Cost effectiveness of imatinib compared with interferon-α or hydroxycarbamide for first-line treatment of chronic myeloid leukaemia

Objective: To evaluate the cost utility of imatinib compared with interferon (IFN)-alpha or hydroxycarbamide (hydroxyurea) for first-line treatment of chronic myeloid leukaemia. Design and Setting: A cost-utility (Markov) model within the setting of the UK NHS and viewed from a health system perspective was adopted. Transition probabilities and relative risks were estimated from published literature. Costs of drug treatment, outpatient care, bone marrow biopsies, radiography, blood transfusions and inpatient care were obtained from the British National Formulary and local hospital databases. Costs (f, year 2001-03 values) were discounted at 6%. Quality-of-life (QOL) data were obtained from the published literature and discounted at 1.5%. The main outcome measure was cost per QALY gained. Extensive one-way sensitivity analyses were performed along with probabilistic (stochastic) analysis. Results: The incremental cost-effectiveness ratio (ICER) of imatinib, compared with IFN alpha, was F26 180 per QALY gained (one-way sensitivity analyses ranged from F19 449 to F51 870) and compared with hydroxycarbamide was F86 934 per QALY (one-way sensitivity analyses ranged from F69 701 to F147 095) [f1 = $US1.691 = E1.535 as at 31 December 2002]. Based on the probabilistic sensitivity analysis, 50% of the ICERs for imatinib, compared with IFN(x, fell below a threshold of approximately E31 000 per QALY gained. Fifty percent of ICERs for imatinib, compared with hydroxygarbamide, fell below approximately E95 000 per QALY gained. Conclusions: This model suggests, given its underlying data and assumptions, that imatinib may be moderately cost effective when compared with IFN(x but considerably less cost effective when compared,with hydroxycarbamide. There are, however, many uncertainties due to the lack of long-term data.