RSNA’s template library brings crowdsourcing to bear on structured reporting

 - Crowdsourcing

CHICAGO—The numbers tell the story of the early successes of RSNA’s radiology reporting initiative: 268 report templates available in the RadReport.org library, including a handful in such languages as Turkish and Chinese—and more than 106 million views and downloads of templates to help radiologists around the world improve their reporting practices.

The latter number is all the more remarkable for the fact that, despite the launch of the library in 2008, RSNA didn’t start keeping count of “checkouts” until a couple of years later.

That’s according to Charles Kahn, MD, MS, a radiologist with the Perelman School of Medicine at the University of Pennsylvania, who presented on the initiative’s accomplishments and new directions last week at RSNA.

Kahn said the reporting initiative is making good headway in its drive to “get people to get a consistent format, to get things that our referring physicians prefer, to maximize our efficiency, reduce our risk of errors, promote compliance with accreditation and certification requirements, and to profit from federal performance incentives where possible.”

He added that the report template library, created by 12 subcommittees of subspecialty experts, is helping radiology practices develop reporting systems that “go beyond the speech recognition systems to really empower the capture of information as radiology becomes more quantitative,” in the process driving value into radiology.

“When we began this initiative, we did not want to put forward some sense that the templates were a standard, that RSNA was dictating to you that this is how you have to report, for example, a chest x-ray,” Kahn said. “In fact, the templates that are in the library are there as best practices. You can change them if you like, you can add things into them and you can modify them to suit your needs.” The library also includes functionality for rating templates, on a scale of 1 to 5 stars, and for offering specific feedback and suggesting improvements.

Kahn stressed that the reports have been coded with RadLex terminology, the vocabulary RSNA has developed for the radiology profession.  

If somebody dictated a report in Turkish using the template for, say, pediatric bone age, and another person dictated a report in Chinese using another template, information from both could be fed into the same database. The coded piece of information that could be extracted out of the database would facilitate the integration of the combined information sources. “People enter a lot of this stuff in free text, which is not necessarily going to be understandable,” said Kahn. “This is part of the tremendous power of having these RadLex terms under the hood, part of the substructure of these reporting templates.”

The radiologist won’t generally see the RadLex terminology when using a template from the library, said Kahn. “It is there to help improve the structure of the report, but it’s not something that you as a radiologist have to see or deal with in any way.”

At the same session, Marta Heilbrun, MD, a radiologist at the University of Utah and a co-leader of RSNA’s reporting initiative, said that real-world use of the template library is producing opportunities to make the library bigger, better and more user-friendly.

“Things are actually getting better now with T-Rex and with MRRT,” she said, referring to the template editor developed for the initiative by Karos Health and the Management of Radiology Report Templates IHE profile, introduced in 2013 and coded in HTML5. Heilbrun explained that the original templates “had some problems" and didn’t work very well.

“We now want your templates to be uploaded so that we can create real crowdsourcing," she said. "What are the elements that we all are using and think are important? We can start talking as an entire community about what a radiology report [should] look like. What are the elements that we all—not just the [ACR] Select experts—think should be there?”

Curtis P. Langlotz MD, PhD, of Stanford University, summarized by stating that structured reporting is a tool to achieve some of the “clinical and quality ends” that radiology has been talking about as a profession for some time.

“Structured reporting is not technology for its own sake, and it really should be led from the top of the department,” he said. “We really need to partner with the vendors to make sure that the technologies are available so that, if you want to bring some of this into your own practice, you