Code
35487
Tang, Wenlong; Sazonov, Edward S (2014) Highly accurate recognition of human postures and activities through classification with rejection. IEEE J Biomed Health Inform 18:309-15