The debate indicates that the pressure to achieve absolute cost saving on a population level has induced great frustration and challenges for participating providers. It is suggested that the emphasis should be placed on value (better quality over the same price) and transparency rather than cost saving.^8,9 Literature review shows that, during initial pilots, bundled payment programs usually have limited effects on the costs.^8,10 Taiwan's experience shows that the 5-year medical payment of the bundled payment group remained stable, but the quality of care was better.⁶ This is not unique as in oncology bundled payment. Achieving cost saving in other Centers for Medicare and Medicaid Innovation programs is difficult too. Only five of the 54 Centers for Medicare and Medicaid Innovation models have resulted in significant financial savings' according to independent evaluations.¹¹ What insurers and providers can focus on is value and transparency to patients. Most APAC countries have universal health care. For the public sector, focus on value could demonstrate a government's commitment to invest into their citizens' health. For the private sector, providers can differentiate in the market by offering better quality of care and price transparency, because patients using the private sector will have financial uncertainty as an extra concern.

Second, build reporting and predictive analytics that are as real time as possible. As alluded in the debate, effective management of cancer requires timely reporting.² While Asian-Pacific countries have gone a long way in defining cancer quality indicators (eg, Japan¹² and Australia¹³), the ability to report these indicators in an automated and timely manner is lacking in most countries.¹⁴ Also lacking is the ability to implement prediction models at the point of care to proactively identify patients risky for adverse events (eg, unplanned hospitalization within 3 months), impeding effective population health management. Researchers from APAC countries have published voluminous prediction modeling papers on various cancer outcomes using retrospective data. However, these models are seldom prospectively validated, not to mention being implemented at routine clinical practice.¹⁵ To achieve these capabilities, it requires advancement in informatics standards (like the US Electronic Clinical Quality Measures, or eCQM), Electronic Health Record, and the workforce.

Third, involve pharmaceutical companies in drug discounts and health economic analytics. Surprisingly, although proponents and opponents both agree that drugs remain the most contentious issue of the Oncology Care Model, pharmaceutical companies are rarely involved in the discussion. Pharmaceutical companies could help in several ways. They could offer discounts on drugs. Discounts could be simply based on the number of committed treatments in a bundle, or based on clinical performance of a drug (outcome-based pricing). Pharmaceutical companies can provide data and health economics expertise to inform the model design, serving as an independent party to the negotiation between providers and insurers. Besides clinical trial data, many multinational pharmaceutical companies have purchased or acquired real-world data. These data could inform the utilization of care, serving as an important reference for model design. It could also help identify risk factors for adverse outcomes. This could become critical for risk adjustment, another contentious topic in bundled payment. We are positive about this multilateral partnership because Asia has been a pioneering region with close collaboration between pharmaceutical companies and insurers in oncology.¹⁶ We could observe a similar trend in the United States, as OCM 2.0 is in discussion with six pharmaceutical companies.¹⁷

Fourth, start small. As alluded in the debate, when the scope of a program is nationwide and for many types of cancers, it creates various challenges, including insufficient adjustment based on cancer subtypes and geographical variation. The alternative is to start small. The bundled payment collaboration between the University of Texas MD Anderson Cancer Center and UnitedHealthcare on head and neck cancer serves as a good example.¹⁸ Its success suggests that insurers in APAC could partner with a small number of providers on pilots that offer manageable risks and administrative burden in a well-understood patient population with predictable treatment pathways.¹⁸

In summary, while the model design details are still under fierce debate, and are going to be different from one location to another, the Oncology Care Model from the United States offers invaluable insights to implement bundled payment models in APAC. We agree with the debate that this is a herculean effort. Start small, build reporting and analytics capabilities, and establish partnership among pharmaceutical companies, providers, and payers. Withhold expectations to achieve saving on absolute cost in the first couple of years. This is a long-term value-based transformation for both public and private insurers and providers.