Award Date: 2018
Project Lead(s): Nicole D. Anderson, Senior Scientist at Rotman Research Institute at Baycrest
Project Title: Deployment and Pilot of a Video-based Safety System for Individuals with Alzheimer’s Disease and Related Dementias
What is the issue?
Long-term care residents with dementia fall an average of four times per year, putting their health and mobility at risk. To prevent more falls from happening, care staff need to understand why and how they happen.
What did we do?
SafelyYou is a video technology that detects falls in real-time. It uses wall-mounted cameras enabled with artificial intelligence in residents’ bedrooms. Care staff reviewed the videos to learn what caused a resident to fall, pinpoint the location and timing of the fall, and why the resident got up. We studied the effectiveness of the technology for detecting falls.
What did we find?
Care staff received 27 alerts, of which only 16 were real falls. The AI had a very high false alarm rate, requiring staff to validate detected falls as true or false. For the incidents staff deemed to be true falls, two out of five were not resident falls but other circumstances, such as a staff member picking something up. SafelyYou did not detect 20 per cent of actual resident falls, far below our target goal of 99 per cent. The videos captured circumstances leading up to falls 81 per cent of the time and informed clinicians’ implementing strategies to help prevent future falls. Family members wrongly thought the solution was for general monitoring of their loved ones and staff members. They also indicated they were reluctant to pay for the innovation and thought the institution should provide it as standard care.