UK’s first randomised controlled trial of AI-assisted fracture detection in ED launched

Oxford University Hospitals (OUH) NHS Foundation Trust has received approval to begin the UK’s first multicentre randomised controlled trial of AI-assisted fracture detection in the Emergency Department (ED), in collaboration with the Danish MedTech company Radiobotics.

The SAMURAI-Fracture study is an OUH-sponsored study that will evaluate the real-world impact of implementing Radiobotics' AI-powered fracture detection tool, RBfracture™ in clinical workflows across four NHS Trusts in the Thames Valley. 

The work is being carried out under the umbrella of OxCAIR (Oxford Clinical Artificial Intelligence Research) and supported by SBRI Healthcare Competition 26, an NHS England Accelerated Access Collaborative initiative focused on improving the delivery of timely and urgent emergency care. 

The SAMURAI-Fracture (Systematic Assessment of the Medical Utility of Radiology Artificial Intelligence - Prospective Evaluation of Fracture Detection) study aims to address key gaps in the evidence generation plan for the National Institute for Health and Care Excellence (NICE) Early Value Assessment for AI Technologies to Help Detect Fractures on X-rays in Urgent Care, published in January 2025.

Mistakes made by clinicians in X-Ray interpretation are the most common error made in NHS UK Emergency Departments, and the most frequent reason for successful litigation. This study will provide detailed evaluations of the impact of using the RBfracture™ tool on clinical errors. Previous small-scale studies have suggested that RBfracture™ can reduce the number of missed fractures in emergency departments by up to 86 percent, with an accuracy, specificity and sensitivity of 94 percent and a median processing time of 15 seconds per exam. 

The SAMURAI-Fracture study will generate detailed prospective evidence of the impact of using the tool, from patient and clinician-level outcomes though to technical performance and wider system effects including health economic analysis. 

The study aims to go beyond conventional diagnostic performance studies to deliver evidence of AI’s real-world effect on clinical workflows, patient care and healthcare resources. The study will measure whether RBfracture can reduce avoidable follow-ups or delays in care by enabling faster, more accurate decision-making in the emergency setting. 

“We're excited to begin this important study in collaboration with Radiobotics. As pressure continues to build in emergency care, we need to explore technologies like RBfracture that can safely streamline decision-making and improve care for patients,” says Professor Alex Novak, Co-Chief Investigator for the study and Co-Director of OxCAIR. 

The project will employ a prospective cluster randomised cross-over trial — a robust study design in which different hospital sites or departments will switch between using RBfracture and standard care at different times. This allows researchers to fairly compare the two approaches and understand RBfracture’s practical benefits across diverse settings. 

“We’re thrilled to officially move forward with this important work,” says Michael Lundemann, Chief Clinical and Scientific Officer at Radiobotics. “This project is about so much more than testing RBfracture for performance — it’s about exploring how AI can be used to support frontline NHS staff and improve care delivery for patients when it matters most.” 

The SAMURAI project also closely aligns with broader UK Government priorities, including: 

• Ensuring the NHS has the right tools and workforce when and where patients need them,
• Reducing lives lost to major health threats through faster, smarter interventions, and
• Creating a fairer healthcare system where everyone can live well for longer. 

Dr. Sarim Ather, Co-Chief Investigator and Co-Director of OxCAIR, adds: “Our goal is to generate robust clinical evidence on the impact of AI on patient pathways, which is critical for responsible adoption at scale and understand how AI delivers meaningful impact.” 

This work was commissioned and funded by SBRI Healthcare, an Accelerated Access Collaborative (AAC) initiative, in partnership with the Health Innovation Network.