OBJECTIVES: Several outcome-based agreements (OBA) have been signed in the past years with health plans in the Unites States for the reimbursement of the PCSK9 inhibitor evolocumab. We developed a model to determine what factors may impact the outcome measured under these OBAs in patients with athereosclerotic cardiovascular disease.

METHODS: A Bayesian model was built in HOPE (Health Outcomes Performance Estimator), a tool that predicts drug outcomes under different real world scenarios. Assumptions on efficacy of evolocumab versus placebo in terms of hazard ratio to time to first cardiovascular event, defined as cardiovascular death, myocardial infarction, stroke, hospitalization for unstable angina, or coronary revascularization, came from a phase 3 randomised clinical trial (RCT) . Two virtual cohorts of 10,000 patients: one approximating the health plan population, one with characteristics as in the RCT, were created with different distributions of baseline age, weight, smoking status, LDL and HDL cholesterol levels, type of atherosclerosis, use of cardiovascular medications,and other factors including adherence. The impact of factors on baseline risk was extracted from the literature. To account for variability in inputs, we performed 1,000 simulations of cardiovascular events in each of the virtual cohorts.

RESULTS: The model estimates provided a good fit to time to cardiovascular event reported in the RCT. The estimated event rate 6 months after drug initiation in the health plan population was predicted to be higher than in the Phase 3 trial on average, with more variability due to higher variability in drug use and population characteristics.

CONCLUSIONS: We investigated how outcomes chosen to define the OBA terms on evolocumab could be affected in the population of health plan subscribers. Additional variability in population factors impacted the event rates which may in turn impact the financial results of the OBA, depending on the financial terms that were used.