Non-Automated Intervention to Improve H. pylori Testing in Active Peptic Ulcer Disease does not Work

Abstract

Purpose:Helicobacter pylori (H. pylori) is a common cause of peptic ulcer disease (PUD). Evidence shows that ulcer remission rates for duodenal and gastric ulcers are much higher in patients successfully eradicated of this infection. The American College of Gastroenterology recommends that all patients with active PUD be tested and treated for H. pylori infection. Our clinical experience suggested testing rates in this clinical scenario were suboptimal at our institution, so we designed a quality improvement project focused on increasing the testing rates of H. pylori for patients with bleeding PUD.Table: [1075] Patients presenting with signs of perforation after diagnostic/therapeutic colonoscopiesMethods: An e-mail communication was sent to all staff that perform endoscopy in patients with active bleeding, requesting they provide clear recommendations for H. pylori testing in their endoscopy reports if PUD was identified (previously recommendations were never included in these reports). All charts for patients undergoing endoscopy over the next two months were reviewed to identify patients with bleeding PUD. After a one month washout period an e-mail reminder was sent and charts were reviewed for an additional two months. Results: In the initial review, 16 patients were identified with bleeding PUD. H. pylori testing was appropriately completed in 9/16 (56%) patients, however, only 5/16 (31%) had recommendations included in their endoscopy reports (vs 0% during baseline analysis, P<0.001, Fisher exact test). When recommended in the endoscopy report, H. pylori testing was done in 5/5 (100%), while only 4/11 patients (36%) without recommendations were tested. After a second e-mail communication, an additional 25 patients with active PUD were identified. Testing was done in 16/25 patients (64%). Recommendations regarding H. pylori testing were made in 11/25 patients (44%) and performed in all 11 of these patients (100%), while only 5/14 patients (36%) without recommendations were tested. The rate of recommending H. pylori testing after the second e-mail was not significantly improved from previous (P=0.52), but remained significantly improved from baseline (P<0.001). Conclusion: Testing rates for H. pylori in patients with clinically significant bleeding due to peptic ulcer disease are excellent (100%) when recommendations were explicitly stated in endoscopy reports. However, compliance with implementing this non-automated intervention was low, and there was only marginal improvement in compliance by sending an additional reminder to staff. Poor clinician compliance is a major limitation of any intervention reliant on a non-automated clinician response. In this situation, an alternative would be an automated message system that would provide the testing recommendation to ordering physicians if their patients are found to have bleeding PUD.