Policy Expensive Medicines in the Netherlands from 2006 to 2012

Spending on expensive medicines in Dutch hospitals was from managed 2006 to 2012 through a policy that listed specific expensive medicines for which extra budget was available. Coverage with evidence development, for the Netherlands a new policy instrument, was implemented for the medicines listed on the policy. In this paper the Dutch policy expensive medicines will be discussed in two parts. In the first part the policy is described including the history, the intention, the associated procedures and assessment criteria. Also the execution of the policy in practice is discussed. To be added to the policy, and to stay on it after four years, medicines were assessed on therapeutic value, cost or budget impact, real-life use and cost-effectiveness. Although the guidelines ask for a cost-effectiveness estimation, in many cases the initial application for addition to the policy expensive medicines was approved without a well described incremental cost-effectiveness ratio (ICER). This led to uncertainties about the cost-effectiveness (CE) of all four reassessments that were finalized until mid-2013. Nevertheless pragmatic solutions, such as price and risk-sharing arrangements, are being negotiated after the reassessments. The second part of this paper describes the initial assessment of cetuximab in third-line use in chemotherapy refractory colorectal cancer (CRC) patients, for which no estimation of the ICER was presented. Despite that the application was approved and cetuximab was added to the policy for this indication. To investigate if a robust cost-effectiveness estimation was possible a Markov model was created to estimate the CE of cetuximab in third line metastatic colorectal cancer (mCRC) using data available in the public domain. The model estimates that compared to best supportive care (BSC) the ICER of cetuximab is €191.417/QALY. In a subgroup of patients with KRAS WT genotype for which cetuximab shows better efficacy, the ICER is more favourable: €144.716/QALY. Deterministic sensitivity analyses show that the model is most sensitive to changes in data on survival. Probabilistic sensitivity analyses show that the actual ICERs might be somewhat lower than the primary analysis showed. The Dutch system of assessment of new expensive medicines has been thorough by taking into account the relevant factors for decision making. However, when implementing coverage with evidence development as a policy instrument it should be clear at the start which data needs to be collected for an effective reassessment after four years. More stringent CE assessment at the start would have been able to prevent uncertainties at the reassessment. This research shows that, at least for cetuximab in third-line mCRC patients, it is possible to give a robust estimation of the cost-effectiveness of a drug, even without the availability of patient level data.