During her 20 years at University Hospitals (UH) in Cleveland, OH, Beverly Rosipko, Director of Radiology Informatics at the academic health system, has used several different analytics software packages to optimize radiology operations. These tools have been tremendously helpful in managing the 1.5 million radiology exams that UH produces annually originating from 12 hospitals, 40+ ambulatory centers and teleradiology partners.
In 2013, when I first met Beverly, she was a pioneer in using analytics software. For over a decade, she has been using data to improve the efficiency and quality of the radiology services provided across the UH system. Fast forward to 2020, where she is one of the early adopters of workflow orchestration software—taking analytics one step further. Rosipko describes the huge amount of radiology exams produced, together with hundreds of radiology staff, and how she needed to move analytics to the next level.
The analytics dashboard for tailored views and alarms
In early 2019, Rosipko along with several UH team members implemented Sectra’s analytics dashboard, one of the tools in the Sectra workflow orchestration solution. In addition to providing a visualization of the operations for various roles, the dashboard was configured to set up alarms and flagging mechanisms based on specific thresholds and rules. This meant display of a complete transparent analytics overview, while also informing where action was needed. “We went from monitoring and manually analyzing the data, to getting access to a tool that can communicate when something is wrong. It shows exactly where action is needed. So far, this has saved significant amounts of time,” says Rosipko.
The beauty of the analytics dashboard is its ability to be tailored for specific needs. The users of the dashboard today are a combination of radiologist section heads, chief radiologists for each subspecialty, radiology administrators, managers for all the radiology departments, and the analytics staff.
Today, UH uses the analytics dashboard for a variety of purposes. They track weekly volumes by modality and site to identify where workloads are too heavy, and if action is needed. Rosipko points out that the dashboard has been valuable during the COVID-19 pandemic, as volumes have been constantly changing. They also plan to use the dashboard as an alarm tool if certain service level agreements (SLAs) risk not being met. “It guides the radiologists to select the exams with the highest priority, based both on the clinical status, the current workload as well as the SLA restrictions,” describes Rosipko.
The informatics team has also created views to understand reading patterns among various subspecialties. Rosipko explains that the health network includes several community hospitals as well as academic centers, and the analytics dashboard has helped them to understand who is reading which kinds of studies and how to identify areas of improvements. “The most complex exams might be better read by the academic hospitals,” she says.
We noticed that some radiologists read double the number of exams than average. It turned out they picked the newer exams, despite hundreds of aging exams in the worklist.
Rosipko highlights the value of creating tailored views of the dashboard. As an example, she explains that they have created one view to monitor and keep track of all CT head exams, and one for all lung cancer screening exams. “Each section has a radiologist in charge, and they have access to a personalized configured view to keep track of their area. This brings tremendous value since they can quickly identify if certain studies should be rerouted to other sites to balance workflows and meet specific SLAs.”
The tech QA tool for continuous improvement
UH has also incorporated another component of Sectra’s workflow orchestration suite, the technologist quality assurance tool (tech QA). “It is a great solution for providing feedback between staff. Within Sectra PACS, radiologists can directly send feedback to technologists’ supervisors if the quality of an exam is not good enough, or if images are missing,” comments Rosipko. She describes how the tech QA tool helps them keep track of errors on specific modalities, areas, or locations, to quickly find areas of improvement and if training is needed. “All feedback is handled by the supervisor of that group. The system prompts that a resolution is reported back to the radiologist on how the issue has been managed. This provides a closed feedback loop and has considerably improved learning,” she explains.
The benefits of workflow orchestration software
Learning more about their overall business and areas that need attention is the key benefit that UH has experienced from their use of the analytics dashboard and the tech QA tool. “Regarding the quality assurance workflow, the process we used previously was too cumbersome and now, the tech QA tool makes our feedback loop so much more efficient,” she says. They are currently evaluating the impact on the quality, but according to Rosipko, it has absolutely improved. “All in all, these tools have enabled us to much better understand our business. We have learned about radiologists’ reading patterns and exam distribution, and been able to begin to communicate back to staff where it is most needed,” she summarizes.
Another key benefit of the workflow orchestration tools has been the breaking down of silos between departments. This spans over multiple user groups and sites, including workflows that have previously been managed separately, providing a holistic overview for the health system.
The use of automated workflow orchestration is most likely where we will see the greatest benefits of AI in radiology in the short-term.
More opportunities ahead
In the future, Rosipko is certain they will extend the use of the workflow orchestration solution to include automatic reallocation of exams. “This is the biggest opportunity I see for the moment, since reallocation of exams is a very time-consuming task. The use of automated workflow orchestration is most likely where we will see the greatest benefits of artificial intelligence in radiology in the short-term,” she says.
As previously mentioned, Rosipko’s team will also begin using the dashboard as an alarm if certain SLAs risk not being met, something that will guide the radiologists to select the exams with the highest priority.
Additionally, these tools will provide visibility to many others in the organization based on their role. “We must remember that these tools will bring about a culture change. We will continue to shift from Excel sheets and pivot tables to real-time analytics. Every step forward toward real-time monitoring provides significant benefits in terms of efficiency for the organization,” Rosipko concludes.