Ashley Clary, Vice President, Imaging Services, Ochsner Health System
Machine learning, augmented and assisted intelligence, artificial intelligence (AI)– not too long ago these descriptive phrases were seen as nothing more than visionary catch phrases embedded in the “future” section of an imaging division’s strategic plan. Fast forward to the hallowed halls (aka vendor showcases) of recent Radiology Society meetings and one may have thought they accidently booked a ticket to HIMSS, until of course they turn a corner and bump into a strategically placed CT scanner on the showroom floor.
"One caveat to this is the division actively developing or “training” these systems, but in those cases proper time and resources must be allocated to this work and expectations must be set that establishing the expectations/ outcomes of this work"
As an imaging executive, plethora of intelligence and knowledge engineering products is exciting and instills optimism in a provider shortage environment along with expectations of faster and more accurate diagnostic decisions and consumer demand for affordable, coordinated care for their loved ones. But how we do filter through the hype? It is important to prioritize – and we must stick to strategy and the core commitment to the production of reliable, high-quality images and accurate, complete diagnostic interpretations.
Before health system imaging divisions jump into the world of dynamic analytics and computer intelligence,it is important to reinforce its core platform. Three critical components to this foundation include:
1) Robust enterprise imaging management solution (integrated best of breed or single version) that can accommodate encounter-based solutions of all imaging types beyond standard radiology services.
2) Modalities and device solutions that promote data flexibility and sustainable upgrade solutions. Imaging equipment with strongly executed AI systems continue to be expensive, even as reimbursements continue to shrink. Faced with this constraint, a company’s strategic plan should aim for a machine fleet that can survive for more than a decade while continuing to be updated as needed, and these become requirements when evaluating fixed assets that require significant construction costs. Successful vendors will focus on upgrade paths, flexible software subscription solutions, peripheral development, and cross modality integration solutions – all capable of exporting data or promoting machine-based data solutions.
3) Most importantly, a clinically driven imaging informatics governance structure. As healthcare systems continue to grow through a variety of partnership models, a significant challenge they must overcome is aligning diverse stakeholders, including employed clinicians, contracted providers and operations/informatics team members. It is critically important to build service line networks enhancing standardization and integration opportunities.
While the showroom floor continues to be a great place for ideas to come into formation, the focus must be on developing the strategic plan and examining how integrating AI systems can help achieve identified priorities. What strategic initiatives associated with medical imaging will support an organization’s mission? What barriers or challenges are disabling this strategy? AI tools should support these initiatives by breaking down barriers, expediting processing or augmenting services.
Service line strategies typically focus on cost reduction, standardization, quality control, growth, and consumerism. Knowledge engineering and data analytics tools can be applied well here. For example:
• Using machine data from sonography units and the electronic medical record (EMR) to detect and display information by sonographer scanning time and protocol variations by exam. This tool enables leaders to quickly identify sonographers that may need additional education and training as well as machines that are not functioning optimally.
Using AIto identify pneumothorax on a portable x-ray unit allows the technologist to notify the critical care team immediately in the Critical Care unit so they can review the image on the machine in tandem with the Radiologists reading formally, promoting a faster patient response time.
• These applications exist today and over the next 18 to 24 months there will likely be hundreds more. It is imperative that leaders properly align these tools with key strategic initiatives. Furthermore, those strategic initiatives should be aligned between clinical, operational, and informatics teams. Failure to generate alignment prior to implementation of these types of products/programs will overwhelm the division leading to frustration, burnout, and “solutions” that are not optimized to meet the department needs. The governance council must provide guidance on which tools are most applicable to your areas of focus and then prioritize them by using methodology associated with scalability and value proposition.
As with most healthcare system initiatives, clinical buy-in will make or break the successful application of these tools. If interacting with these tools require additional steps, clicks or mental processing power to complete a task, then the implementation will fail, or at the very least, be prolonged resulting in a lower financial or value return. One caveat to this is the division actively developing or “training” these systems, but in those cases proper time and resources must be allocated to this work and expectations must be set that establishing the expectations/outcomes of this work.
Medical Imaging is a core component to the delivery of care within integrated health providing diagnostic guidance to support clinical management of patients. Clinical informatics and knowledge engineering tools coupled with solid modality and enterprise solution platforms will provide the much-needed support to providers to keep up with the rapid pace of healthcare management.
Hesham Abboud, MD, PhD, Director of the Multiple Sclerosis and Neuroimmunology Program and staff neurologist at the Parkinson’s and Movement Disorder Center at University Hospitals of Cleveland, Case Western Reserve University School of Medicine