S1365 Reducing Inappropriate Utilization of CT Angiography in GI Bleeding

Introduction: Computed tomographic angiography (CTA) of the abdomen and pelvis is frequently obtained in patients presenting with gastrointestinal bleeding (GIB). Although there is a role for CTA in the diagnosis of GIB based on ACG guidelines, it can be over utilized which can lead to potential adverse consequences such as increased costs and unnecessary patient exposure to intravenous contrast agents and radiation. We noted that there was a high baseline utilization of CTA for diagnosis of GIB at our institution. Thus, we conducted a quality improvement project aimed to identify trends regarding CTA utilization for the diagnosis of GIB and then implemented measures to decrease inappropriate utilization. Methods: All inpatient consults for GIB were retrospectively analyzed over a period of 17 months. If a CTA for GIB was obtained, its appropriateness was determined based on ACG guidelines. For example, appropriate utilization would include a patient presenting with ongoing lower GIB and hemodynamic instability whereas inappropriate utilization would include a vitally stable patient presenting with hematemesis. If a CTA was obtained for reasons other than for GIB, it was not included in the analysis. Preliminary findings over the first 8 months of data collection found the emergency department had the highest utilization of inappropriate CTA as compared to other departments. A focused education was provided to the emergency department providers regarding optimal utilization of CTA for GIB. In addition, an alert was integrated in the electronic health record (EHR) if a CTA was ordered for GIB. Data was collected for an additional 8 months after these measures were implemented. Results: The number of inappropriate CTA ordered by ER providers decreased by 27% (from 30 in the pre-intervention time period to 22 in the post-intervention time period). The proportion of inappropriate CTA ordered by ER providers as compared to all other providers decreased from 59% to 29% pre- and post-intervention, respectively. There was no impact on inappropriate CTA utilization on other service lines. Conclusion: By implementing focused education along with integrating alerts targeting a specific order in the EHR, we were able to decrease inappropriate utilization of CTA in the diagnosis of GIB. Further education of other service lines and system changes are likely to yield additional improvements in the optimal utilization of CTA and avoid unnecessary health care costs and patient risks.