CT audit highlights shortcomings of radiology reports

Information about the precise imaging examination and use of contrast are still being omitted from radiology reports, and some referring clinicians fail to include a clinical question for the radiologist to answer, an audit of CT reports has found.

Referrers must be encouraged to provide as much clinical information as possible to allow radiologists to effectively assist with interpreting scans, noted the authors of a new study presented at this week's U.K. Radiological Congress (UKRC) in Manchester. Furthermore, development of a computer-based form can serve to encourage referring clinicians to pose clinical questions for an interpreting radiologist to answer.

"Radiology reports carry the importance of medicolegal implications, hence a clear structured report is vital," stated Dr. Vishal Bhalla, a specialist registrar in radiology at the University Hospital of North Staffordshire (UHNS) in Stoke-on-Trent, U.K. "The development of a computer-assisted request service has highlighted the importance of posing clinical questions for reports to refer to. Also, referring to a uniformly adopted reporting structure could aim to provide more structure, readability, and, in some cases, accuracy to the report."

A radiology report is an essential communication tool for radiologists and referring clinicians, and it should incorporate advice for further management or investigations to pinpoint a diagnosis, he added.

Along with Dr. Biju Thomas, consultant radiologist at UHNS, Bhalla evaluated the quality and contents of reports. Based on the 2006 recommendations of the U.K. Royal College of Radiologists (RCR Standards for the Reporting and Interpretation of Imaging Investigations), they devised local standards. These highlighted the importance of a clear, concise report as it was assumed there is a clear relationship between the structure of reports and their accuracy.

The authors retrospectively studied 102 CT reports over a two-month period in 2011, and sought to determine the cause of any discrepancies or failure to meet the standards set. The outcome was a high quality of reporting standards, and only a small percentage of requests failed to have an adequate conclusion.

Of this sample, 22 requests omitted a clinical question for the report to answer. Where a question was posed, eight reports failed to respond to the question. Of these eight requests, four had multiple unrelated clinical questions, although strictly speaking not all the clinical questions were answered. It could be argued that the most relevant and likely ones were addressed.

In the 12 cases in whivh a conclusion was omitted, 10 of the reports were surveillance scans post-therapy, so only brief reports were required. Two reports failed to include a conclusion, compromising readability and a clear structure, and this remains a possible area of improvement, Bhalla said.

In the 90 cases that did contain a summary, 13 repeated what was contained in the body of the report. However, structure was still maintained, and readability was not compromised. The summary still contained the most salient points.

"Although the standards affecting the description can be quite observer-dependent and can vary with reporting style, the overall description was of a very high standard, often brief, relevant, and helpful," he observed, noting that RCR-based templates have been devised at UHNS for trauma scans to help the structure of reports and make them easy to read.

It is often quite difficult to determine if any suggestion for further imaging is deemed appropriate, so this was omitted from the statistical analysis, according to the authors.

There is a close relationship between the structure of a report and its accuracy, and reports need a conclusion to refer to the initial clinical questions posed, they concluded.