Co-Design of a Patient Fall Risk Prevention Service Powered by Machine Learning

In hospitals, patients often experience falls at night, especially when they're under the influence of painkillers and anesthesia. Instead of seeking assistance, these patients, in their confused state, might impulsively try to get up, endangering their safety. Such incidents can lead to severe injuries, even fatalities, or cases where patients wander away from their rooms. These occurrences not only impact the health of the patients but also increase insurance costs for hospitals. Centro Chirurgico Toscano (CCT - Tuscany Surgery Center) and the Interaction Design Lab of the University of Siena have explored a fall prevention system using Machine Learning (ML) and image recognition. This system, installed in a CCT room, monitors patients' positions in real-time (whether they're lying, sitting, or standing) and raises an alert for any risky postures. The project favored a human-centered approach to meet the specific needs of stakeholders. The setting of the fall prevention system, and four dashboards to view, control, manage and modify the system, were co-designed for and with the CCT. The tested system has shown reliability, leading to considerations for its broader application. The goal is to promote a preventive culture in health care by leveraging technologies such as image recognition to minimize fall risks.