SIIM veterans talk about the big issues facing imaging informatics

May 16, 2008

It wasn’t intended as a formal roundtable, but a question from the radiology news media -- what are the biggest challenges facing the field -- led some SIIM veterans, participants on a panel assembled by the association, into discussion Thursday of a laundry list of issues facing radiology informatics. Their comments also produced some insights into hot topics to watch in the future.

It wasn't intended as a formal roundtable, but a question from the radiology news media - what are the biggest challenges facing the field - led some SIIM veterans, participants on a panel assembled by the association, into discussion Thursday of a laundry list of issues facing radiology informatics. Their comments also produced some insights into hot topics to watch in the future.

Not all their answers were exactly new. Image overload and better integration between disparate PACS/informatics systems have been topics of interest for several years at least. But the issues seem to be taking on more urgency as technology advances and the demands on imaging increase.

Others were new: How radiologists report their imaging findings will need to change; pay for performance, until now a small factor in radiology, will gain in importance and prominence; and radiologists and vendors will need to work harder to validate what they assert in their reports and software programs.

Reforming the radiology report: SIIM is known for PACS information, but the next wave in imaging informatics will be the way radiologists report their information, said Curtis Langlotz, M.D., outgoing chair of SIIM. A key question will be whether radiology reports can be integrated into the electronic medical record and cross-correlated with other information.

Researchers may want to mine the reports and correlate findings with biomarkers. How the reports are designed will be critical to these activities, he said.

There are already financial incentives to encourage radiology report reforms. Under its Physician Quality Reporting Initiative, the Centers for Medicare and Medicaid Services is offering a 1.5% bonus to practices that report specific types of data, including whether they observe guidelines for reporting carotid stenoses and certain stroke findings, Langlotz said. That program will be expanded and what is now a bonus could become a holdback for practices that fail to meet reporting standards, he added.

George Bowers, MBA, who has worked as CIO at a large radiology practice and a big academic medical center, said CIOs are concerned about the relationship between radiology and the electronic health record. One of the big challenges for SIIM will be defining what imaging informatics is going to mean for the electronic health record.

"It's not just images popping up electronically, it's much more complex than that," he said.

Pay for performance: Right now payers are not holding radiologists accountable for meeting quality standards, but that won't last much longer, Siegel said. As time goes on, radiologists will be required to collect data and report it in a standardized way that no radiology department is able to do at this point.

An example of one possible measurement is radiation dose, he said. The ACR is working on a mechanism for creating national databases to benchmark quality measures. In the case of radiation dose, practices would probably need to capture the information and store it. This is not easy, and, as the pay-for-performance concept progresses, practices will need informatics solutions to gather and provide that data, he said.

Clinical validation: "Too frequently when we look at images, what we think we see becomes ground truth without having ground truth," said Bradley Erickson, M.D., incoming chair of SIIM. "Getting validated data sets is a critical component."

This will be particularly important as informatics moves from displaying images on a screen to having a computer help figure out what's going on," he said.

"If I end up doing a measurement - whether it's the size of tumor or vascular disease - I need to have a mechanism to know whether that's accurate," Siegel said. "I need rigorous science to make sure we move from radiology as an art form, which means producing good-looking images to be subjectively reviewed, to being something we can actually use in the digital, or computer, age."

Right now, software systems often provide widely varying figures for such things as measurements of volumes or carotid stenosis, he said.

System Integration: "When I go into work in the morning, I have to log on to three different systems and I have two keyboards and two mice. It's too much," Langlotz said.

This is something computers ought to do well, but finding ways to do that in a multi- vendor - or even same vendor - environment is a challenge, he said.

Another system integration problem, sharing images, concerned Siegel. The problem is solved right now by putting images on CDs, but they are often incompatible with different PACS. The issue has roused even the AMA to put forward a resolution, particularly supported by neurosurgeons and oncologists, calling for a better system.

Imaging overload: Erickson noted the problem of imaging overload, particularly the high volume of images coming out of today's CT scanners.

Katherine Andriole also cited the issue of imaging overload. Part of the solution will involve finding new ways to navigate the data. Some concepts for doing this may be imported from the gaming industry and Hollywood, she said.