Baek wins grant using AI to reduce falls among elderly

Geb Thomas

Professor Stephen Baek received a $15,000 grant from the Centers for Disease Control and Prevention (CDC), National Center for Injury Prevention and Control for a pilot project entitled "In-home gait and balance screening for the risk assessment of falls in elderly populations." He will lead a research team comprised of researchers at CCAD and UIHC to develop AI algorithms to predict the fall risk in an older adult based on his/her gait pattern measured by smart-shoe sensors. 

Falling is an increasingly important public health problem with 30% of older adults who live in the community experiencing a fall each year. Fall death rates begin to rise around age 65 and increase with age. Of all the unintentional injury deaths in adults aged 65 and older in the United States, falls account for 55% of those deaths. When not fatal, falls often result in hip fractures, brain injuries and loss of independence. The CDC estimates Medicare expenditures for fall injuries at $31 billion annually.

The CDC has developed the Stopping Elderly Accidents, Deaths and Injuries (STEADI) initiative to assist in the screening for, assessment of and interventions for fall risks. This initiative provides an algorithm for use by the primary health care provider which includes history reports from patients and validated fall risk self-assessment questionnaires. For patients with history findings consistent with fall risk, the initiative also recommends a clinician’s evaluation of patients’ gait, strength and balance. The vast majority of primary care health providers lack the time or expertise to evaluate the older adult’s gait. Even experienced geriatrician’s evaluations of gait and balance are rudimentary.

There is a need for a more precise way to assess older adult’s gait and balance to initiate known interventions that prevent falls. This primary prevention measure ideally will measure aspects of gait and balance in the older adults’ home environment rather than in the foreign and somewhat artificial health clinic facility. Falling is caused by a combination of intrinsic (personal) and extrinsic (environmental) factors. Monitoring gait and balance with the use of technology such as smart shoes in the home may provide more valid information and improve fall risk assessment because gait is analyzed while the older person is interacting with the environmental hazards in the home.

In this pilot project, the UI researchers will develop AI algorithms and protocols to predict the fall risk in an older adult based on the gait pattern measured by smart-shoe sensors. Motion data of 100 adults 65 years and older will be studied to capture the latent patterns and trends behind sensor signals. These underlying patterns will then be used for classifying individuals with high fall risks. 

Photo of a proud Stephen Baek.