Health Services Research, Clinical Informatics, and Quality of Care
Modeling an oncology outcomes-based contract using a blockchain database approach: Cost and technology considerations.
Background: Cost of care is an ongoing concern for all oncology stakeholders. Outcomes Based or Risk Sharing contracts are increasingly discussed, but difficult and costly to implement. The literature notes barriers of outcomes definition, reliable data sources, and technology enablement. The authors sought to design and test software that could enable faster, lower-cost, and auditable administration of such contracts. Methods: The authors developed software using blockchain databases and transactional proof of work to simulate such a contract and to compare the financial result to usual fee-for-service reimbursement (FFSR). The software processed a synthesized Medicare claims dataset and Average Sales Price (ASP) data from 2008-2010, looking for use of the Bevacizumab(BV)/Carboplatin/Paclitaxel(CP) regimen in non-small cell lung cancer (NSCLC). The contract hypothesized a scenario that offered payers a discount for “underperforming” BV doses (defined as doses given to a patient with < 9 mo. BV duration) and required a bonus payment for “overperforming” BV doses (defined as doses given to a patient with > 14 mo. BV duration). These parameters were selected based on survival data supporting the 2007 FDA approval of BV/CP in 1L tx of advanced nonsquamous NSCLC. Results: The software successfully processed the claims dataset and projected financial results (additional/saved cost) to the payer in this hypothetical contract compared to FFSR. The software also enabled comparison of different hypothetical contracts, inclusion/exclusion rules for claims, and discount structures. The software accurately categorized doses according to the defined logic. Conclusions: Outcomes based contracts have potential for better aligning oncology reimbursement with meaningful results, particularly for costly therapeutics and where patient response or outcome is difficult to predict. While such agreements are recognized as difficult to implement, a software platform that facilitates efficient and scalable design, simulation, and implementation of such agreements, under constraints of real world data availability and sharing, may advance their adoption.