Translating AI-generated Fall Prevention Models to Inform Organisational Decision-Making

Opportunity

Patient falls in hospitals are a serious and common problem that can result in injury, longer hospital stays, and even death. In Australia, over 55,000 patients experience fall-related injuries in hospitals each year, leading to significant physical and emotional trauma for patients and high costs for the healthcare system. Despite substantial investment in fall prevention, fall rates continue to rise. We will use artificial intelligence (AI) to help prevent patient falls in hospitals. We will develop and test the implementation of an AI system that can analyse patient electronic medical record data combined with workforce data to predict which patients are at higher risk of falling. This model, based on another dataset, will be refined and validated according to Alfred Health’s own dataset. This information will be displayed on a hospital dashboard that healthcare staff can easily monitor. The project will be conducted at Alfred Health and involve working closely with hospital staff to design a system that is practical and useful. The dashboard will provide near real-time alerts about patients at risk of falling, helping staff make informed decisions about fall prevention strategies and where to allocate healthcare workers and resources. If successful, this approach could significantly reduce patient falls, improve patient safety, and decrease healthcare costs. With additional refinements according to individual hospital requirements, the system will be translatable for use in other Australian hospitals and internationally.

Project Objectives

• To externally validate an electronic falls risk prediction model that identifies and stratifies ward patients at risk of hospital harm.
• To co-design with health service leaders and clinicians a falls risk alert reporting system and implementation strategy.
• To develop, implement and test a near real-time dashboard for falls risk and an alert reporting system.