|Search CDDW Abstracts|
COST-EFFECTIVENESS OF IV PPI’s IN THE TREATMENT OF NON-VARICEAL UPPER GI BLEEDING FOLLOWING URGENT ENDOSCOPY
K Herba, W Kennedy, AN Barkun
McGill University, Health Centre, Montreal Quebec
Recent clinical studies support the use of intravenous proton pump inhibitors (IVPPI’s) for the treatment of patients with ulcer bleeds exhibiting high risk stigmata at the time of endoscopy, thus requiring endoscopic hemostasis. Amongst such patients, those receiving an IVPPI infusion after endoscopic hemostasis had lower 30-day re-bleeding rates.
METHODS: We assess the cost-effectiveness of starting an IV infusion of pantoprazole (80-mg bolus followed by 8mg/hr for 3 days) for a high risk, non-variceal ulcer lesion after therapeutic endoscopy has been performed. The cost-effectiveness analysis was conducted by creating a decision tree model in Data 3.5. The assumptions of probabilities and costs were derived from the literature, a local cost database, and a national Registry of patients with Upper Gastrointestinal Bleeding undergoing Endoscopy (RUGBE). The unit of effectiveness adopted was the proportion of patients with an episode of re-bleeding. We tested the robustness of the estimates by using both threshold and sensitivity analyses. The time horizon was 30 days following the admission to hospital.
RESULTS: Using the RUGBE data and calculated per diem costs, it was estimated that hospitalisation costs for patients with uncomplicated and complicated ulcer bleeds were respectively CDN$1546.08 and $3275.63 per patient. Over the range of probabilities covered by the 95%confidence interval assigned to the re-bleeding rate, the optimal strategy was found to be one that included the IVPPI infusion versus one that did not. The IVPPI strategy was dominant exhibiting higher effectiveness (17% decrease in re-bleeding) at lower cost ($67 less per hospitalised patient). Sensitivity analyses on the probability of complications indicated that the estimates are robust across a wide range of clinically relevant variables. Assumptions about hospitalisation costs had the greatest effect on the decision to start/not start the IVPPI.
CONCLUSION: Based on the assumptions of our model, the most cost-effective approach is to start an IVPPI infusion for a patient with a high risk ulcer bleed having undergone urgent endoscopic therapy. Further analyses are required to determine the appropriateness of therapeutic approaches and any regional variations in the care provided. RUGBE is a partnership of the CAG and Byk Canada Inc.