How Technology and Data Can Revive Radiology

CHICAGO - Everyone used to have to come to radiology. No one had to read reports. There was true collaboration. “But then people like me came along, IT geeks with PACS and IT that actually fostered a kind of a peripheralization, a decoupling between that normal collaboration…I’m trying to ask forgiveness as an IT geek,” said Paul J. Chang, MD, FSIIM, of the University of Chicago, at RSNA 2014.

While technology may have fostered the detachment of radiology, Chang said “it’s inappropriate to blame technology for our own failures. We allowed ourselves to be peripheralized, we allowed ourselves to be commoditized, in my belief.”

Technology isn’t just part of the problem, as Chang explained, it is also part of the solution. Radiology might not return to its old day traditions, but IT can be leveraged to cultivate the next generation of radiology.

“We are at a crossroads in everything, when it comes to how health care is reimbursed, how its organized, how its governed, and when it comes to radiology, we are at a crossroads: are we going to become commoditized or are we going to take a high road?” Chang said.

But, Chang said, this challenge is also an opportunity. Radiologists will no longer be valued for just interpreting images, radiologists will be valued for managing the role of imaging in a complicated, risk-sharing system. Radiologists need to be irreplaceable in this aligned model, and radiologists need to provide believable evidence that they are providing that quality, he said.

When it comes to quality in radiology, there is a general consensus that it is difficult to define. To measure quality, he said, radiology needs to be evidence-driven. Evidence-driven radiology, Chang said, needs a fully leveraged electronic-based workflow and practice management infrastructure that can support all of the quality-driven initiatives.

“The goal is…to try and achieve the efficient, measurable improvements in whatever you deem that you want to achieve,” Chang said. “That’s how we engineers think…but the methodology by which you can achieve what you want to do, I think, can be reproducibly taught.”

For example, performance modeling needs to be based on true key performance indicators (KPIs), and critical to this process is business intelligence analytics, Chang said. But that’s not radiology’s current focus.

The accepted framework today, Chang said, is that one cannot improve a process unless one can measure a process.

“This is the problem I have with analytics today…I know we are 15 years behind when it comes to evidence-drive workflow and analytics, and when I look at other industries, they laugh when they see this,” Chang said. “They find it quaint.”[[{"type":"media","view_mode":"media_crop","fid":"31248","attributes":{"alt":"Paul J. Chang, MD, FSIIM","class":"media-image media-image-right","id":"media_crop_9246877301757","media_crop_h":"0","media_crop_image_style":"-1","media_crop_instance":"3293","media_crop_rotate":"0","media_crop_scale_h":"0","media_crop_scale_w":"0","media_crop_w":"0","media_crop_x":"0","media_crop_y":"0","style":"height: 178px; width: 160px; border-width: 0px; border-style: solid; margin: 1px; float: right;","title":"Paul J. Chang, MD, FSIIM","typeof":"foaf:Image"}}]]

Today’s status quo, Chang said, is the scorecard or dashboard. Something that provides feedback for prior performance with a message to do better next time, he said.

Many hospital administrators, Chang said, are unclear about what is a dashboard and what is a scorecard. He defined a dashboard as a prospective monitoring tool that is operational and tactical in scope and provides situational awareness in real-time (are we winning the battle?). A scorecard, Chang said, is a retrospective management tool that is linked to KPIs and is strategic in its perspective (are we winning the war?).

Dashboards are rare in radiology, Chang said, scorecards are much more common, but scorecards aren’t meant to improve process. Chang said that scorecards are easier to build, which is why they are easier to find. Dashboards, on the other hand, need to be deeply integrated into the workflow engine and require interoperability, making for a more complicated system. 

Chang argues that properly integrated dashboards can help radiologists improve efficiency because they are project management tools that are tactical and operational. The problem with dashboards and scorecards, he said, is that they both need to be consumed by people. “Dependency on knowledge workers can be a liability, especially for complex real-time processes,” he said.

The Case For Business Intelligence
Business intelligence analytics (BIA), however, are more than just static or monthly reports retrospectively viewed from your RIS or EMR, he said. BIA requires a comprehensive and strategic perspective.

“Analytics is not something you buy, it’s not something you stand up. Analytics, business intelligence is what you are, it’s who you are, it’s how you do your business,” Chang said. “There is a reason other industries call it B-I-A, and not A-B-I, analytics is the least important in other industries, it’s the business intelligence [that is important], if you get [business intelligence] right, the analytics come free.”

The modern BIA approach doesn’t start with buying, it starts with defining what you are and who you are, he said. He argued that a practice’s vision statement is critical, and goals need to be defined and extracted from that vision to create an appropriate business operational model. “Notice,” Chang said, “you haven’t bought a thing yet.”

“Once you define [your model], then you say ‘what are the meaningful metrics that we can extract from our various IT systems that are germane to these operational goals and our vision?’” Chang said. “It’s a very rational approach, evidence-driven, that’s what we call the key performance indicator, we extract certain things that are meaningful that address directly how we see our business process model.”

“That’s how it is at the University of Chicago, when I get marching orders from my CIO…it has to link to a [key performance indicator] that has to link to an operational model that has to link to one of our goals in our vision statement,” Chang said. “That’s how we understand what we are doing actually makes sense.”

At this point, Chang said, the IT geeks come in and build the system. And then, he said, the most important part: “you do it again, and again, and again, every day, every hour, every second. BIA is not something you buy, it’s not something you set up, BIA is what you do every day…The most important part is defining an efficient, agile and sustainable governance model. That’s the hard part, buying stuff is easy.”

Analytics tools in radiology have too much emphasis on the tools or widgets, Chang said, and they have a radiology-centered or RIS-bias. “[This was] fine for a fee-for-service world where we didn’t have to care about anyone else…we were a revenue generator. Now that we are going to be a cost center, we have to be everyone’s…best friend,” he said. “And that means the kinds of questions I’m going to have to ask from an analytics perspective go beyond what a RIS knows.”