UnitedHealthcare's two-year pilot program to test whether a bundled reimbursement scheme can reduce the high costs of cancer treatment is slated to report results later this summer.
However, while costly pharmacogenomically guided treatments are common in oncology, the preliminary information released on UHC's pilot program doesn't indicate whether it will yield any cost-effectiveness and clinical utility information about personalized medicine interventions.
In November 2010, UHC launched its "episode payment" pilot, which set out to evaluate a reimbursement model that pays oncologists upfront for an entire cancer treatment program based on the expected cost of standard treatment. The pilot involves five large, undisclosed medical oncology groups that each selected specific chemotherapy regimens they believed to be superior in treating 19 clinical episodes of breast, colon, and lung cancer. The oncologists are required to comply with these agreed-upon treatment strategies 85 percent of the time. They can deviate from the plan only when a patient has a contraindication to the regimen or is enrolled in a clinical trial.
In a Health Affairs article published in April, Lee Newcomer, UHC's senior VP of oncology services, highlighted the bundled payment model as a way to shift physicians' incentives from a model that rewards them for prescribing costly treatments. In the article, Newcomer noted that although the number of effective cancer therapies on the market has continued to grow over the years, the current "buy and bill" reimbursement system for chemotherapies is no longer sustainable amid changing treatment modalities and current pressures to reduce healthcare costs.
"Chemotherapy drugs now target specific genomic abnormalities; about 150 effective drugs are available for cancer therapy; and new surgical techniques minimize negative consequences while producing better results," Newcomer wrote.
Yet despite these advances in treatments, the reimbursement scheme for cancer chemotherapy "remains unchanged," Newcomer wrote in the Health Affairs article. "Medical oncologists, the physicians who prescribe and administer these drugs, buy them at wholesale prices and, in effect, 'sell' them to payors at a profit."
Although cancer-related death declined 22 percent for men and 14 percent for women between 1990 and 2007, the cost of care continues to rise. According to National Cancer Institute figures cited by UHC, annual direct costs for cancer therapy were $104 billion in the US in 2006, but are projected to balloon to $173 billion by 2020.
The Patient Protection and Affordable Care Act includes mandates for insurers aimed at reining in runaway medical expenditures. One strategy proposed in the law is the episodic or bundled payment scheme, under which insurers give a fixed amount to healthcare providers covering a number of major components of patient care. Medicare is slated to launch a bundled payment pilot effort in 2014.
Through UHC's pilot program, the insurer is attempting to cut the ties between oncologists' income and the use of high-cost drugs. Under the current "buy and bill" system of payment, doctors purchase chemotherapies and seek reimbursement from insurers for drug costs based on average sales price plus an additional percentage of the ASP to cover the costs associated with administering the treatments.
Medicare reimburses oncologists for the ASP plus 6 percent, but some surveys suggest that physicians charge private payors as much as 18 percent plus ASP. Under such a payment structure, doctors stand to earn more by administering higher volumes of expensive drugs.
In its pilot project, UHC is paying oncologists for the expected cost of acquiring and administering the treatment plus a "patient care fee" for each of the 19 cancer episodes. "We wanted to reward people that did consistent evidence-based care and we wanted to keep oncology incomes at present levels," Newcomer said during a presentation on bundled reimbursement at the American Society of Clinical Oncology annual meeting last year. "We wanted to break the dependency in the oncology world between drug selection and income for the practice."
Newcomer emphasized that while bundled payments might preserve physician incomes, it may reduce revenue for the oncology practice. "I said incomes, not revenues," he noted during the ASCO presentation. "We have got to pay attention to what you spend as an oncologist. So, revenues may go down, but we plan to keep income levels the same."
A move away from the "buy and bill" model and toward the bundled payment scheme will likely have implications for the adoption of pharmacogenomically targeted drugs, which tend to be very expensive and require the aid of genetic tests to identify which patients will respond to treatment. For example, Pfizer's non-small cell lung cancer Xalkori costs more than $115,000 per year and the average Medicare reimbursement for the companion ALK mutation test is $440. Similarly, a six-month course of Roche/Plexxikon's melanoma treatment Zelboraf costs $56,400, and the companion BRAF mutation diagnostic costs between $120 and $150 per test.
Proponents of personalized cancer treatments have said that PGx strategies can in the long run save healthcare dollars by avoiding treatments and unnecessary toxicities in patients unlikely to respond. However, others have pointed out that in order to ensure that these new costly treatments are truly being given to patient populations most likely to see a benefit, more research is necessary to compare the efficacy and cost of PGx drugs and genomic tests to standard cancer interventions.
In the past, Newcomer has spoken about the need to conduct comparative effectiveness research before broadly adopting personalized medicine strategies. Two years ago, just before launching the bundled payment pilot, Newcomer defined personalized medicine at a conference on comparative effectiveness research as the practice of "selecting evidence-based therapies based on the unique characteristics of the patient" (PGx Reporter 12/8/2010).
"It's evidence-based ... It does not [mean] we have a carte blanche to try the theory outside of the clinical research and clinical trial areas," Newcomer said at the event, which was hosted by the ECRI Institute. He further defined "comparative effectiveness research" as the process of "developing the evidence to support personalized medicine decisions."
At the ECRI conference, Newcomer cited an example involving the use of Herceptin — a drug that is often upheld as the poster child of personalized medicine — to point out unnecessary healthcare spending in this arena. He pointed out that although published evidence suggests that 52 weeks of Herceptin treatment yields a reduction in the risk of breast cancer relapse comparable to 12 weeks of treatment, "we continue to treat with a full year of [Herceptin] therapy."
UHC did not reveal whether any of the treatment modalities in its pilot project include PGx strategies, such as Herceptin for the treatment of HER2-overexpressing breast cancers, or prognostic genomic tests such as Oncotype DX or MammaPrint, which help doctors discern whether patients are at high risk of recurrence and need chemotherapy. However, the structure of the pilot certainly offers an opportunity to garner much needed comparative effectiveness data for genomic interventions.
The medical oncology teams in UHC's pilot project have agreed to meet certain performance measures with regard to their chosen treatment modalities — for example, in terms of the total cost of care, relapse rates, survival rates, time to progression, and hospital admissions. Each year the groups meet and compare how their patients did on the various regimens.
"If the data identify a best practice, UHC anticipates that all groups will adjust their therapies accordingly," Newcomer wrote in the Health Affairs paper. "With different groups choosing different drugs but meeting annually to compare results, the system provides for real-time comparative-effectiveness evaluations for competing medications."
Newcomer told PGx Reporter via e-mail that "there are no rules or guidance on genetic testing for the pilot payment program – it is left to the discretion of the attending physician."
Early results from the pilot project also haven't provided any hints with regard to how the oncology teams might be using PGx treatments and genomic tests. In the Health Affairs paper, Newcomer reported generally that all five groups chose to treat early-stage breast cancer with docetaxel and cyclophosphamide chemotherapy, but the groups' treatment costs "varied by 100 percent."
At the ASCO annual meeting last year, Newcomer reported that 14 months into the pilot, the oncologists have "had no problem making 85 percent compliance between the five groups … but there has been a fairly substantial clinical trial enrollment." From UHC's perspective, he noted this was "a good thing." It is likely that even if PGx strategies don't feature heavily in the treatment modalities for the bundled payment pilot, patients will have a fair opportunity to receive molecularly targeted treatments under investigation in clinical trials.
Tackling Patient Incentives
In discussing the pilot project, Newcomer has previously acknowledged that the bundled payment strategy may not be "the final answer" to reducing the cost of cancer care. "It's another example," he said. "It, too, has its flaws."
Indeed, other groups are looking at different ways to better align the oncology reimbursement model with the complex therapeutic landscape.
In a paper published in the Journal of the National Comprehensive Cancer Network in January, researchers led by Jonas de Souza of the University of Chicago observed that a fixed, episodic payment scheme addresses physicians' incentives to administer certain treatments, but doesn't necessarily change the cost structure for patients.
The researchers point out that the price per milligram of Avastin is the same for patients with metastatic colorectal cancer, who stand to experience a median overall survival benefit of more than four months, as it is for patients with advanced non-small cell lung cancer, who can hope to see a median overall survival benefit of only two months. Similarly, the drug price remains unchanged when it is given for metastatic breast cancer patients, who in one clinical trial experienced a median progression-free survival benefit of more than five months – a finding that hasn't been confirmed by other studies.
"I'm fine with bundled payments where you pay a fixed amount of money for the whole package," De Souza told PGx Reporter. "But I still don't think that Avastin in the first line of metastatic breast cancer should be valued the same and priced the same as metastatic colon cancer."
Furthermore, while the bundled payment strategy influences physicians' prescribing behavior, it does little to change patient attitudes regarding high-cost cancer treatments with limited impact on outcomes, De Souza and colleagues point out.
In the JNCCN article, the authors advocate so-called value-based reimbursement designs that waive or lower insurance copayments for people based on how well the intervention impacts their quality of life or survival. Essentially, the authors believe that value-based insurance design could be a complementary strategy alongside other insurance reform ideas, such as bundled payments.
"The bundled payment tries to cut incentives to the provider" to prescribe expensive drugs, "but it still does not cut the incentives for the patient to ask [for] more ... You need a way of prioritizing [treatments]," De Souza explained. "I think that's where the value-based insurance design would come in."
For example, despite all the data attesting to the fact that breast cancer patients whose tumors overexpress the HER2 protein respond to Herceptin, up to 20 percent of women receiving the drug in the US did not have test results for HER2 status, the researchers point out in the paper. "Value-based insurance design, by varying reimbursements based on available evidence and different values, would promote the use of [Herceptin] in patients with HER2/neu-positive metastatic breast cancer … compared with use in patients who are either not tested or who would not benefit from the drug," the authors suggest.
Payors are starting to adopt more stringent rules to make sure that expensive pharmacogenomic treatments are being adopted in a cost-effective manner. For example, most insurers have instituted prior authorization schemes around prognostic genetic tests, such as Oncotype DX and MammaPrint, to stipulate under what circumstances patients should receive testing. Additionally, insurers are also closely tracking whether patients are genetically tested before receiving personalized cancer treatments.
For example, a recent survey of oncologists reported no barriers to reimbursement for BRAF gene mutation testing for melanoma patients, because payors most likely want to make sure that the costly BRAF inhibitor Zelboraf is given only to those with a BRAF-mutated tumor. Furthermore, the survey suggested that doctors are genetically testing melanoma patients in earlier stages of the disease to help with treatment planning should the illness progress (PGx Reporter 5/16/2012).
Still, to date, value-based insurance design strategies haven't been readily implemented because the focus of such programs is on changing patient behavior – a difficult task.
"There is the drug company who wants to sell their drug. You have the physician who wants to treat the patient and may have incentives to prescribe specific chemotherapies. There is a patient who wants to be treated, and the patient a lot of time wants to be treated at any cost, even if the treatment has a 1 percent chance of making him better," De Souza said.
Value-based insurance design is not used that readily "because it's hard [to implement a cost-containment strategy] that depends on giving incentives to a patient to accept a therapy," he reflected.