Robots in Nursing Homes: Helping Nurses Detect and Prevent Falls.

Abstract

Falls are a leading cause of morbidity and mortality in older adults, especially among nursing home residents. Falls occur more commonly among older adults with dementia than among those without dementia. Moreover, half of nursing home residents have moderate to severe cognitive impairment. While less than 5% of older adults live in nursing homes, they account for 20% of deaths from falls in this age group. In addition, 78% of older adults who fall need help in getting up from the floor. The consequences of falling, such as prolonged lying on the floor, can produce severe and prolonged health effects. The acute shortage of staff in nursing homes, especially during evening, night and weekend shifts, can delay the detection and response to falls. There are various systems designed to detect falls and alert staff, including those utilizing wearable devices, ambience sensors and cameras (vision) as well as fusion systems. They each have their advantages and drawbacks. In this NIA-funded SBIR grant, we are developing and testing the feasibility of a fall detection and prevention system that addresses the drawbacks of previous systems. We anchor our approach on the deployment of an autonomously navigating robot equipped with a mounted infrared camera and machine learning software designed to detect the risk of falls and falls themselves. The robot will patrol resident rooms during evening and night shifts and alert the staff, allowing them to evaluate the fall risk or fall alert presented by the robot video camera and determine whether indeed a resident has fallen or is at risk of falling, and take appropriate action.