OBJECTIVE: In 1989, Germany was first country to introduce reference based pricing (RBP) to cap drug costs. Initially RBP was only applicable to identical ingredients. In 1992 and 1993, RBP was extended to chemically similar drugs of the same indication. In 2004, the German government seeks to extend RBP to patent-protected drugs. With that measure the government expects major cost-savings. Taking PPI (protonpumpinhibitors) in the treatment of GERD (gastroesophageal reflux disease) as an example we demonstrate that the expected savings are not going to be realized. METHODS: A markov model was constructed to evaluate the effect of RBP on the clinical outcomes and costs in the maintenance therapy of GERD. It is assumed that payments by the patients in case of a drug price above the reference price have an effect of on the continuation of the therapy. Refusal of payments by the patient can lead to a change or discontinuation of the initially chosen therapy. RESULTS: Due to changes in treatment patterns and the more frequent use of less effective PPI, only 18% of the expected savings are realized. Hence the likelihood of remission per year decrease from 79% to 73% and 6% more hospital admissions are to be expected. CONCLUSIONS: Only a fraction of the expected costs are going to be realized, while the overall effectiveness decreases substantially. The inclusion of patented drugs in reference price schemes should be assessed individually.