Free-Form to Structured Reports: Wave of the Future

With increasing numbers of images to read, and more information to integrate into reports, shouldn’t the radiology report structure change with the times? This is what researchers from Memorial Sloan-Kettering Cancer Center concluded in their study, Improving Communication of Diagnostic Radiology Findings through Structured Reporting, to be published in the July print issue of Radiology.

The study compared content, clarity and clinical usefulness of 330 randomly selected conventional (free-form) and structured radiology reports of body CT scans. These scans were read by three surgeons, two medical oncologists, three radiologists and three radiology fellows, none of whom read the same reports. Nonradiologists read reports relevant to their specialty. Each physician read 15 reports from each format, rating them from 1 (very dissatisfied or confusing) to 10 (very satisfied or very clear).

Researchers found that the conventional reports had a mean content satisfaction score of 7.61 out of 10, and a mean clarity satisfaction rating of 7.45. Structured reports scored 8.33 for mean content satisfaction, and 8.25 for clarity. Nonradiologists were more satisfied with the standardized reports than the radiologists, but the difference wasn’t found to be significant. The physicians gave more “10” ratings for clarity on structured reports (44 ratings at this level) than for conventional ones (13 ratings).

The CT scan reports used in this study were all from 2009, all from Memorial Sloan-Kettering Cancer Center. Their radiology department adopted standardized reports mid-year, so they were able to use equal numbers of both types of reports for the study, written in a similar timeframe.

Co-author, David Panicek, MD, Vice Chairman for Clinical Affairs for Radiology at Memorial Sloan-Kettering was intimately involved in changing the reports to standardized ones at the cancer center. He said that while there was some initial resistance from some doctors, acceptance went to about 80 percent very quickly.

New trainees especially “took to it like fish to water,” he said. “People’s ability to express themselves in English is declining. New trainees are less able to write cohesively.”

Authors describe conventional reports as free-form text, starting with an introduction (the exam type and clinical history), followed by the main body (findings), and then an overall impression. Structured reports, on the other hand, use a template with standardized headings, similar to a check-list style report that many referring physicians prefer. It also uses standardized language, to reduce miscommunication and make the data easier for research and mining. While structured reports for breast imaging have been used for almost 20 years, it hasn’t been adopted in radiology as a whole. 

An unintended byproduct of the standardized reports at Memorial Sloan-Kettering Cancer Center, was its helpfulness in oncology. Panicek said that a lot of the imaging they do are serial examinations, looking at how things change during therapies. Since they need to look at prior reports, the radiologists can immediately look to see what was said previously about the various organs, like the kidneys or the lungs. “In previous reports, people ramble on in giant paragraphs and it was impossible to find things.”

Panicek said that the response from clinicians at his center has been uniformly great. “The clinicians find it easier to scan,” he said. “They have to review pathology and laboratory data, and radiology is only one part of that picture. They appreciated being able to pick out the things of interest to them.”

There are a few downsides to standardized reports. Some physicians feel that standardized reports are less creative. Panicek noted that one fellow in their facility said ‘This takes the romance out of making reports.’ Do we go into radiology for romance? No, it’s for science! There are some people who enjoy the language,” he said, however “the old way (of writing reports) is mind numbing.” Panicek said that in the conventional reports, sentences constantly started with “there is” when describing things.

Panicek said that for 95 percent of reports, standardized ones work well. However for a complex case, where a tumor infects several organs, “it can be awkward,” he said, noting that the radiologist would have to describe the tumor in one organ section, and then say “see above” for the other organs.

He noted that their hospital has gone “full steam ahead” with standardized reporting. “Everyone must use templates,” he said. He estimated that 15 percent of other facilities in this country are using them, but the movement is going in that direction. Getting acceptance from the physicians depends on the mindset of the faculty, and the leadership. He added that the Radiology Society of North America (RSNA) has committees developing templates, but the limiting factor is software.