OBJECTIVES : The price of imatinib for Chronic Myeloid Leukemia (CML) raised societal questions on fairness at introduction. Since then, several parameters changed, including patent expiration. This led to the following objective: to model the varying cost-effectiveness of imatinib versus standard of care for first-line treatment of CML in the Netherlands over time, from 2004 till break-even point (Incremental Cost-Effectiveness Ratio (ICER) below the Willingness-To-Pay (WTP) threshold) due to new available/updated data and guidelines.

METHODS : A cost-effectiveness model for CML was developed containing four health states: chronic CML, accelerated phase, blast crisis and death. A lifetime horizon was used. Data input varies over time based on new available/updated data, including costing manuals, CML guidelines, transition probabilities, costs and utilities. Efficacy data was obtained from the pivotal IRIS-studies. Costs were discounted at 4% and outcomes until 2009 at 4%, thereafter 1.5%. One way and probabilistic sensitivity analyses were performed to assess robustness of the results. A break-even analysis was used to show the time period for which the implementation of imatinib earned back the reimbursement costs, transparently demonstrating the societal return-on-investment of imatinib.

RESULTS : The ICER of imatinib was deemed not cost-effective at the time of reimbursement in 2004 for a WTP threshold of €9,076/Quality-Adjusted Life Year (QALY), however would become cost-effective later, using the updated WTP threshold of €80,000/QALY. In 2010, the ICER substantially decreased due to the changed outcome discount rate. In 2016 another substantial decrease was observed due to patent expiration. The ICER increased 4 times due to an increased price. The societal break-even point is 16 years.

CONCLUSIONS : This analysis shows that the cost-effectiveness of a product changes over time due to various factors. In the case of imatinib, the cost-effectiveness improves over time, demonstrating that assessing the cost-effectiveness only at moment of introduction is not sufficient.