Wearable sensors may help to quickly spot patients at risk of falls

A recent study from the University of Oxford suggests that wearable sensors, when used in a brief gait and posture assessment in the clinic, can accurately predict the risk of falls in Parkinson’s disease patients. The study, which involved over 100 Parkinson’s patients without a history of falls, found that artificial intelligence (AI) models analyzing sensor data could distinguish between patients who would fall within two to five years and those who would not. The researchers used sensors placed on patients’ chest, waist, wrists, and feet to track movement during a two-minute walk test and a 30-second postural sway task. AI models revealed that features like stride length were significant indicators of fall risk, with the most accurate predictive model showing 84%-92% accuracy for predicting falls within two years, and 78% accuracy over five years. This digital approach offers a more objective and efficient method of identifying those at risk for falls, potentially aiding in the development of targeted prevention strategies. The study suggests that wearable sensors could play a crucial role in predicting falls and informing interventions such as exercise programs or medication reviews.