Value-based contracts between manufacturers and payers rely on correct assumptions around clinical benefit. We use the example of Entresto to illustrate potential economic impact of improper benefit characterization under four hypothetical VBC structures. We compare economic outcomes under assumptions drawn from RCT data versus outcomes based on real-world evidence (RWE). VBC structures were an absolute-threshold-based VBC where the manufacturer refunds drug costs (1a) or medical costs of excess outcomes (1b) if the outcome rate exceeds a pre-specified threshold, and a comparative performance contract where the manufacturer refunds drug or medical costs (2a, 2b) if relative performance of drug versus comparator fails an agreed-upon threshold. We drew the absolute and comparative performance thresholds from the PARADIGM-HF RCT, and statistics on risk observed in real-world scenarios from published RWE studies. We estimated the differential cost of care – the basis of a refund amount – if the real-world benefit varied from the benefit recorded in the RCT. For our calculations, we assumed 2 million HF patients, 20% uptake of Entresto, and a US$12,000 cost per HF hospitalization. In PARADIGM-HF, the annual rate of cardiovascular death or HF hospitalization was ~10 per 100 person-years, varying by risk score, and the relative rate reduction versus enalapril was 20%. In some real-world studies risk scores were higher. With higher baseline risk, all four scenarios would require the manufacturer to provide a refund. Under an absolute-threshold-based VBC (1b), a 1% absolute higher baseline risk in the real world versus the trial would translate to 4000 excess HF hospitalizations per year and economic impact to the manufacturer would be ~$29 million. Our results suggest a potential for increased financial risk to manufacturers if VBC terms are extrapolated directly from RCTs. Rather, they should be informed by RWE on baseline risk among treated and comparator patients.