AI-assisted CT scans in the Emergency Department

The Challenge

Emergency Departments (EDs) across Australia face mounting pressure from overcrowding and long waiting times, which can compromise patient safety and the quality of care. Head injury is a common reason for ED presentation, often requiring a CT brain (CTB) scan before a patient can be safely discharged.

While the use of CT imaging has increased significantly, around 90% of CTB scans ordered in the ED show no clinically significant findings. Low-risk patients often wait hours for a formal radiology report, unnecessarily prolonging their ED stay. This delay contributes to overcrowding, strains radiology resources, and diminishes the patient experience.

This project led by RPA Green Light Institute, directly addresses this inefficiency by aiming to create a streamlined care pathway that uses Artificial Intelligence (AI) to help clinicians quickly and safely identify patients who can be discharged sooner.

Project Overview 

The project is implementing and evaluating an innovative AI-assisted model of care for non-contrast CTB scans in the ED. Its primary goal is to accurately identify clinically unremarkable scans, allowing senior ED clinicians to safely discharge low-risk patients sooner, with follow-up coordinated through the RPA Virtual Hospital.

The project comprises three stages:

1. Retrospective validation – Testing the AI model’s accuracy on approximately 4,000 historical CTB scans from RPA.
2. Prospective Randomised Controlled Trial – Evaluating the impact of AI-assisted workflows on ED length of stay.
3. Cost-effectiveness analysis and scaling strategy – Assessing the economic benefits and planning rollout to other hospitals.

By combining cutting-edge AI with innovative care models, this project seeks to improve ED efficiency, reduce overcrowding, and enhance patient experience without compromising clinical safety.

Measuring Impact

Hearts and Minds measures its impact against six core categories as developed by the Association of Australian Medical Research Institutes. Key highlights include:

Advancing Knowledge

• Collaboration and expertise: Strong engagement with internal leaders from ED, Radiology, plus external clinical guidance from a neurosurgeon‍
• Media and dissemination: Featured by Sydney Local Health District and SBS News highlighting its innovative approach.

Research Capacity Building

• Funding award: $50,000 from The Pitch, recognising the project’s innovation and the fellow’s research capability
• AI and digital health skills: Expanding capability among clinical and research teams‍
• Training and knowledge transfer: Empowering ED clinicians and radiologists to use AI-enabled care models.

Health Impacts

• Faster care: Targeting reduced ED length of stay for low-risk head injury patients
• Safe discharge: Maintaining diagnostic accuracy while expediting care‍
• Improved patient experience: Coordinated virtual follow-up after early discharge.

Economic Impacts

• Cost savings: Anticipated reduction in costs from shorter ED stays‍
• Scalable implementation: Developing a framework to roll out the AI workflow nationally.

Informing Decisions

• Policy and clinical guidance: Providing evidence to inform hospital and health policy, as well as clinical decision-making on CT brain scan outcomes.

‍

This project represents a significant step forward in optimising emergency care. By reducing unnecessary delays, enhancing patient safety, and providing robust evidence for AI integration not only addresses ED overcrowding but also demonstrates the transformative potential of innovative research and technology in improving healthcare outcomes.

‍Funding support from Hearts and Minds Investments, as nominated by Core Fund Manager, Regal Funds. This content was last updated in August 2025, for further information visit RPA Green Light Institute for Emergency Care.