OBJECTIVES : In the outcome-based risk sharing agreements (OBRSA), performance in a defined patient population is tracked over a specified period of time at the individual patient level and the amount or level of reimbursement is determined on the basis of the outcomes achieved. We applied OBRSA models to the national level data of two targeted drugs reimbursed in the Czech Republic.

METHODS : Total, 661 mBC patients and 1201 first line mRCC patients treated with everolimus and sunitinib respectively were retrospectively evaluated according to the lengths of treatment and volume of reimbursed drugs in the Czech National Register of Reimbursed Health Services 2014-2019. The lengths of treatment in the real-world clinical practice (RW) was compared to an outcome in a corresponding registration clinical trial and enabled us to classify each patient from the RW to the group of responders or non-responders to the treatment. Different scenarios of cost-sharing between payer and manufacturer according to responders / non-responders ratio were modelled afterwards.

RESULTS : In the model when 90% length of treatment in CT (4.95 months) separates responders and non-responders, 43.1% patients treated with everolimus would be reimbursed by manufacturer which made 14.8% of total cost of treatment (in 80% and 70% model: 36.6% and 33.4% patients which made 11.3% and 9.8% of costs). In sunitinib patients, the model when 90% length of treatment in clinical trial (9.9 months) separates responders and non-responders, 57.2% patients treated with sunitinib would be reimbursed by manufacturer which made 23.6% of total cost of treatment (in 80% and 70% model: 52.7% and 47.4% patients which made 19.9% and 16.0% of costs).

CONCLUSIONS : We proposed and tested feasible reimbursement model on national level data that closely align price and value of reimbursed product and reduces the risks associated with high-cost pharmaceuticals with uncertain real-world outcomes.