Direct costs analysis of hepatocellular carcinoma treatment using a population-based Melbourne clinical cohort.

Introduction: Many developed countries have identified the increasing incidence and mortality of hepatocellular carcinoma (HCC) as targets for intervention, with various preventive, early detection, and therapeutic options available. Current cost-benefit analyses that inform reimbursement of new therapeutic agents have been derived from administrative databases. We provide real-world costing for a clinical population of patients with HCC. Method(s): Patients were prospectively recruited over a 1-year period, in a previously published population-based clinical cohort from Melbourne, Australia. Direct costings were derived from hospitals for admissions, outpatients, and emergency attendance for 24 months of follow-up. Analysis was performed by phases of disease: initial phase (3 months before to 1 month after diagnosis), terminal phase (6 months before death), and continuing phase (time in between). Result(s): There were 136 patients from three tertiary hospitals, including the state liver transplant unit, representing 50% of the original cohort included in the costings. The total cost of treatment was $6 663 149 over 24 months, with a median total cost of $6046 in the initial phase, $33 737 in the continuing phase, and $19 395 in the terminal phase. The majority of costs were incurred by inpatient episodes (69%), with emergency attendance 22%, and outpatients 9%. Liver transplantation was most expensive overall, with best overall survival (21% of total costs, eight patients, median cost, $199 924; median survival, 24 months), followed by local ablation (19%, 20 patients, $31 011; 19 months), transarterial chemoembolization (15%, 24 patients, $41 576; 18 months), best supportive care (14%, 29 patients, $19 856; 3 months), resection (11%, 19 patients, $43 908; 20 months), and sorafenib (6%, 17 patients, $16 216; 13 months). Treatment received and Child-Pugh score at diagnosis were significant (P < 0.001) predictors of high cost on multivariate analysis. Tumor stage did