Over the past century, radiologists have witnessed, and even embraced, many revolutionary developments — from digital imaging to PACS and Radiology Information Systems. Yet one aspect of imaging has remained virtually unchanged since Wilhelm Roentgen discovered the X-ray: the radiology report.
Now, experts are questioning why nothing’s been done to improve reports considering their well-known imperfections. Errors in transcription are common, and the written format is cumbersome when it comes to data-mining, coding, and billing.
Has the time finally come to transform the radiology report to meet the specialty’s modern needs? What role will structured reports and multimedia play in this process? ACR members and researchers weigh in on the changes ahead.
For more than 100 years, radiologists have generated “narrative, descriptive reports,” says David J. Vining, MD, professor of diagnostic radiology and medical director of the Image Processing and Visualization Laboratory at The University of Texas MD Anderson Cancer Center in Houston. “As radiologists, we are like art critics,” he adds. “We look at images and describe what we see.”
Yet this traditional method of reporting has been criticized. “Trying to decipher the information from a narrative report is tedious and time consuming,” says Vining, who adds that narrative reports can vary in content and clarity depending on the radiologist.
Annette J. Johnson, MD, MS, associate professor of radiology at Wake Forest University School of Medicine in Winston-Salem, NC, agrees and says that these challenges and errors — often a product of voice-recognition software or transcription — can be costly, both to patient care and a practice’s bottom line. “I’ve read studies suggesting reporting errors are present in 20 to 40 percent of finalized, signed radiology reports,” Johnson explains. “Most errors are minor and don’t impact care. But if the report has dropped a word, and that word is ‘no,’ it’s not trivial.”
Aside from such mistakes, the traditional report is difficult to use for billing, coding, and research purposes, according to Johnson. “It takes a lot of time and effort to manually go through content,” she says. In the era of data mining, this hassle has not gone unrecognized.
As a result of administration and referring physician complaints, many imaging departments encourage the use of structured templates that are more easily searchable. “Structured reports” rely on standardized formats or templates to frame free text or select findings from a menu of choices. But even this uniform design is not without pros and cons.
Structured for Benefits
Several potential benefits support the rationale behind instituting a structured reporting system. According to Johnson, “There is a lot of interest in developing standardized reporting systems. They would use the same format and language, be produced quickly, and yield data that are easily coded or mined for clinical, administrative, or research purposes.”
Many groups are working to make standardized templates a reality. RSNA is in the midst of creating a library of templates for radiologists seeking to standardize report format, content, and structure through the “Radiology Reporting Initiative.” These templates were created by subspecialty expert panels and are available for download at www.radreport.org.
RSNA isn’t the only organization involved with structured reports. Since October 2008, the College has been developing several structured products, which include the Liver Imaging Report and Data System (LI-RADS). LI-RADS has not only provided a comprehensive grading system for liver findings; it will also soon include a structured template and lexicon for liver reports.
“We initially started out to develop minimal reporting guidelines but are now moving toward a template that will include specifics on reporting individual observations and a standard way to create a synopsis,” explains Benjamin M. Yeh, MD, professor of abdominal imaging at the University of California San Francisco, assistant chief of radiology at San Francisco Veterans Affairs Medical Center, and chair of the LI-RADS Reporting Subcommittee.
The LI-RADS templates are in development and will be based on those created by RSNA. Once completed, Yeh predicts that they will make a positive and immediate impact on patient care. “The templates will condense and standardize relevant information so that clinicians can easily triage patients for local therapy, transplantation, or systemic treatment,” he explains. “Secondly, [the LI-RADS templates] will enable powerful retrospective or cross-institutional studies and the creation of a registry to help aid diagnosis and risk stratification [for liver patients].”
Despite these advantages, many individuals in the radiology community are still skeptical of structured reports. In fact, in 2009, Johnson conducted a research study for Radiology to compare the accuracy of residents’ reports using one traditional and one structured system.1 Expecting results in favor of structured reporting systems, Johnson was surprised to find just the opposite: traditional, free-text reports were more accurate and complete than reports using a structured system.
One of Johnson’s primary complaints with the structured system tested in the study, and many similar programs, is the “number of clicks” needed to enter findings. Simply put, the greater the number of clicks, the more such a system is viewed as disruptive to radiologists’ workflow — a major deterrent to using structured reports. Furthermore, Johnson says reports that include only drop-down menus without free-text options are overly limiting. “If you’re going to describe something abnormal … it’s sometimes easier for a physician to review [free text],” she explains. “Many even prefer it.”
According to Johnson, more research and technology development are needed before structured reporting is fully embraced. “The system we tested was difficult to use,” Johnson explains. “At this point, I haven’t seen a fully-structured system out there that’s user friendly enough [to be adopted]. It would be a mistake to force radiologists to use systems that don’t work very well.”
Toward a More Image-Centric Report
In addition to workflow concerns, Vining has another grievance about radiology reports, both traditional and structured: neither is linked to images. “An image is worth a thousand words, yet we expect a clinician to obtain a complete understanding of a patient’s condition in just a few words,” explains Vining, who is also the developer of ViSion, a multimedia structured reporting system. “Images and reports are often disconnected,” he says, “and you can’t always believe what you read in a report without looking at the actual images. Certain diagnostic entities, such as mixed response of tumors to targeted therapies, aren’t always captured accurately in a narrative report.” For a referring physician to fully understand a radiologists’ interpretation — especially when the diagnosis or finding is complex — images are crucial, says Vining.
If images are that valuable, imagine what multimedia content, including images, video, and audio descriptions, could add to reports. It’s possible with Vining’s ViSion software app, which can be downloaded onto any PACS or 3-D workstation. As a radiologist reviews images, the ViSion software captures screenshots and records a radiologist’s verbal descriptions of image findings. “It’s sort of like Facebook for medicine,” says Vining, because image “tags” describing the anatomy and pathology of a finding are extracted from the verbal description and associated with the image finding, similar to the way a photo on Facebook can be tagged with the names of family and friends. “I’m not displaying the images being interpreted by a radiologist; I am simply taking ‘snapshots’ from any PACS or 3-D workstations,” Vining adds.
These snapshots are fed into a database of image findings. “We can link these key images to anatomical locations on a graphic of a patient’s body,” says Vining. “So, in a single view, we can visualize the entire radiological history of a patient. If a finding is linked to prior studies, we can generate disease timelines over the course of a patient’s disease at that location.”
Vining’s multimedia reporting system is certainly innovative, but neither structured nor more creative reporting forms are likely to achieve widespread use until imagers believe these tools can be used efficiently and accurately.
Ultimately, despite the great need to revolutionize radiology reports, it appears that additional developments in structured reporting have yet to adhere to the famous design principle, “form following function.” The radiology report of the future will one day be realized. But first, concerns with limiting templates and workflow must be addressed so as to better meet the needs of radiologists, physicians, and their patients.