OBJECTIVES : The implementation of value-based healthcare is challenging given varying interpretations and stakeholder perspectives. As medical device evidence is accumulated through clinical studies and real-world studies, perceived value may evolve. Outcomes-based agreements (OBAs) between providers/payers and industry are increasingly viewed as a mechanism to address uncertainty around medical devices outcomes. As OBAs become more common, outcome-risk evaluation takes on greater importance. The objective of this study was to apply basic statistical principles to risk evaluation in OBAs.

METHODS : Hospital-specific data and the peer-reviewed literature were reviewed to identify baseline data required for quantifying risk under an OBA for the reduction of surgical site infections (SSIs). Reviewed data included patient demographics, procedure types, baseline SSI rates, and anticipated device effect parameters. Given these data, the probabilities of meeting OBA targets were quantified by drawing from anticipated statistical distributions.

RESULTS : We anticipated that the device subject to the OBA would reduce surgical site infections (SSI) at the participating hospital with a relative risk (95% CI) of 0.60 (0.52; 0.68). The participating hospital performs 200 procedures/year with an historical SSI rate of 10%. We considered three targets for reducing SSIs in this OBA: 20%, 30% and 35%. Assuming a binomial distribution for the anticipated effect, as well as hospital-specific parameters, we estimated the probability of achieving each respective target to be 90.5%, 77.8% and 68.5%. Considering the required level of confidence for mutual stakeholder benefit, a final target for SSI reduction was set at 20% within this OBA.

CONCLUSIONS : Practical policy to implement OBAs and establish methods to evaluate risks are critical steps to systematic deployment of mutually beneficial OBAs. Risk transparency is beneficial to contracting parties, fostering a spirit of partnership and a shift to discussions of value instead of price alone.