|
|
Back to 2016 Annual Meeting
Does Giving Pasireotide to Patients Undergoing Pancreaticoduodenectomy Pay for Itself?
Yuan Fang1, Amiram Gafni2,4, Chu-Shu Gu3, Deepak Dath1, Michael Marcaccio1, Leyo Ruo1, Ved Tandan1, Pablo E. Serrano*1,2
1Surgery, McMaster University, Hamilton, ON, Canada; 2Clinical Epidemiology and Biostatistics, McMaster University, Hamilton, ON, Canada; 3Ontario Clinical Oncology Group, McMaster University, Hamilton, ON, Canada; 4Centre for Health Economics and Policy Analysis, Hamilton, ON, Canada
Background: Postoperative pancreatic fistula is the most common serious complication following pancreaticoduodenectomy. A recently published randomized trial showed that the use of pasireotide could reduce its occurrence by 49% (95% confidence interval: 0.25 to 0.95). The aim of our study was to evaluate if the use of pasireotide in this patient population is cost saving or cost neutral.
Methods: Clinical and cost data were obtained from a cohort of patients undergoing pancreaticoduodenectomy at a single institution from 2009-2013. Health care utilization, including unit costs for hospitalization, operating, and emergency room visits up to 90 days following surgery were obtained from hospital administrative databases. Cost of pasireotide was obtained from the institution’s pharmacy purchaser and it includes the indirect cost of administration. The total cost for patients who did not receive pasireotide (our original cohort) was compared to 10,000 simulated cohorts of patients who received pasireotide (administered at a standard twice daily dose for 7 days following pancreaticoduodenectomy). We simulated these hypothetical cohorts receiving pasireotide by randomly generating an event rate (clinically relevant pancreatic fistula) for each cohort based on the distribution of the relative risk from the original randomized trial. In each simulated cohort, patients with and without events were randomly sampled from our original cohort. Costs for the original and simulated cohorts were summarized and compared using descriptive statistics. Cost analysis was undertaken from the perspective of a third-party payer (i.e., Ministry of Health) in Canada.
Results: There were 279 patients who underwent pancreaticoduodenectomy for which costs were obtained. The mean total cost (primary hospitalization, re-hospitalization and emergency room visits) for the simulated group was CAN less expensive than the group not receiving pasireotide (control) (2.5% to 97.5% quantile: -4,066 to 3,895). The mean cost for the primary hospitalization in the simulated group was CAN,910 compared to CAN,058 in the control, a mean difference of CAN,148 (2.5% to 97.5% quantile: -1,797 to 3,585). The mean cost for re-hospitalizations in the simulated group was CAN,611 compared to CAN,106 in the control, a mean difference of CAN (2.5% to 97.5% quantile: -2,830 to 2,857). The main difference between groups was the cost of the drug, which was calculated at CAN,300.
Conclusions: Despite the added cost of the drug in the simulated sample of patients who were given pasireotide, their total cost was similar to those who did not receive pasireotide. This was mostly due to a decrease in the cost of hospitalization and rehospitalization in the simulated cohort. At its current price, pasireotide is cost neutral. Cost saving could be achieved by lowering the price of the drug.
 Cost difference distribution between control (no pasireotide) and simulated (with pasireotide) cohort. The probability of the total cost difference to be cost saving (cost difference more than 0) is 59%.
Back to 2016 Annual Meeting
|
