Cost-minimization analysis of piperacillin/tazobactam versus imipenem/cilastatin for the treatment of serious infections: A Canadian hospital perspective

BACKGROUND: In 1998 we reported the first Canadian double-blind, randomized , clinical trial involving a comparison of piperacillin/tazobactam (P/T) wi th imipenem/cilastatin (I/C). The present study was conducted to determine the feasibility of replacing VC at our institution. OBJECTIVE: TO describe the outcome of a pharmacoeconomic analysis of the cl inical trial from the perspective of a tertiary acute-care institution. METHODS: A total of 150 consenting adults originally prescribed I/C were ra ndomly assigned to receive either P/T 4.5 g iv (n = 75) or I/C 500 mg iv (n = 75) every six hours. Actual direct medical resources used in relation to the treatment of bacterial infections were prospectively assessed during a clinical trial; these included cost of study and ancillary antibiotics, ho spitalization, diagnostic testing (radiology, laboratory assessments), and labor, as well as treatment of adverse drug reactions, antibiotic failures, and superinfections. RESULTS: While costs for successful treatment courses were similar across treatment arms, hospitalization costs for treatment co urse failures were higher for P/T recipients. Direct medical costs for trea tment courses associated with a superinfection were also higher in the P/T arm. Overall costs for treatment failures with either study drug were at le ast twofold those observed fur successful treatment courses. Mean total man agement cost per patient in the P/T group was $15 211 ($ CDN throughout) (9 5% CI $11 429 to $18 993), compared with $14 232 (95% CI $11 421 to $17 043 ) in the I/C group (p = 0.32), resulting in a mean cost difference of $979. Sensitivity analyses revealed that the superiority of VC over P/T for succ essful treatment of serious infections was sensitive to changes in the cost of hospitalisation and drug efficacy for either drug. CONCLUSIONS: Based on the results: of the clinical trial, P/T and I/C offer similar clinical, microbiologic, and toxicity outcomes in hospitalized pat ients: with serious infections. Under base-case conditions, our pharmacoeco nomic analysis showed that VC was a cost-effective alternative to P/T at th e dosage regimens studied. However, this finding was sensitive to plausible changes in both clinical and economic parameters.