This analysis is part of the USC-Brookings Schaeffer Initiative for Health Policy, which is a partnership between Economic Studies at Brookings and the University of Southern California Schaeffer Center for Health Policy & Economics. The Initiative aims to inform the national health care debate with rigorous, evidence-based analysis leading to practical recommendations using the collaborative strengths of USC and Brookings. The Commonwealth Foundation has provided a grant to the Brookings Institution to support the work on which this post is based. This post originally appeared in Health Affairs on August 2, 2019.
Medicare spending for physician-administered drugs under Part B has grown rapidly. Between 2009 and 2017, Part B drug spending has increased at 9.6 percent per year on average, with the majority of this caused by price growth. This is an unusual environment in which physicians are in the business of purchasing drugs, administering them, and earning a margin from payment by Medicare and beneficiaries—often referred to as “buy and bill.”
There are few competitive forces to restrain spending for Part B drugs, in contrast to many drugs obtained from pharmacies and financed under Medicare Part D. Even when physicians have a choice among therapeutic alternatives for drugs they administer, under buy and bill there are no incentives to choose drugs that are more cost-effective and no tools, such as formularies, to do that. As a result, drug manufacturers face little constraint in pricing for drugs with therapeutic alternatives—and effectively no pricing constraint for drugs with no therapeutic alternatives.
Recently, this issue has been getting more attention from policy makers. They have advanced ideas to more aggressively use administered pricing, where a public payer sets the price it will pay, as well as ideas to draw on market forces. In an advanced notice of proposed rulemaking(ANPRM) with comment, the Trump administration proposed a Medicare demonstration using both approaches. The demonstration would include a ceiling on what Medicare will pay for drugs based on an international price index (IPI)—analyzed in a recent Health Affairs Blog post—and also a new role for vendors designed to remove buy and bill incentives for physicians.
Separately, the Medicare Payment Advisory Commission (MedPAC) recommended administered pricing changes to address high launch prices and rapid increases in prices for existing drugs, along with a Drug Value Program (DVP) that would work through market forces. The DVP would encourage physicians to affiliate with a vendor that would develop tools to counteract the incentives of buy and bill, put pressure on manufacturers to lower prices, and encourage more efficient prescribing. The vendor would reward affiliated physicians with a portion of savings achieved through a shared savings contract with Medicare.
A voluntary vendor approach incorporating shared-savings contracts with Medicare can be an effective way to reduce spending for physician-administered drugs through market forces, although it is complex to undertake and raises important policy issues related to beneficiary choice. But, under the administration plan to limit Medicare payment rates based on international prices, there is little opportunity also to employ vendors in an effective manner. Vendors would be unlikely to achieve additional price reductions beyond the IPI, and distorted incentives for physicians under the current system could be addressed through other, less complicated, measures.
Physician practices or hospital outpatient departments purchase physician-administered drugs from distributers, wholesalers, or the manufacturers themselves at negotiated prices. Under Medicare Part B, the program pays physicians an administered price based on the average sales price (ASP), a figure published by the Centers for Medicare and Medicaid Services (CMS) that reflects most rebates and discounts. Medicare would normally pay 106 percent of the ASP, but this is reduced to 104.3 percent due to a 2.0 percent budget “sequester” reduction in the Medicare payment. The physician, in turn, must collect the standard 20 percent beneficiary coinsurance.
Commercial payers frequently adopt the ASP with higher markups but have more tools to help control drug spending, employing utilization management tools such as step therapy and prior authorization. Physician practices and hospital outpatient departments can often acquire a drug for less than the ASP, so the Medicare amount results in a profit for the provider. Specialty societies point out that small physician practices may lose money since they pay prices higher than the ASP due to lack of negotiating leverage and have higher unit overhead (mostly storage).
The buy-and-bill system, under which physician practices and hospital outpatient departments earn a margin on the use of drugs they administer, has been criticized because physicians have incentives to choose the drug with the higher ASP in situations where there are therapeutic alternatives. In some cases, physicians will find that their “spread,” the difference between the Medicare reimbursement and the acquisition price they pay, will vary across drugs; they will thus have an incentive to choose the drug with the largest spread. To the extent that more expensive drugs tend to have larger spreads, the add-on and difference between Part B reimbursement and acquisition cost each reward choosing high-price drugs, causing insurers and the patients to pay more.
But the problem does not stop there. These distorted incentives undermine competitive pressures on manufacturers to constrain drug prices. If a manufacturer’s lowering the price of a drug leads to only small increases in sales, then a price cut will not increase revenue. So prices stay high, with incentives to increase rather than reduce prices. This is in striking contrast to drugs obtained by patients from pharmacies, in which formulary incentives for patients (including exclusion of a drug from coverage) have substantial effects on a drug’s volume. Unlike Part B, when drugs in Part D have therapeutic alternatives, market processes have the potential to restrain drug prices net of rebates and discounts.
Creating A More Competitive Environment
Creating a more competitive environment for the use of Part B drugs would require program rules that balance financial incentives, the ability of physicians to trade off price and effectiveness, and the interests of patients. This could include utilization management requirements, such as prior authorization and step therapy; it could also include formularies that group clinically similar drugs into preferred and non-preferred tiers, along with differential cost sharing that rewards patients for using preferred drugs.
Indeed, these tools have long been successfully used in Part D. But Part B and Part D differ in important ways:
- In Part D, drugs are prescribed by physicians but dispensed at retail pharmacies chosen by patients. On the other hand, Part B largely involves drugs in which physicians both prescribe and dispense products through buy and bill, allowing them to benefit from the margin between the reimbursement they receive and what they pay to acquire a drug.
- Beneficiaries in Part D annually choose their drug plan, and CMS makes available detailed information on each plan’s formulary and utilization management tools, so patients can elect a plan with features that best meet their needs. In contrast, were Part B to allow formularies, tiered cost sharing, and utilization management, it would be difficult to make that information transparent to beneficiaries and enable them to consider these factors when choosing a physician.
In addition to resolving these two policy challenges, creating a more competitive Part B drug program would require addressing two additional issues: establishing entities to develop formularies and attract physicians and permitting physicians to share in savings associated with prescribing clinically appropriate but lower cost drugs.
Under current Part B rules, savings associated with prescribing a lower-cost mix of clinically appropriate drugs would accrue to the Medicare program and, through coinsurance, to beneficiary secondary payers, or infrequently to beneficiaries directly—but not to an entity or its affiliated physicians. To provide physicians with some of the savings arising from a more efficient utilization pattern, the entities would need to have shared-savings agreements with Medicare and the ability to pass a portion of the savings on to participating physicians. These could resemble shared-savings agreements for accountable care organizations and for bundled payments. A benchmark would be established for the cost of a patient’s treatment by diagnosis, such as rheumatoid arthritis, and spending would be compared to the benchmark. Entities and physicians could choose to accept either one-sided (bonus only) or two-sided risk.
This basic structure to bring competition to Part B has surfaced in several prominent proposals, including MedPAC’s DVP proposal and as a component within the Trump administration’s proposed IPI demonstration. Both proposals use an entity called a “vendor” that would serve as an intermediary between physicians and drug companies.
Vendors under the DVP would not get involved in physical distribution of the drugs, which would be unchanged, nor would they take ownership of the drugs; they would administer the tools that would influence physician prescribing, and they would negotiate prices. Physician margins from drugs would come from a share of the profits of the vendor, as Medicare would pay the DVP negotiated price of the drug and an administrative fee. This would remove ASP spreads from the equation for participants. Vendors that successfully shifted volume to lower-price drugs would earn profits from a shared-savings contract with Medicare that would reward them for both lower negotiated prices and lower spending from a lower-cost mix of drugs. These profits would be shared with the physician practices and hospital outpatient departments that affiliate with a vendor.
Unlike the MedPAC DVP proposal, vendors under the administration’s proposed demonstration would take ownership of the drugs from distributors; the physician practices and hospital outpatient departments would no longer own the drugs. Medicare would pay the vendor 80 percent of the price determined by the IPI, and the prescriber would be responsible for collecting the 20 percent coinsurance from the patient.
The proposal would end buy and bill. Instead, Medicare would convert the 6 percent add-on under the current system to a fixed “drug add-on amount.” The ANPRM offered options such as paying the same dollar amount for drugs that are grouped into a therapeutic class. Thus, for example, the payment to physicians per injection of anti-VEGF agents for macular degeneration might be the same dollar amount regardless of the specific drug that is chosen.
On the one hand, this proposal would avoid the problems with buy and bill by taking the physician out of the financial transaction except for payments for administration from the physician fee schedule and the add-on payments in lieu of past buy-and-bill margins. However, Medicare would have to establish the rules for compensating the vendors. To the extent that they would be at risk (and could lose money), the initiative would have to include adequate compensation, which would include a margin for assuming risk.
Within the framework outlined above, there are a number of design choices to be made.
How Can Vendors Lower Drug Prices?
As discussed above, a vendor could lower prices through its ability to engage affiliated physicians in using a formulary. If successful, the vendor would have leverage in negotiating drug prices. In the DVP proposal, the shared-savings portion of these gains, along with those from a lower-cost mix of drugs, would accrue to the vendor and its affiliated physicians through the contract with Medicare.
In contrast, under the administration proposal, which precludes the use of tools such as formularies, there is limited ability to negotiate lower prices, perhaps because the IPI would be expected to achieve lower prices directly. Indeed, we are skeptical about whether the vendor component of the administration proposal has the potential to bring additional value, given its lack of tools and the role that administered prices based on international comparisons play in lowering prices by an expected average of 30 percent.
Should Provider Participation Be Mandatory?
Should all providers use vendors, or should it be a choice? Voluntary use of vendors would work better under a DVP-like approach because vendors and providers would have a complex working relationship. Vendor tools for managing use and creating incentives for physicians to choose lower-price drugs need to be acceptable to physicians who work with them and, in theory, their patients. When vendors and providers are working together to create savings to share, physician choice of vendor could be important.
But under a voluntary approach, it might be important to limit the number of entities that Medicare accepts as vendors to reduce administrative complexity for the program and to assure that vendors would be large enough to have leverage with manufacturers. To the degree that large practices or hospital outpatient departments might qualify as vendors (discussed below), the need to avoid fragmentation would argue for allowing only very large practices or outpatient departments to participate as vendors. Should a DVP-like approach work well, making physician participation mandatory—although letting physicians choose a vendor—might be pursued to expand the impact of the strategy.
However, vendors would play a different role under the administration’s proposal. Vendors would take ownership of drugs from manufacturers and be paid by Medicare, and physician financial involvement would be limited to receiving a fixed per drug add-on from Medicare in addition to payment for administration under the physician fee schedule. Physicians would have to choose a vendor to provide Part B drugs in Medicare. But it is not clear what basis physicians would have to prefer one vendor over another other than possible differences in the fees vendors might charge for their services. Thus, whether participation by physicians in arrangements with vendors should be mandatory depends critically on the nature of the vendor role.
Should Physicians Choose A Single Vendor?
When vendors are voluntary—such as in a DVP-like approach—there are compelling reasons to have physicians choose a single entity for all their Medicare patients. Multiple vendors would add complexity for physician practices and likely undermine the vendors’ negotiating leverage with manufacturers. Associating different patients with different vendors could open the shared-savings contracts with Medicare to problematic risk selection. The partnership envisioned in using tools such as formularies would appear to be undermined if physician loyalties were not focused on a single vendor.
This does not preclude having the vendor market evolve with vendors specializing in drugs used by different physician specialties, for example, one vendor might be perceived to be particularly effective for oncology, another for rheumatology. Having specialization emerge would not interfere with requiring each physician to choose a single vendor, although multi-specialty practices might require special treatment, such as being allowed to contract with vendors by specialty.
Should Vendors Take Ownership Of Drugs?
Having a vendor take ownership of drugs—as in the administration’s IPI proposal—creates two substantial problems. First, the vendor would need substantial working capital to cover the period from the date the manufacturer must be paid to the date that payment is received from Medicare. Second, purchasing inventory could add an additional element of risk to becoming a vendor. The administration’s proposal may have included ownership of the drugs by vendors to eliminate buy and bill. But the combination of limiting drug prices through the IPI and paying physicians a uniform dollar amount for a therapeutic class of drugs largely eliminates concerns with buy and bill, making the contribution of creating vendors superfluous from a policy perspective.
In addition to taking ownership of medicines, the administration’s proposal would also allow vendors to choose whether they wish to take possession of the drugs and to provide delivery to practices. Involving vendors in physical distribution does not necessarily contribute to their ability to secure either lower prices for Part B drugs or savings from shifting to lower-cost therapeutic alternatives.
Indeed, allowing vendors to deliver the drugs directly could create several complications. First, many of these medications are fragile and have precise storage requirements, so existing distribution systems would not be easy to replace. Vendor distribution would fragment the physician’s supply chain, with supplies for Medicare patients coming from a different source than supplies for other patients. Presumably, this would increase inventory costs as well since separate inventories would be needed for different patients. It would increase distribution costs by involving additional distributors for a practice. These complications could be circumvented depending on the type of vendor organization, and the administration’s proposal does allow flexibility on which vendors are eligible specifically to minimize disruptions in the supply chain.
Could Medical Practices Serve As Vendors?
Some of the specialties that are the largest dispensers of physician-administered drugs—for example, oncology, rheumatology, and retina subspecialists in ophthalmology—are characterized by a number of very large practice groups. Some of these physician groups, as well as the outpatient departments of large hospital systems, might have the wherewithal to function as vendors, at least as envisioned in the DVP proposal.
Many of these large entities already are sophisticated in purchasing drugs and presumably do well under ASP-based reimbursement by obtaining drugs at prices below the ASP. Under a shared-savings contract with Medicare, large practices or systems might be particularly effective at changing patterns of drug use to respond to the shared-savings incentives. These entities likely are better positioned to succeed as vendors than entities that work with unrelated small practices. Indeed, a large practice might also engage smaller practices in its specialty in the community to participate in its vendor operations.
However, if too many practices were to serve as vendors, it could not only limit negotiating power with manufacturers (discussed above) but also be a significant burden on CMS staff to administer the shared-savings agreements.
Could Commercial Payers Also Work With Vendors?
If Medicare were able to persuade physician practices to work with vendors under shared-savings arrangements, it might be attractive for commercial payers to contract with the same vendors under their own shared-savings approaches. If a physician practice worked with a Medicare vendor, it could be attractive operationally as well as financially to deal with the same vendor and formulary for commercial patients as well.
Medicare and commercial payers could work in parallel—even though they would not formally be working together—enhancing the market power of a vendor. This appears to have happened in settings such as accountable care organizations and bundled payments, where providers have separate shared-savings contracts with Medicare and commercial payers. These contracts would differ in details such as the size of margins above the ASP and how savings are shared, but the incentives to physicians would be similar.
A voluntary vendor approach for Medicare could diminish the incentives of buy and bill and engage physicians in trading off cost and effectiveness of Part B drugs within a class. But it would be a complex undertaking for physician practices, CMS, and patients requiring a significant change from Part B rules. It probably is best seen as an alternative to administered pricing strategies, such as the IPI, based on the shortcomings discussed in this post associated with the administration’s proposal to do both.
An implication of this is that it would be quite difficult to straddle the fence between competitive approaches and administered pricing, at least in this area. Policy makers need to embark on a course, plan it carefully, and perhaps not pursue the “toe in the water” practice of creating a demonstration—better to legislate a new policy and revise it as needed.