The Future of Radiology Reports: How Structured Reporting Is Rewriting the Rules

When Cincinnati Children’s Hospital Medical Center made the bold decision in 2010 to launch a structured reporting system, leaders within the radiology department knew they were bucking more than 100 years of history. That’s how long prose reports have been the radiologists’ definitive work product, a fierce source of professional pride and personal identity that had stubbornly resisted change even as new reporting techniques sprouted around them—techniques that could provide greater consistency and more robust, mineable reports that facilitate faster payment.

“We knew there was this feeling among radiologists that to standardize was to take away a lot of their autonomy,” says David Larson, MD, MBA, who was director of quality improvement in the hospital’s radiology department at the time and is now associate chair of performance improvement for the department of radiology at the Stanford University School of Medicine in Stanford, Calif. Larson, who spearheaded the change initiative at Cincinnati Children’s Hospital, was aware that persuading a sizable group of radiologists in his organization to jettison the traditional free-text reporting style in favor of one built on fixed templates, a common lexicon and pre-populated fields could pose a greater challenge than the technical demands of converting to the new system.

That is, until Larson and his team played a strategically clever hand: Consensus building. They created a structured reporting committee with members from all divisions who were charged with developing report templates in line with the panel’s general guidelines. 

“A big part of it was giving our radiologists the opportunity to weigh in,” Larson says. “We balanced the need for department-wide standardization with the flexibility for them to be able to dictate abnormal findings as they saw fit.”

That thoughtful approach proved pitch perfect for the department’s 36 radiologists. Initial skepticism quickly gave way to widespread acceptance and adoption as 66 percent of all radiology studies had structured reports available on the dictation system within the first six months; that number jumped to 94 percent within two years.

In addition, feedback to the leadership team showed that most radiologists at Cincinnati Children’s Hospital had developed a preference for structured reporting department templates. Perhaps most revealing of all, they said they believed structured reporting had positively affected their efficiency. Even referring physicians became aware of the advantages. “When [physicians] started requesting report modifications from us, they were amazed at how quickly and uniformly we could change our reporting to accommodate their needs,” Larson says.

Despite the prospects of greater consistency and less variability, structured reporting has hardly touched off a groundswell movement. While many academic institutions are known to be using or experimenting with the format, the record elsewhere is spotty, even with the advent of speech recognition software allowing for automatic report population. Many radiologists are still suspicious that any system that requires them to essentially fill in the blanks or check off boxes will improve the quality of radiologic reporting or, ultimately, patient care. There is even the fear it could compromise quality by promoting a fragmented approach to reporting. More specifically, a smorgasbord of menus and templates (not to mention catchy and colorful graphics) could end up distracting radiologists, impairing their cognitive reasoning and forcing them to miss the big picture. Finally, converting a radiology practice to structured reporting requires a huge upfront commitment—both technologically and organizationally.

Structured Reporting, Defined

But structured reporting is hardly an alien concept for radiologists. Most have used structured headings such as “clinical history,” “comparison,” and “findings” for years and have become quite facile with the process. Still, a giant chasm exists between the use of common headings and full-blown structured reporting, which typically features standardized or “constrained” language and consistent formatting that incorporates pick lists and drop-down menus.

The authors of a 2014 article in Applied Radiology described a three-tier system: the first contains common headings, the second takes it a step further with sub-headings and the third requires the use of standardized language, pick lists, buttons and other form elements (Appl Radiol. 2014 August;43:18-21).

“The important thing is that there be uniformity, organization and consistency to the reports,” says Justin Cramer, MD, assistant professor at the University of Nebraska Medical Center in Omaha, and lead author of the article. “If you get to the point where you’re using standardized language, then you can do data mining and statistical analysis on virtually all your medical data. That’s the holy grail for structured reporting, though hardly anyone is currently doing that.”

Structured reporting is not without inroads. The Breast Imaging Reporting and Data System (BI-RADS) was the first example of a structured reporting lexicon, and it remains widely used.  But breast imaging admittedly encompasses a limited range of clinically relevant pathology, making it particularly well suited to structured reporting. When the field widens, the benefits of structured reporting may become more problematic. “In the case of the chest radiograph, there is simply too much anatomy and the range of pathologic processes is almost as wide as medicine itself,” wrote Richard B. Gunderman, MD, PhD, and Logan R. McNeive of the Indiana University School of Medicine in a 2014 commentary for Radiology (Radiology. 2014 Oct;273(1):7-9). “The situation is magnified even further in abdominal CT, which includes at least portions of most anatomic and physiologic systems.”

One way around that challenge is to develop lexicons tailored to limited ranges elsewhere in the body. Here, the American College of Radiology has been a leader through development of the Liver Imaging Reporting and Data System (LI-RADS) for reporting liver masses and the Head Injury Imaging and Reporting and Data System (HI-RADS) for traumatic brain injury. And some in the industry believe that RadLex, a comprehensive lexicon created by the Radiological Society of North America (RSNA), is well suited to become the standard language of structured reporting. But Marta Heilbrun, MD, MS, vice chair of the RSNA Radiology Reporting Committee, sees RadLex as more of a “road map” than a final destination.

“RadLex attempts to be broad enough to serve as an anchor for synonyms with the same underlying meaning, like ‘hypoattenuating’ and ‘less attenuating,’ rather than dictate that only one word is possible,” she says. 

In that context, Heilbrun, associate professor in the department of radiology and imaging sciences at the University of Utah and a co-author of the 2014 Applied Radiology article, views RadLex as a “good underpinning” for the evolving library of structured reports, which the Radiology Reporting Committee is creating. These “best practice” reports are intended to serve as a starting point for internal template development by radiologists and IT teams anxious to improve their report quality through standardizing format, content and structure. The library currently consists of more than 200 radiology report templates.

Reimbursement as a Driver

When the radiology department at NYU Langone Medical Center in New York decided nearly five years ago to improve the traditional narrative style of its own reports, it opted to create new templates, with each sub-specialty group deciding on content rather than tapping into the RSNA resource. The medical center’s structured report for abdominal ultrasound examinations, for example, was developed by members of the imaging section incorporating input from societal guidelines and the school’s billing department. The format consists of a tabulated template with each organ listed on a separate line. The template, in turn, was made available on the radiology department’s dictation software program.

What drove NYU Langone to structured reporting was more than a whim; it was a matter of reimbursement and charge capture.

“When we looked at our data we found that almost 10 percent of our completed abdominal ultrasound reports were being incompletely coded," says Kristine Pysarenko, MD, assistant director of quality for the department of radiology. “In coding terminology, they were ‘limited.’”

For an abdominal ultrasound to be considered a complete exam, she elaborates, eight elements need to be described in the report—liver, gallbladder, bile ducts, kidneys, pancreas, spleen, inferior vena cava and aorta. Omit any of these elements and the exam must be coded as "limited." That, in turn, can lead to a decrease in reimbursement by payers ranging from 2.5 percent to 5.5 percent, according to one study.

NYU Langone’s answer to the problem—a structured reporting system—provided a handsome payback. A follow-up study showed that the average number of erroneously coded limited abdominal ultrasound examinations fell from 14.67 to 2.58 per month after implementation of structured reports.

Pysarenko says there was another major benefit of making the change. “Our referring physicians really love it because they know they can find the information right away,” she says. “It’s well organized, and that’s why radiologists love it, too. [Structured reporting] gives us the facts right away in a concise format.”

It didn’t take long for Coastal Radiology Associates, serving eastern North Carolina, to also uncover a strong bottom line advantage to structured reporting. Since moving to the new system more than six years ago, the 15-radiologist practice has been in full compliance with CMS’ Physician Quality Reporting System (PQRS), which pays an annual bonus for adhering to the quality reporting program’s guidelines. (In 2015, the program began applying a negative payment adjustment to individual professionals and group practices that did not satisfactorily report data on quality measures). 

According to Catherine Everett, MD, MBA, managing partner of Coastal Radiology, PQRS (and more recently, MIPS) requirements are now built into the practice’s structured reporting system, removing the onus from radiologists to remember them.

“It won’t let me sign the report unless I’ve met all the requirements,” she says, and that reinforcement has helped Coastal qualify for the maximum PQRS bonus each year.

In addition, structured reporting has opened up valuable data mining opportunities for the North Carolina radiology practice. In one instance, it was able to pull from its database cases coded for stroke and identify those where physicians at the three hospitals it serves were using MRI with contrast to rule out stroke. “Probably 80 percent of the hospitalists were ordering contrast, which is very expensive and can involve some risks,” Everett says. “We pointed that out to doctors who were writing these tests inappropriately, and since then have reduced usage to the point where we rarely see one that gets contrast.”

Adding Value

Heilbrun thinks resistance to change and measurement is one of the biggest hurdles structured reporting has to clear to become mainstream.

“Physicians resist the idea of being measured, yet going to a templated report structure will give us the tools to actually compare them,” she says, noting that the government is increasingly interested in this comparative detail.

Pysarenko sees another force at play, one that could soon lead to an upsurge of interest in structured reporting over the next five years. “We know referring physicians prefer a structured format, and if it’s going to make their lives easier and allow us to provide better care,” she says. “As radiologists, we need to constantly add value to our services.”