ISRA May 2022

Appropriateness of Chest-Abdominal-Pelvis Computed Tomography Use in the Emergency Department

Clara Singer 1 Dorith Shaham 2 Talya Markus 2 Clara Singer 1 Jacob Sosna 2 Jacob Asaf 3 Sharona Vaknin 1 Nama Ahmed 3
1Health Technology Assessment and Policy Unit, The Gertner Institute for Epidemiology and Health Policy Research, Israel
2Department of Radiology, Hadassah Hebrew University Medical Center, Israel
3Emergency department, Hadassah Hebrew University Medical Center, Israel

Background & Aims: We assessed the appropriateness of Chest-Abdominal-Pelvis (CAP) Computed Tomography (CT) scan use in the Emergency Department (ED), based on expert physicians and the ESR-Iguide, a clinical decision support tool (CDSS).

Methods: We included 100 consecutive cases of CAP CT scans ordered at the ED in a tertiary hospital in 2020. Four experts (two radiologists and two emergency medicine physicians (EM)) rated the appropriateness of the use of CAP CT on a 7-point scale. The ESR-iGuide score recommendation for the CT CAP protocol was used to evaluate the appropriateness of the ED decision.

Results: The mean rating of the four experts was 5.2 ± 0.106, (Median-5.5). Based on the mean rate, the experts considered only 63% of the tests as appropriate (using a threshold of 5). Using a three-class variable (1-2; 3-5; 6-7), the Overall Percentage Agreement between the experts was 46.5%. Agreement was higher among EM (51%) experts than among radiologists (44%). According to the ESR iGuide system, for a high proportion of the cases (85%;n=85), CAP CT use was inappropriate, while if imaging of the chest was omitted then the Abdominal-Pelvic protocol was recommended for 65 out of those 85 cases. The Pearson Correlation coefficient between the average expert rating and ESR-scores was 0.09 (p>.05) for the CAP protocol and 0.22 for the Abdominal-Pelvic protocol (p=0.029).

Conclusions: According to both the experts and the ESR-Iguide, inappropriate testing was prevalent. The findings raise the need for unified workflows that might be achieved using a CDSS tool.