Skip to main content
Premium Trial:

Request an Annual Quote

PMC Paper Asks Policymakers to Consider Impact of Alternative Payment Models on Personalized Rx

physician payment

NEW YORK (GenomeWeb) – If alternative payment models are structured in a way that doesn't allow doctors the flexibility to prescribe and gather evidence on innovative drugs and tests, they could end up impeding access to molecularly targeted treatment strategies, warns a Personalized Medicine Coalition report.

One of the main objectives of the Affordable Care Act is to reduce rising healthcare costs while providing Americans quality care. The traditional fee-for-service system of payment is often blamed for driving costs, where physicians are rewarded for the volume of care they provide — in the form of office visits and administered tests — rather than the quality of care. Increasingly, payors are moving away from fee-for-service payment to alternative payment models (APMs), and policymakers are hoping that these new mechanisms will place the focus not just on reducing costs (as earlier capitation models did), but also on improving quality of care and patient outcomes.

Molecularly targeted, individualized treatment strategies, if implemented in an evidence-based manner, have the potential to reduce costs and improve patient outcomes, according to the PMC. However, as the PMC pointed out in its recently released report, early designs of pilot programs aimed at studying the impact of APMs — through mechanisms such as bundled payments, accountable care organizations, and medical homes — aren't usually structured to capture the value that personalized treatment strategies can add to patient care.

According to Andrew Shin, author of the PMC report, two dominant events in the healthcare sector are currently emerging independently: the proliferation of personalized drugs and advanced diagnostics driven by a growing understanding of the genome; and the rapid move among payors to implement APMs. But to date, "personalized medicines have been largely outside the grasp of APMs," Shin, director of healthcare policy and life sciences at consulting firm ML Strategies, told GenomeWeb.

Meanwhile, as part of broader healthcare delivery reform efforts, the Obama administration said earlier this year that it wants the Centers for Medicare & Medicaid Services to make half of the Medicare payments to doctors and hospitals through APMs by 2018. HHS has set even more aggressive goals of tying 85 percent of fee-for-service payments to quality-improvement and cost-reduction metrics by 2016 and 90 percent by 2019.

In addition, Congress passed a bill in April to scrap the problematic Medicare payment formula under which doctors were facing a 21 percent payment cut. This bill, which the President has now signed into law, also includes ways of improving and measuring quality, reducing wasteful spending, and incentivizing physician participation in APMs. Although participation in APMs is voluntary under this law, "I'd argue that a 5 percent bonus and then possibly a reduction later for physicians who don't participate is not really voluntary," Shin said. "That's saying, 'You will do this.'"

As APMs expand in scope, stakeholders are concerned that they will continue to focus on tying payment to lowering cost and not to whether medical interventions — tried and tested methods and newer, innovative strategies — are providing value to patient care. CMS and private payors are thinking about this, and have asked stakeholders to weigh in on how to factor in drugs, devices, and diagnostics when gauging the total cost of care as part of payment models. But in Shin's view, stakeholders haven't yet ironed out the metrics around value. Absent this, APMs may end up discouraging providers from prescribing often expensive personalized therapies with the help of advanced diagnostics that lack the level of clinical evidence backing their use compared to older interventions.

"We have yet to have a real discussion among policymakers and stakeholders of what value is," he said. "We always talk about moving from volume to value, but if you don't define value in a way that is universal, then you run into some serious problems."

Challenges for personalized medicine

The PMC report cites payors' reported experiences with several APM demonstration projects and pilots to highlight some potential challenges for the implementation of personalized medicine. For example, CMS's Medicare Shared Savings Program (MSSP) allows doctors and hospitals to form so-called accountable care organizations (ACOs), under which providers agree to meet certain cost and quality metrics for a group of Medicare beneficiaries. ACOs can earn "shared savings" when spending is below a historical cost benchmark for that particular Medicare population. But data released by CMS in 2014 on the MSSP shows that of 49 ACOs that qualified for shared savings, 59 percent performed below the national average quality measure for those programs.

This "demonstrates that ACOs do not necessarily need to improve quality to earn payment," Shin wrote in the report. If the impact of quality measures on physician decision-making is a fuzzy aspect of the ACO model, then this doesn't bode well for personalized treatments that may fall outside quality measure thresholds in these programs in the first place. "Although personalized medicines often represent the potential for significant quality improvement, depending on how quality measures are designed and payment parameters are structured, the ACO model could end up limiting access to them," the PMC paper states.

Another difficulty for the adoption of personalized medicines is bundled payment, a scheme that provides payment to a healthcare provider for a single episode of care for a defined time. Under these efforts, when doctors spend below the payment level they can pocket the difference in cost, but if they go over the limit, they are responsible for the difference. "Within these various models you have a timing issue," Shin said. "Bundled payments traditionally go for 30, 60, or 90 days. The most you'll see is for 120 days. But we know that with many personalized medicines, the benefits are not going to be seen [in that time frame] … A lot of these gains are going to be years down the road."

Holding doctors accountable over such short periods pits the APM incentives against the goals of personalized medicine, Shin said.

Ultimately, current APMs are looking at tried and tested treatment modalities, measuring the waste, and incentivizing doctors to improve on that. But this doesn't account for new treatment modalities, such as personalized drugs and advance diagnostics, that might "change the entire game," Shin said.

"I don't think there is necessarily anything wrong with the former way of doing things, because it will improve quality and reduce costs," he continued. "But the problem is you're saying, 'Only within the technologies and therapies available in that time in history is what we're measuring you against.' It's not accounting for new technologies and innovations that are rapidly coming to market."

Flexibility for innovation

One example of a bundled payment pilot in the report is a 2009 UnitedHealthcare effort in which five medical oncology groups agreed to select specific chemotherapy regimens based on evidence that they were the best options for treating 19 clinical episodes of breast, colon, and lung cancer. The oncologists had to stick to these treatment strategies 85 percent of the time, deviating only when a patient was contraindicated for a regimen or enrolled in a clinical trial.

With this pilot, UHC was trying to show that this bundled payment program could sever oncologists' financial dependency on the amount of expensive drugs they prescribed. After three years, data collected from the pilot showed a 34 percent reduction in cancer treatment costs, but a surprising 179 percent increase in spending on chemotherapy drugs. Despite the greater spending on chemo treatments, oncologists saved around $40,000 per patient.

This pilot suggested, contrary to conventional thinking, that greater spending on drugs doesn't necessarily have to translate to higher overall treatment costs. "Utilizing therapies can be cost reducing," Shin observed. And this may also prove true for high-cost personalized therapies, particularly in cancer, he suggested. But the way APMs currently measure quality or value doesn't provide much leeway for doctors to discuss innovative treatment options or prevention strategies with patients.

"If you're not giving providers the ability to go outside the historical cost of a particular procedure or treatment, then you really don't have the ability to use more innovative, new treatments that may be outside that historical cost," he said.

One of the biggest challenges that personalized medicine has is to be able to disseminate information as close to real time as we can.

Moreover, under some APMs, insurers provide financial incentives to doctors for pursuing certain pathways of treatment. Anthem (formerly WellPoint) gives doctors $350 for selecting a treatment strategy on the insurer's pathway program and for coordinating care for the patient. Personalized medicine options do show up in certain pathways in this program, such as Roche/Genentech's Tarceva (erlotinib) and Boehringer Ingelheim's Gilotrif (afatinib) for first-line metastatic non-small cell lung cancer patients with EGFR-mutated tumors.

Anthem, which spends $5.4 billion on cancer care for members, estimates that this pathway program will yield modest savings of around 4 percent. While the impact of this program on long-term patient outcomes remains to be seen, Shin again highlighted the risk that too-stringent pathways might bar use of personalized drugs and advanced tests to identify best responders.

In order to be included in a pathway, the interventions must have evidence backing their use in a clinical setting. But traditional, large randomized studies for diagnostic tests or molecularly targeted drugs indicated for a small subset of patients can be a long and expensive proposition. In this regard, collecting data on individual patient outcomes through the development of registries, for example, could help gather such evidence. While CMS has recently granted conditional coverage to certain molecular diagnostics, requiring data collection on patient outcomes via registries, healthcare providers aren't currently incentivized under APMs for submitting and sharing data in these registries.

At a February meeting hosted by the FDA on regulating next-generation sequencing tests, Levi Garraway of the Dana Farber Cancer Institute pointed out that data-sharing "takes work" for medical institutions, and often these institutions don't have people assigned to such a task. To this, Heidi Rehm of Harvard Partners, who is one of the experts leading NIH's variant data sharing and classification effort, quipped "imagine a CPT code that said, 'shared your variant with ClinVar; enter your patient in a case-level repository.'"

"One of the biggest challenges that personalized medicine has is to be able to disseminate information as close to real time as we can in order to give providers the latest information possible and utilize technology to the fullest extent," said Shin. But he noted there hasn't been much effort from payors to offer incentives for sharing genetic data, particularly since pathway and other APMs generally deal with interventions for the general population.

"We're doing a disservice to the field of medicine by not building in flexibility, as well as infrastructure," Shin said. This would "improve the data that is available to providers and patients, so they can use these technologies and grow the body of knowledge and have a better idea of how these technologies can be utilized so they lower costs."

A way forward?

In the report, the PMC points to one APM idea — patient-centered medical homes — as providing the best example of a payment model that might facilitate personalized care, particularly in oncology. The medical home provides comprehensive care, including prevention and wellness programs; coordinates care within hospitals, the home and the community; and attempts to improve quality of care by following best practices and measuring performance.

Under the patient-centered oncology medical home scheme, for example, providers receive a fixed per-member per-month fee for care management in addition to a fee-for-service payment. Because the care for the patient is coordinated and comprehensive, including different specialties, and the care management fee helps cover what the fee-for-service doesn't, medical homes stand to have a "modest impact" on the utilization of personalized medicines, PMC believes.

The ideal APM for spurring the use of personalized treatments, in Shin's view, would have to have a few key features, such as a focus on measuring patient outcomes, and measuring quality improvements over as long a window of time as possible. "In terms of bundled or clinical pathways you'd try to provide as much flexibility to adopt new technologies if the patients and providers so choose," he noted.

One way to factor in innovative therapies would be to carve out an exception within APMs for patients that doctors feel might be good candidates for diagnostic testing and subsequent personalized treatment. The costs wouldn't count against them if they measure these patients' outcomes and collect the data. "After a number of years, you'll have enough of a historical cost baseline to be able to include a genetic profile and targeted therapy in that cost calculation," he said.

Another way to encourage use of personalized treatments, Shin said, would be through a "medical home-blended payment model" that measures outcomes over a longer period and employs a regional shared savings model. Because patients go to different physicians' practices, a shared savings model that includes participating providers in a region would more accurately account for the total cost of care and quality for each patient, Shin explained. Such a model will also give doctors the room to prescribe personalized drugs and diagnostics and incentivize data sharing on newer technologies.

"At the end of the day, there's no perfect model ... short of complete global population-based payment," said Shin. "And of course, even then, there are the details."