Researchers at the University of Missouri are working to develop and evaluate a mountable in-home system that can detect falls, as well as measure changes in movement that might signal an increased fall risk. This system includes a radar component and a modified Microsoft Kinect (originally designed to work with the Xbox video game system). This type of non-wearable fall detection system could be invaluable for a number of settings where falls are a major risk, such as hospitals, long-term care, and congregate senior housing. However, this radar and Kinect system provides the additional benefit of also being able to measure fall risk as individuals go about normal daily living. Measuring fall risk in this way allows for notification and interventions when greater fall risk is detected by measuring everyday movements prior to an actual fall. Measuring fall risk in this way can also provide evidence for the positive impact of strength training or other exercise-based interventions, a potentially valuable source of encouragement. Lastly, such a system can potentially keep health care providers and family members informed about any changes in falls or fall risk.
At this point, researchers are field-testing this system to perfect the algorithms that this home-based system employs to assess fall risk. In a recently published study, researchers compare this system to results obtained from a standard, commonly used, reliable, and well-validated set of tasks that make up a standard fall risk assessment, as well as to results from the GAITRite® Electronic Walkway, the most accurate validated measurement tool for gait and fall risk. This comparison was done with Kinect and radar systems installed in residences of 15 participants who lived in a senior living community. In these apartments, the Kinect was installed on a small shelf above the front door, and the radar system was installed in a decorative wooden box nearby.
Comparing the Kinect and radar system to the battery of tasks included in the fall risk assessment commonly used by health care providers, the Kinect and radar data correlated consistently in the expected direction with the fall risk assessment tasks. For each of the tasks in the fall risk assessment, these correlations were statistically significant for at least 4 out of the 5 measurements taken by the Kinect and radar system.
Compared to GAITRite, correlations with the Kinect and radar system were again in the expected direction, but these did not reach statistical significance. It should be noted that the GAITRite measurements were taken in a controlled laboratory environment, while the Kinect and radar system was installed in an actual living space. The authors also suggest that the lack of statistically significant findings could be due to the small 15-person sample size in this study.
Overall, this study concludes that “an automated, continuous, unobtrusive, environmentally mounted, in-home monitoring system is possible and has potential for success.” In addition to providing alerts for actual falls without the need for a wearable device, a number of advantages to such a system exist, such as the continuous measurement of fall risk, the lack of a need for a health care professional to physically measure fall risk, and the possibility of having a much more fine-grained view of how fall risk may be changing over time, as opposed to measuring fall risk at longer intervals.