In 2021, I-MED partnered with annalise.ai to implement a comprehensive decision-support AI solution for chest X-rays Australia nationwide, with a view to empowering its radiologists to make more accurate and faster decisions.
I-MED Radiology is Australia’s largest medical imaging service provider with over 250 clinics nationwide. Their investment in state-of-the-art systems, technology and imaging innovation underpins their commitment to delivering expert diagnosis, high quality care and better health outcomes for patients.
I-MED
250+ clinics
nationwide
More than 6 million patient procedures performed each year
350
radiologists
4,000+ clinical and non-clinical staff
The challenge
I-MED engaged annalise.ai to implement a comprehensive AI decision-support solution for chest X-rays across its national network of clinics. The solution needed to:
- Integrate with I-MED’s IT systems and reporting workstations
- Provide fast and accurate results for reporting radiologists
- Be user friendly, easy to learn and seamlessly integrate with current reporting workflow
- Be scalable
Annalise Enterprise CXR
Current AI decision-support solutions only identify a small number of clinical findings – most fewer than 15* – which limits the clinical utility of decision-support AI. Radiologists require more clinically diverse AI tools that can detect a wide range of clinical presentations. However, implementing multiple tools creates a lack of scalability and inconsistent user experiences. annalise.ai engaged clinicians early in the product development process to gain a deep understanding of the real challenges faced in medical imaging. The solution:
- detects a comprehensive 124 suspected findings relevant to ICU, ED, inpatient and outpatient settings
- is one of the only decision-support AI tools for chest X-ray that analyses up to three images, including laterals*
- Annalise Enterprise CXR’s customisable user interface (UI) integrates seamlessly into RIS/PACS
- was developed on one of the world’s largest chest X-ray datasets of more than 780,000 studies
*As of March 2022
The implementation
annalise.ai rolled out the implementation in two key phases, an initial six week pilot followed by an enterprise deployment across I-MED’s national network.
Initial pilot
Over a six week period, 11 I-MED radiologists tested and validated Annalise Enterprise CXR as a clinical decision-support solution while reporting 2,972 chest x-rays. Each participating radiologist was asked to provide feedback on the impact of Annalise Enterprise CXR to their final report. The evaluation revealed:
Results
CXR reports were significantly altered with use of the AI model
patients experienced a significant change in patient management
of radiologists felt that the AI model improved their accuracy when reporting
of radiologists felt more positive towards AI in general having used Annalise Enterprise CXR
Source: Jones et al. BMJ, Dec 2021
Enterprise deployment
Having established real world clinical benefits, annalise.ai and I-MED developed a strategy for a large-scale implementation of Annalise Enterprise CXR. annalise.ai worked in consultation with I-MED to gain a deep understanding of I-MED’s IT infrastructure, the way their Radiologists work and I-MED’s experiences with other service providers. This ensured a thorough deployment strategy which involved:
- A comprehensive IT plan including PACS integration and installation of Annalise Viewer on all reporting workstations
- A commitment to end-to-end change management with I-MED’s vision at the centre
- A series of interactive webinars and a bespoke multimedia communications plan to drive clinician engagement
- An easily accessible onboarding system with clinical training materials for Radiologists
- Comprehensive training and ongoing support for I-MED’s IT team
The annalise.ai and I-MED projects teams rolled-out Annalise Enterprise CXR in a staged approach across its business units. The teams convened regularly to assess progress, review feedback and address any necessary changes to the rollout process. This ensured any issues were identified and resolved quickly and effectively to minimise impact to clinicians.
Given the scale of our business, working in close collaboration with annalise.ai throughout the entire process was key to the success of the implementation. The annalise.ai team were consistently proactive and invested time and resources to ensure our clinical and technical staff felt confident and well supported in using Annalise Enterprise CXR.
Dr Ron Shnier
Chief Medical Officer, I-MED
Long-term success
Throughout the transition process, annalise.ai continued to support I-MED’s internal teams while monitoring performance and feedback. When Radiologists were asked about their experience with Annalise Enterprise CXR after an initial 3 month period it was found:
Results
Source: Internal report, 2021
annalise.ai continues to support I-MED and implement product improvements through planned system upgrades, along with ongoing communications and training initiatives.
annalise.ai partner network
annalise.ai offers flexible deployment solutions including both on premise and cloud. For those customers that wish to realise the benefits of cloud computing, including availability and redundancy, storage and scalability, annalise.ai works closely with Amazon Web Services (AWS). Through this partnership, annalise.ai has access to a dedicated Solution Architect to consult on the future direction and development of the annalise.ai suite of products. This relationship has been key to annalise.ai realising the success of its cloud deployments, including the I-MED project.
Disclaimer: The information contained herein is for healthcare professionals only. Annalise Enterprise CXR is not intended to provide direct diagnosis. For detailed device information, including indications for use, contraindications, and warnings, please consult the user guide prior to use. Annalise Enterprise CXR is not for sale in the US. Not all features are available in all regions, check current regulatory status with an annalise.ai employee or contact info@annalise.ai.
All product names, logos, brands, trademarks and registered trademarks are property of their respective owners.
OPT-PRM-024 V1