CHICAGO--Radiology is witnessing the maturation of digital image management technologies and has entered a new phase, according to Paul J. Chang, MD, chair of radiology informatics at the University of Chicago Medical Center, during a session on Nov. 29 at the 97th Scientific Assembly and Annual Meeting of the Radiological Society of North America (RSNA).
Chang dubbed the next phase of digital image management as PACS 3.0, or “meaningful innovation for meaningful use.” Image management, he said, needs to go beyond the walls of the enterprise and also recognize that the emphasis is no longer on images. Radiology needs to deliver a value proposition.
Dashboards and scorecards, the primary product of business intelligence, play a significant role.
Chang noted that the technical aspects of radiology dashboards were actually “quite easy,” when he entered the business more than a decade ago. “Back then the problem was essentially image management. Now, it’s a lot more complex.”
Those complexities come in multiple forms: workflow management, enterprise image sharing, patient-centric data sharing and non-radiology image datasets. In addition to internal pressures, radiology must grapple with external economic and quality factors. Hence, said Chang, the era of PACS 3.0 is here.
Central to PACS 3.0 are new workflow tools to support a more complex imaging environment. Business intelligence and analytics are critical, he said.
A fair number of radiology departments are stuck in 1960’s era business intelligence, utilizing static reports from the PACS, RIS and EMR as business intelligence. “We need to catch up to our colleagues in other verticals,” said Chang.
Dashboards and scorecards
Emphasizing that the time for business intelligence is now, Chang outlined a basic process for bringing analytics to radiology.
Departments need to define goals and form a vision as they create a business model for dashboards and scorecards. Although even experts occasionally interchange the terms, dashboards and scorecards do differ, said Chang.
Dashboards are tactical in scope. They are rich with operationally focused performance monitoring tools and leverage graphics, charts and gauges. For example, a reading room dashboard is typically integrated into daily workflow via PACS or RIS and shows tactical data such as studies waiting to be read or reports requiring correction.
Scorecards, in contrast, are strategic in nature and focus on performance management and utilize metrics, such as key performance indicators (KPI), related to goals. Scorecard data tools stress strategic awareness of the department. For example, a scorecard typically displays information such as modality utilization, physician referral patterns or turnaround time. These data, said Chang, should be tied to a referenced goal or KPI.
Despite the traditional segregation between dashboards and scorecards, Chang said newer systems are blurring the distinction. For example, a system might provide the dashboard perspective of radiation dose to a specific patient during a single exam as well the scorecard perspective of longitudinal exposure to patients.
Chang’s formula for business intelligence
Radiologists, unfortunately, fall a bit short in training when it comes to business intelligence, said Chang. “It’s one area where consultation might be helpful.”
The core of business intelligence is KPIs, which serve as a framework for data aggregation and analysis. Chang cautioned against oversimplifying metrics, warning that overly simple metrics may not provide a useful picture of operations. In addition, the number of KPIs should be limited and manageable.
KPIs often require consolidation and correlation of multiple data elements. Chang admitted that data aggregation, extraction and storage also represent a potential pitfall, either because the department can’t access the data in traditional health IT systems or the data are not meaningful.
“A separate IT entity is usually required,” he said. Chang delved into the informatics infrastructure, explaining that service-oriented architecture can be used to extract data from existing systems and normalize it in a middle layer before outputting a graphical user interface for the dashboard or scorecard.
For example, at the University of Chicago the business intelligence system extracts orders, pathology results and additional data, and feeds them to radiologists in both a dashboard format (e.g. the pathology