09.23.2019

Efficient Cancer Care? Yes. Forced, Problematic Rule? No

A new cost-cutting bundled payment methodology? Good. Directing that model toward 17 different kinds of cancer care for the benefit of patients? Good. But turning upside down the way such payment tests have been constructed in the past – that is, mandating participation rather than seeking a voluntary group of hospitals to test the risky venture – that’s very bad indeed.
  
CMS has issued a proposed rule that includes a radiation oncology (RO) model to test whether prospective, bundled payments for 90-day radiotherapy episodes of care would reduce Medicare spending while preserving or improving care. The problem isn’t with the model, bundled payments or accountable streamlined care – all things in which hospitals are participating every day. The problem arises from the fact that the model would carry high risk and be mandatory for physician group practices, hospital outpatient departments, and freestanding radiation therapy centers that deliver radiotherapy services for 17 types of cancer in certain areas of the country. (CMS will only reveal which parts of the country in its final rule.)
  
Because different hospitals of varying sizes are on different points along the path of transitioning to value-based care, the American Hospital Association (AHA) says a blanket rule should not be forced on all of them. Mandating that the program be up and operating on January 1, 2020 is impractical, AHA added.
  
Most disturbing is the risk-versus-reward of the proposal. “The AHA recognizes that, in crafting the proposed regulation, CMS attempted to achieve a balance between offering incentives for providers who achieve success and fulfilling CMS’s obligation to protect taxpayers and the Medicare Trust Fund,” the association wrote. “However, as proposed, the balance would be significantly misaligned and put the achievement of savings well out of reach for many providers. Specifically, the rule would require providers to take on 100% risk immediately upon starting in the model, without any stop-loss protections or adjustment for actual versus historical case mix.”