A retrospective study showed that the increase in head CT referrals in an Italian hospital suggests that an increase in the number of unnecessary scan referrals could be mitigated by improving the quality and appropriateness of the requests.
In an article published on 26 August by Insights into Imaging, a team led by Dr. Marco Parillo, a diagnostic radiologist at the Unità Operativa Multizonale di Radiologia Ospedale di Rovereto e Arco in Trento analyzed 2,908 imaging requests, assessing referral quality using the Reason for Exam Imaging Reporting and Data System (RI-RADS), and referral appropriateness using the American College of Radiology (ACR) criteria.
The authors assessed referral quality before assessing appropriateness; they note that they wished to emphasize “the critical need for thorough imaging requests to convey clinical justification.”
Of the 2,908 requests, only 620 (21%) were found to be of adequate quality, according to the RI-RADS metric (RI-RADS categories A or B); 2,288 (79%) were assessed as inadequate (RI-RADS categories C or D).
For the subsequent appropriateness assessment, the authors noted that of the 2,908 referrals, 410 (14%) could not be evaluated using the ACR guidelines due to a lack of clinical data. Of these, 43 had been assessed as RI-RADS A or B and 367 as RI-RADS C or D.
For the remaining 2,498 requests, 1,816 (73%) were classified as “usually appropriate” by ACR guidelines, while 640 (25%) were classified as “usually not appropriate”.
When broken down by quality, 84% of those assessed as RI-RADS A or B (high quality) were appropriate according to ACR criteria, while 70% of those assessed as RI-RADS C or D (low quality) were assessed as appropriate.
When requests assessed as inappropriate were evaluated with clinical findings, 96% of head CTs that had been classified as RI-RADS A or B were negative. Headache (26%) and trauma (25%) made up the majority of these requests. Four patients had positive imaging findings: one finding each of a tumor, subgaleal hematoma, arteriovenous malformation, and intraparenchymal hemorrhage (diagnosed on a follow-up CT scan).
Of requests assessed as inappropriate that were classified as RI-RADS C or D, 99% were negative. Syncope and headache (22% each) were the two most common reasons for these requests. However, there were eight patients with positive imaging findings: three tumors and one subarachnoid hemorrhage were found on the initial CT scan; one aneurysm and two carotid stenoses were found on CT angiography; and one ischemia was found on MRI.
When evaluated in total, 98% of all patients with inappropriate requests were negative; in contrast, among patients with appropriate requests, 84% were negative, with 16% having positive imaging findings. Of these, there were 198 positive findings on the initial CT scan, 29 on 24-hour follow-up CT, 31 on CT angiography, and 24 on MRI.
Chart showing the percentage of positive outcomes (patients with acute cerebral radiological findings) in relation to the appropriateness criteria of the ACR guidelines, across all cases and grouped by the quality of the requests according to the RI-RADS metric. Courtesy Parillo et al, Insights into Imaging. Image available for republishing under Creative Commons license (CC BY-NC-ND 4.0).
The authors noted that headache and syncope were the two leading reasons for inappropriate requests. For appropriate requests, the two leading reasons were trauma (47%) and stroke (23%).
The differences -- and the outcomes in findings -- highlight the need for better education in appropriate use of CT, Parillo et al write. The results showed that when CT was used for appropriate indications, there was a greater likelihood of clinically significant findings. However, the proportion of inappropriate requests in the study was high, which reinforces the suggestion that the uptick in CT scans may reflect unnecessary referrals.
“Inappropriate CT procedures represent an unnecessary expense in economic, environmental, and unjustified radiation exposure terms,” the authors conclude, as well as using radiology staff’s clinical time inefficiently. Better training, a “no-blame culture,” and integrating guidelines into workflows would improve the quality and appropriateness of referrals.
Read the study here.
