Do Hematologists Believe APMs Afford Fair Value? An Open Discussion at ASH

Alternate payment models, bundled payments, risk-sharing, value-based payments—these issues have found a permanent place in clinical oncology meetings. Initially developed for primary care or common surgical procedures, payment models are now being developed for patients with hematologic diseases. On the second day of the annual meeting of the American Society of Hematology, physicians got together to discuss the impact of alternate payment models—proposed by CMS and by private health plans—on clinical practice.

Moderator Steven L. Allen, MD, from the North Shore-Long Island Jewish Health System, Manhasset and New Hyde Park, NY, and chair of the ASH Committee on Practice, said, “Physicians need to be aware of how insurers pay for their services, even if their income is entirely based on salary. Our speakers will address bundled payments, using hematopoietic stem cell transplantation as the model, followed by a review of resource management, since controlling costs is crucial in all payment models, and finally, we will hear from an insurer who will discuss what is happening in this new environment and how insurers plan to handle the transfer of risk from the insurer to the provider.”

Using hematopoietic cell transplantation (HCT) as a model, Michael Lill, MD, from the Cedars-Sinai Medical Center, Los Angeles, said that HCT is a high-cost, high-risk, and high-benefit procedure, with high resource utilization—but patients can be cured.

Lill listed several barriers for entry of institutions into HCT:

• Shortage of trained physicians
• Need for accreditation
• Restrictions rendered by insurance companies, such as the need to be a center of excellence and the need for outcomes data
• Substantial infrastructure needs

Providing a historical context to the evolution of payment models in HCT, Lill explained that payment models for HCT were initiated in the 1990s, but the incentives were poorly aligned, with a 100% risk on the payer. “So the concept of bundled payment came about, to share the risk between payers and providers, and to incentivize providers to be more efficient,” Lill said. It continued to leave HCT as a profit center at most hospitals, and “It helped transplant center directors focus their attention on costs of care,” he explained.

Explaining the nuance of an HCT bundled payment, Lill explained that a transplant episode is divided into the following phases: assessment, identify donor, administer conditioning therapy and other supportive care, and post-discharge follow-up.

The nature of bundles varies from one contract to the next and each phase of the bundle described above involves a technical fee (institution) and a professional fee (physician). The professional fee is very minor, usually 5% to 10% of the cost of care of HCT. “The professional fee can be quite straightforward if it is not included in the bundle,” explained Lill. “However, inclusion in the bundle makes things complicated. What do providers bill the overall bundle? Should they charge the Medicare rate or the PPO rate, or the cash pay rate?” These complications create a need for negotiation and then a separate contractual arrangement with the private practice providers, Lill said.

In addition, long-term management and payment problems pop up for patients who might see a different provider 6 months after their transplant, but are still on the case rate. Providing an example of a transplant patient who might come down with pneumonia and seek care at a local emergency department, Lill said, “We need clear definitions of what is and what is not covered in the bundle.”

Provisions such as stop-loss payments add to the complication. “A typical global contract will specify payments for a well-defined episode of care, specify a rate that will start after the contract episode ends,” Lill said, explaining that stop-loss payments helps ensure that the hospital or institution does not shoulder the entire risk. The stop-loss clause, he said, states that once a certain threshold in charges is reached, the payer will pay a percentage of charges for a particular episode of care.

Lill then provided the pros and cons associated with bundled payments. The advantages of bundles, he said, include:

• Physician control over patient care plan
• Absence of pre-authorization requirements
• Close attention to expenditures and outcomes by those in the know
• Incentive for innovation
• More predictable cost structures for insurers.

The disadvantages, however, include:

• Being in control, the physician is responsible for decisions made, not the payer (especially relevant when holding back care)
• Treatment strategies include cost-benefit-profit analyses
• Physicians need to be more in tune with the business of medicine
• Physicians may not be comfortable with these concepts
• Potential for conflicts of interest

“We have had 25 years of experience with this payment model in the transplant field,” said Lill, emphasizing that while cost decisions are made explicit to providers, we need to ensure that providers are not inappropriately influenced by the payment model. “Outcomes data become very important with bundles and they can also lead to innovations in the clinical practice,” he said.

The next presentation by Joseph Alvarnas, MD, Hematology/HCT, City of Hope National Medical Center, was titled Measuring Episodes of Care and How to Turn Them into Payment. Alvarnas, the editor in chief of Evidence-Based Oncology, asked the question, “Is episode of care more a modality-driven care in patient’s care?”

Emphasizing the influence of the Affordable Care Act on changes in healthcare payments and infrastructure, Alvarnas said that non—fee-for-service payment models are shifting the ownership of the entire care continuum onto providers. “Providers are now responsible for direct care costs, readmissions, and complications of treatment,” he said. While this concept creates negative incentives for ineffective, expensive, or duplicative treatments, it simultaneously shifts the risk onto consumers through higher deductibles, copays, co-insurance payments and out-of-pocket expenses, Alvarnas clarified.

For diseases like acute leukemia, increased risk-sharing funds the alignment between knowledge, risk, and reimbursement, Alvarnas said.

While cancer care costs are rising rapidly, specifically for acute leukemia, which represents 1.7% of all new cancer diagnoses, studies have shown that it accounts for $5.4 billion in cancer care costs. “But do we really know how much it costs to treat someone with acute leukemia?” Alvarnas asked the audience. A lot of these figures, he said, are derived from international trials and additionally they do not consider the high interpatient variability associated with this disease. “Further compounding the complexity is the fact that none of these data are linked to meaningful outcomes. For the cost to be meaningful, it should be linked to how it affects the patient,” Alvarnas stressed.

How can we come up with a payment model that can encompass all of the factors associated with disease treatment modalities, such as patient age, demographics, disease status risk, cytogenetic and molecular risk, and treatment modality-related financial risk, asked Alvarnas.

Citing CMS’ episode of care payment model, he said that the model defines an episode as 6-months of care, with the clock set in motion with a new disease diagnosis, relapse, or disease progression.

“While the sensibilities behind the episode of care payment model are right—namely providing longitudinal care and creating economic incentives for data-driven care delivery— it is unclear if this model is suitable for acute leukemia,” Alvarnas declared, suggesting that the shared savings model might be a better approach.

As we move towards value-based care, we need cost-insensitive care delivery, with an emphasis on achieving systemness in care, said Alvarnas.

The final presenter of the session was Michael Kolodziej, MD, national medical director for Oncology Solutions at Aetna. He provided a flavor for the payer approach to new cancer payment methods.

Kolodziej emphasized that cancer patients do care about the quality of care they receive, and providers and payers should be wary of that.

“The major healthcare challenges today are a rapidly-ageing population, and the growing expense of healthcare, a lot of which is contributed by high drug prices,” said Kolodziej. But it’s not just the sticker price of the drugs, in his opinion, that’s responsible for escalating costs. Cancer is the most costly medical item, he said—not just drugs but the overall care of cancer patients is expensive.

What most health plans care about, Kolodziej said, are end-of-life use of medical services. He shared statistics showing that a third of people spend their life in the ICU in the last month of their life—something that he said was quite unusual a few years back while he was still a practicing oncologist.

He listed a few strategies used by health plans to combat the rising cost of cancer care:

• Pay less
• Manage more (prior authorization)
• Narrow networks
• Increase co-payments
• Incorporate process measures and link them to outcomes (pay for performance)
• Shift risk (capitation)

Kolodziej emphasized that generating value is THE solution in oncology. “We can provide quality care and optimize the right treatment to the right patient at the right time according to the patient’s needs,” he said. Additionally, adherence to evidence-based guidelines decreases cost without a negative impact on outcomes.

Using well-validated clinical pathways also help provide structure and contain costs, he said, but pathways are just a part of the process measures. “Medical homes are also a part of the solution,” according to Kolodziej, and he believes CMMI’s Oncology Care Model can draw parallels to the Oncology Medical Home (OMH). “And OMH is an ACO solution.”

Performance on process measures and outcomes measures can have a significant impact on pathways, explained Kolodziej. “Ultimately, pathways are just a scaffolding that help structure and manage patient care. We have to remember that you cannot improve data that you do not measure,” Kolodziej concluded.