About one-third of individuals over 65 fall at least once each year – and that increases to two-thirds for those who have already had a prior fall (Report from the Annals of Internal Medicine). Injuries from a fall are often severe and can lead to long-term disability and even premature death. With an aging baby-boom generation, this is of increasing significance to the health of this population and to greatly rising health-care costs.
Texas Tech University (TTU), the Texas Tech University Health Sciences Center (TTUHSC), and Texas Instruments Incorporated (TI) have partnered to conduct research on the development of a fall-prevention system. There have been a number of attempts to develop sensor systems that can automatically detect and report a fall – or sensors that detect when a person attempts to get out of a bed or a wheel chair and alert caregivers. This research has taken things a step further to work on the creation of an electronic system to detect instability prior to a fall allowing the person to take actions to help prevent the fall. If a fall occurs, like other systems, it can also automatically report the fall, for example by a text message to a relative or a signal to a nurse’s station. Additionally, such a smart system can distinguish between an actual fall and just “plopping down” in a chair.
With funding and collaboration from TI, Prof. Donald Lie, Prof. Tim Dallas, and their graduate students at TTU, Dept. of Electrical Engineering, have built a prototype system. The system utilizes commercial accelerometers and gyroscopes, a TI ultra-low-power MSP430 microcontroller (MCU) to digitize and process all the signals, and includes a TI CC2500 radio frequency transceiver for wireless transmission of the signals to a computer or other wireless device. The wireless communication uses the SimpliciTI network protocol and embedded software and utilized TI’s eZ430-RF2500 development tool for system integration.
Upon basic development of the electronic system, Prof. Lie and his engineering research team began collaboration with clinicians at TTUHSC to determine the best placement of the sensor system on the body. They tried various placements, including leg-mounted, foot-mounted (on the body or in shoes) and torso-mounted. They found that the torso-mounted system delivers the best results. The current prototype system can be worn on a belt or clipped onto a pocket or other piece of clothing. In the future, a commercial system should be able to be miniaturized to be less intrusive, so that it could be routinely worn on a daily basis.
With the working prototype in place and best body placement determined, the team began studies using volunteers at the TTUHSC Geriatric Education and Care Center. They began analyzing the dynamics of posture and walking gait to build a database of normal movement (including sitting and rising) versus patterns of instability that precede a fall, which can include the pattern of people’s response to the initial onset of falling. These efforts are not yet complete, but the team hopes to identify and use conditions that occur before a fall to help warn people before they actually fall. The warning system could be an audible signal or vibration to tell the person that they should “grab hold” of something or sit down before they fall. This was the part of the program of most interest to TI; this took the research beyond just fall detection and aspires to use electronics to take preventative measures to help reduce falls.
This system, of electronic system and software analytics, should be helpful in reducing falls in elderly and people with balance/vestibular disorders and diseases of the central nervous system, such as Parkinson’s, epilepsy and dementia. Use of the sensor system should be able to help identify those at highest risk for falls – since the sensor data can be wirelessly streamed to a computer system for collection. Ultimately, the team hopes the research will lead to an inexpensive commercial device that incorporates the best software analytics for early warning and fall prevention for any patient with compromised balance, and that it can be implemented widely for reduction of falls in clinical care and at-home use.
Dr. Lie noted, “We believe that this project will really make a significant difference in the clinical care of geriatric populations.” Dr. Lie credits his team of collaborators at TTUHSC for the motivation and clinical pull for this research. These TTUHSC collaborators include Drs. Tam Nguyen, Steven Zupancic, Andrew Dentino and Ron Banister. Prof. Lie’s engineering research group and their TTUHSC collaborators have published a conference proceedings’ paper about the early system development. It is “A Fall Detection Study on the Sensor Placement Locations and the Development of a Threshold-Based Algorithm Using Both Accelerometer and Gyroscope”, by J. Jacob, T. Nguyen, et al., in the Proc. IEEE International Conf. on Fuzzy Logics (Fuzz’11), pp. 666-671, June 27-30 (2011).
Allen Bowling is a TI Fellow at Texas Instruments, a company that provides integrated circuits for many applications including wireless communication, embedded processing, and a wide variety of analog functions.