Editor’s Note: This analysis is part of the USC-Brookings Schaeffer Initiative for Health Policy, which is a partnership between the Economic Studies Program at Brookings and the USC 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. Adler and Fiedler’s work on this piece was supported by a grant from Arnold Ventures. This piece was updated on May 25, 2023 to correct the characterization of a June 2022 analysis published by the Medicare Payment Advisory Commission.
On April 26, 2023, the House Energy and Commerce Subcommittee on Health held a bipartisan hearing on strategies to improve transparency and competition in health care markets. At the hearing, the Subcommittee considered seventeen bills on a range of topics, notably including proposals to expand site-neutral payments for certain ambulatory services and to improve health system transparency.
This article analyzes the site-neutral payment and transparency proposals in depth, and then briefly comments on other proposals related to the site of care delivery and prescription drug rebates. We argue that many of these proposals, particularly those that would expand site-neutral payments in Medicare and increase transparency around the ownership of health care providers, would improve the U.S. health care system or enable research that could lead to such improvements. We also discuss how several of the proposals could be improved and how others could create unintended consequences or new administrative burdens that may outweigh the benefits that lawmakers are aiming to achieve.
Site-Neutral Payments for Ambulatory Care
Medicare pays different amounts for ambulatory services depending on where a service is delivered. Medicare’s total payments for a given service are typically much higher when a service is delivered in a hospital outpatient department (HOPD) rather than a physician’s office (roughly twice as high, on average, in 2016). Payments for a given service are also higher when a service is delivered in an ambulatory surgical center (ASC) rather than a physician’s office, although payment differences are typically smaller.
Paying more when a service is delivered in an ASC or HOPD rather than a physician’s office often makes little sense. To ensure appropriate access while containing program costs, Medicare payments should generally reflect an efficient provider’s cost of delivering care. Thus, to justify paying more for a service in facility settings, the clinical needs of the patients treated in facilities must differ in ways that make delivering that service more costly.[1] For the types of services commonly delivered in physician offices (e.g., office visits, imaging, and drug administration), it is hard to see how large cost differences could arise, especially since the differences between patients treated in HOPDs and physician offices appear modest.
Setting Medicare payments based on the lowest-cost site of care would save money for the federal government. It would also save money for Medicare beneficiaries because the premium for Medicare Part B (which covers ambulatory services) is set to cover 25% of Part B costs and because Medicare applies 20% coinsurance to Part B services, which is borne either by beneficiaries directly or by whoever pays for their supplemental coverage. This change would also reduce incentives for hospitals to acquire physician practices, and for physicians to seek hospital employment rather than independent practice, which could generate savings in commercial insurance markets. There is also evidence that reducing Medicare payments can directly reduce commercial payment rates.
The Bipartisan Budget Act (BBA) of 2015 (and subsequent amendments) and a 2019 Trump administration rule together implemented site-neutral payments for a subset of outpatient services: all office visits at off-campus HOPDs and other services delivered at off-campus HOPDs established after November 2015. But Medicare still pays more than it would pay in a physician’s office for ambulatory services delivered in ASCs and on-campus HOPDs, plus many services delivered in off-campus HOPD established before November 2015.
Major site-neutral proposals
Two bills considered at the hearing would substantially expand site-neutral payments in Medicare.
The first would align payments between HOPDs and freestanding physician offices for all off-campus services starting in 2025, effectively eliminating the BBA 2015’s grandfathering of any off-campus HOPD established before November 2015. Site of service payment differentials are arguably hardest to justify for off-campus HOPDs, which often look and feel similar to a physician’s office. In 2020, the Congressional Budget Office (CBO) estimated that a similar policy would save $39 billion over ten years.
The second, and most impactful, proposal would reduce Medicare payments for certain services at HOPDs (on- and off-campus) to the ASC payment rate or the physician fee schedule rate. In general, the Secretary would determine annually whether each specific ambulatory service was most often performed in a physician’s office, an ASC, or a HOPD during the preceding four years. If a service was most frequently performed in a physician’s office, then Medicare would pay a rate intended to be equivalent to the physician fee schedule rate regardless of the site of care. If the service was most frequently performed in an ASC, then Medicare would pay the ASC rate for that service when it is delivered at a HOPD (and still pay the physician fee schedule rate when that service is delivered in a physician’s office). If the service was most frequently performed in a HOPD, then payment would not change. Once a service was assigned to a lower payment category, it could not be reassigned to higher payment category in future years.
This proposal mirrors an approach discussed by the Medicare Payment Advisory Commission (MedPAC) in their June 2022 Report to the Congress (and follows years of related MedPAC recommendations dating back to 2012). MedPAC estimates that this approach would have reduced Medicare spending by $6.6 billion and beneficiary cost-sharing by $1.7 billion, if in effect in 2019. And CBO previously estimated that a similar proposal would reduce Medicare spending by $102 billion over ten years.
While this proposal would be a major step forward, we have two suggestions to improve it:
- Avoid blunting the benefits of site-neutral payments. The Subcommittee’s draft limits the shift toward site-neutral payment for certain hospitals. Specifically, payment rate reductions for hospitals with an above-median share of low-income patients (measured using the methods used when determining Medicare disproportionate share hospital payments) would be limited such that the hospital’s total reduction in Medicare revenue under the policy was no more than 4.1%.[2] This policy appears to be modeled on a stop-loss policy that MedPAC discusses in their June 2022 report. MedPAC estimates that this limitation would apply to 23% of hospitals, reducing the policy’s savings and the overall Medicare hospital revenue reduction by about 10%.
This limitation has two significant downsides. First, for the affected hospitals, the limitation would preserve the perverse incentives to acquire physician practices and shift services into the HOPD setting that currently exists. Indeed, once a hospital generates a large enough portion of its revenue through affected HOPD services, it would face the same financial incentive to acquire more physician practices that it faces today. Second, the limitation would forfeit some of the policy’s savings for taxpayers and beneficiaries—likely disproportionately benefiting hospitals that have most aggressively exploited Medicare’s current site of service payment differentials. Both downsides will grow over time if hospitals continue to convert physician practices into HOPDs (as they likely will since Medicare payment policy is not the sole driver of this trend).
To the extent that lawmakers are concerned about certain hospitals’ overall financial position, there are better ways to address those concerns than partially unwinding the site-neutrality policy. For instance, lawmakers could target additional assistance to hospitals that lawmakers believe are particularly financially fragile. Or lawmakers could provide for across-the-board increases to Medicare rates under the physician fee schedule, outpatient prospective payment system, or inpatient prospective payment system. Increases to physician fee schedule rates would also directly limit the magnitude of payment reductions under a site-neutral payment policy, as most of the affected services would now be paid based on physician fee schedule rates. Alternatively, the alignment of payment rates across different settings of care could be phased-in or a temporary stop-loss could be applied to give hospitals time to adjust.
- Use a less conservative formula to identify services where site-neutral payments are appropriate. A strength of this proposal is its data-driven approach to selecting services that will be subject to site-neutral payments. However, the specific approach used in the draft bill will likely miss some services that could be safely performed in a physician’s office.
In particular, there has already been a large shift of services out of physician offices and into HOPDs as hospitals have steadily acquired physician practices and relabeled them as HOPDs over the last decade, driven in part by the current payment system. For example, MedPAC notes that the share of Medicare chemotherapy administration services delivered in HOPDs grew from 35.2 percent in 2012 to 50.9 percent in 2019. As a result, some services that can in fact be safely delivered in a physician’s office—and historically have been most often delivered in that setting—are now most often delivered in a HOPD.
To avoid this problem, services could be assigned the payment rate for a lower-cost setting as long as some minimum percentage of services was delivered in that setting. For instance, if a service was performed at least 30% of the time in a physician’s office, then Medicare would pay the physician fee schedule rate. If a service was performed less than 30% of the time in a physician’s office, but at least 30% of the time in an ASC, then Medicare payment for care at both HOPDs and ASCs would be tied to the ASC payment system. For other services, payment would not change.
For related reasons, the bill’s prohibition on reclassifying services into higher payment categories over time serves an important function. Otherwise, if hospitals continue to acquire physician practices and convert them into HOPDs, more and more services would get classified as being eligible for the HOPD rate. This, in turn, would further hasten hospital-physician integration.
Finally, we note that there are risks to the bill’s approach of allowing the Secretary to specify a list of services that will always receive the HOPD rate when delivered in the HOPD setting. It is unclear why it is necessary, and it could invite lobbying efforts that create inappropriate exceptions.
Additional site-neutral related policies
Two other proposals related to site-neutral payments were also considered at the hearing.
The first would align Medicare payments to all off-campus HOPDs for drug administration services with what would be paid under the physician fee schedule, ending the BBA’s grandfathering of off-campus HOPDs established before November 2015 solely for these services. The payment alignment would be phased in over four years. This policy is simply a narrower version of the first policy discussed in the previous section, which would eliminate the BBA’s grandfathering of higher payments to already-established off-campus HOPDs for all services (including drug administration services). As such, its benefits are smaller. However, this proposal would still be an incremental improvement to the Medicare program.
The second proposal would require each department of a hospital to bill Medicare under a distinct national provider identifier (NPI). If hospitals are then also required to use distinct NPIs for different departments (in particular, on- vs. off-campus HOPDs) in the commercial market, this proposal may help commercial payers better identify the site of care. In our discussions with stakeholders, some health insurance issuers indicated that some of their contracts with hospitals specify different payment rates for on- and off-campus HOPDs, but that they are not always able to determine which type of HOPD a claim originated from. Therefore, this proposal may produce some savings in the commercial market, although we anticipate that any such savings would be small, particularly once it becomes clear that insurers can differentiate on- and off-campus HOPDs and parties adjust their contracting strategies accordingly.
Health Care Transparency
Many of the proposals considered at the hearing require insurers, providers, or other entities to disclose additional information in hopes of making the health care system more transparent. We consider each of the specific proposals in greater detail below, but first briefly discuss the ways in which greater transparency could improve the health care system, as supporters of these proposals intend.
First, transparency efforts that allow patients, employers, or other purchases to better understand the cost of alternative options could help them comparison shop more effectively. That could directly reduce health care spending by allowing purchasers to find lower-priced sellers and putting competitive pressure on sellers to reduce their prices. There is some evidence that past transparency efforts like some of those discussed by the subcommittee have modestly reduced prices for some services. However, we generally agree with the conclusion of a recent Congressional Budget Office review that this evidence points toward relatively small percentage price reductions (e.g., on the order of tenths of a percentage point in the commercial market). Savings of this size could still amount to billions of dollars per year and thus be worth seizing, but policymakers should have realistic expectations.[3]
Second, greater transparency could enable research and analysis that gives public and private actors a better picture of the health care system as a whole, potentially allowing policymakers to identify opportunities for reform and private actors to identify strategies to reduce spending. We suspect that this may be the most important effect of transparency initiatives over the long run. We note, however, that many of these potential benefits are contingent on future action by policymakers. In this respect, transparency should be viewed as a waystation, not a final destination.
Ownership transparency
Perhaps the most potentially impactful transparency bill put forward by the subcommittee would generate publicly-available data on the parent company and ownership structure of health care providers. Specifically, the proposal would require certain providers—physician groups, hospitals, ASCs, and freestanding emergency departments—to report this information to the Department of Health and Human Services (HHS), along with any mergers, acquisitions, or changes in ownership over the prior year. HHS would then be required to make these data publicly available.
Data like these would facilitate several types of important analyses. For example, by making available information on each entity’s ultimate parent company, it would allow agencies and researchers to more easily and accurately track horizontal consolidation in physician and facility markets and trace its effects. Information on ownership structure would also provide an up-to-date snapshot of emerging trends in different forms of corporate ownership of health care providers, as well as allow for stronger analysis of the effects of private equity, hospital, and payer acquisitions of physician practices.
Some similar data are already being reported to the Centers for Medicare and Medicaid Services (CMS), but the bill would greatly improve what is readily available to researchers, especially with respect to physician groups. In particular, there currently is no comprehensive source of data that makes it possible to identify which physician groups share a common owner or determine the ownership structure of the entity with a controlling ownership interest in the group (e.g., private equity, insurer, or hospital).
In detail, physician groups are identified in Medicare claims data by their tax identification number (TIN). But in many cases, large parent companies own and operate many TINs. For instance, Envision, owned by private equity firm Kohlberg Kravis Roberts (KKR), operates more than 100 TINs in its emergency medicine staffing business. Moreover, matching TINs to their parent company and the ownership structure of that parent company is often complex and difficult to automate. The best starting point at present is the Medicare Data on Provider Practice and Specialty (MD-PPAS) data extract, which researchers can purchase from CMS. These data include the legal name associated with physician group TINs, from the Medicare Provider Enrollment, Chain, and Ownership System (PECOS) data, and some information on the geographic location where a clinician delivers most of his or her Medicare services.
But that information is often inadequate to identify a practice’s parent company. Researchers are therefore forced to identify parent companies by piecing together additional information on ownership from various sources, which make this type of research very resource-intensive and thus limits how much can be done. This manual work also inevitably ends up incomplete, which can result in researchers underestimating the level of market concentration, among other potential problems. Ideally, this proposed legislation would lead to CMS releasing a more robust version of the MD-PPAS data extract that additionally includes the parent company and ownership structure of each physician group TIN (or otherwise link TINs to the relevant parent company and its ownership structure).
Another bill considered by the subcommittee also aims to increase the transparency of health care ownership arrangements, with a focus on vertical integration between health plans and providers and in-home medical assessment entities. The bill requires Medicare Advantage Organizations (MAOs) to disclose certain data, including about average enrollee utilization, diagnoses, risk scores, and spending, broken out by whether enrollees’ had received services from a provider integrated with the MAO. (The bill would also require certain reporting by prescription drug plan sponsors, pharmacy benefit managers, and pharmacies, which we discuss in a later section.)
As discussed above, we believe that greater disclosure of information about ownership arrangements has high value. If those data can be linked with health care claims data, researchers can then answer a wide range of questions about the behavior of, in this case, vertically integrated entities. In the case that no such linkage is feasible, collecting certain aggregated data—as envisioned in this bill—can also be useful if the data offer insight into business practices and are not otherwise attainable. However, many of the comparisons that could be made using aggregate data requested in this bill would be quite vulnerable to confounding from unobserved differences in patient characteristics, which makes us question whether it is worth bearing the (modest) administrative burdens required to collect them.
As an alternative, we suggest pursuing the enhanced reporting on provider ownership discussed earlier in this section in tandem with improvements to the Medicare Advantage (MA) encounter data available to researchers. Most importantly, CMS should release payment and cost-sharing information it collects with the encounter data, which it currently withholds. It should also begin collecting information on plans’ non-fee-for-service payments to providers (e.g., shared savings or shared loss payment), which are not captured in traditional claims data. To facilitate analysis of how vertical integration affects the intensity of diagnosis coding, CMS could also give researchers access to enrollee-level risk scores for years other than 2014.
Price transparency for health care services
The committee also considered bills focused on increasing transparency around the prices of health care services. One would codify the insurer-facing Transparency in Coverage rules and the hospital-facing Hospital Price Transparency rules, while another would extend similar requirements to clinical laboratories. We discuss each bill in turn, as well as some of the limitations of expanding access to information on contracted rates without doing the same for information on utilization or enrollment.
Codifying the Transparency in Coverage rules
The Transparency in Coverage rules require commercial insurers to produce an internet-based tool that enrollees can use to obtain an advance estimate of their cost-sharing obligations, as well as publish machine-readable files that report the plan’s contracted rates for all health care items and services and certain information about out-of-network payments.[4] These rules, which rely on statutory authority provided by the Affordable Care Act, were finalized by the Trump administration in 2020 and began phasing in during 2022. Because the rules are already in effect, codifying them would have little immediate impact, although it would ensure that the rules could not be repealed by a future administration.
The bill does, however, include some provisions aimed at improving the usability of insurers’ machine-readable files, which are often challenging to analyze due to the files’ large size and the lack of a standard format. Notably, while the bill’s standards for the machine-readable files would largely mirror the standards established by the agencies in regulations and guidance, the bill would require that those standards ensure the files “are limited to an appropriate size.” Prodding the agencies to address file size challenges could be useful, considering the problems large files pose. Lawmakers could consider going further and requiring the agencies to establish a fully standardized reporting format, as well as develop a mechanism for auditing insurer compliance, as others have suggested; this would more closely mirror the bill’s approach with respect to the Hospital Price Transparency rules (discussed below).
The bill would also make a specific change to the current reporting regime aimed at reducing file sizes; it would allow insurers to omit a contracted rate if the provider submitted fewer than 10 claims for that service to the plan in the last year.[5] Unfortunately, this volume-based exclusion could substantially reduce the value of these data. While the volume accounted for by any specific excluded provider-plan-service combination would be quite small (by definition), the total volume accounted for by all such combinations might often be quite large. This is particularly true since volume would be measured at the plan level, and each self-insured group health plan would be considered a distinct plan for the purposes of this provision. More generally, sporadically excluding prices from this type of data set would often make analysis more difficult, offsetting ease-of-use benefits from reduced file size.
There are other ways to reduce file size that would avoid these downsides and are worth exploring:
- Requiring a more efficient file format: Insurers have generally published their files in JSON format, perhaps because this is the format for which CMS has published the most robust implementation resources. JSON has strengths; it is human-readable and well-suited to storing complex data, and it has a well-developed software ecosystem. However, JSON stores information very inefficiently, in part because field names are repeated in every record, a major liability in large databases like these. As an alternative, CMS could consider other file formats that are designed to store complex data, but can do so more efficiently, or it could consider adopting a relational format.
- Streamlining reporting of percent-of-Medicare contracts: Many provider-insurer contracts specify negotiated prices by starting with Medicare’s fee schedules (or some other publicly available fee schedule) and then applying a multiplier that is common across services. In such cases, it may be possible to greatly reduce the number of data elements that must be reported by allowing providers to report the single common multiplier rather than the resulting service-specific prices, which typically number in the hundreds or thousands.[6]
For this approach to achieve the desired goal, it would be essential that insurers clearly specify which Medicare prices (or other fee schedule) the multiplier applies to. An ideal approach would be for CMS to publish a suitable reference file of Medicare prices that all insurers could link to in their submissions. Alternatively, insurers could be required to include the relevant fee schedules as part of their submitted files; this latter approach could still generate some reduction in file sizes because the same fee schedule may often be used for multiple different contracts.
- Reducing duplication of prices within and across plans: There are also opportunities to reduce file size by reducing duplication, as others have also emphasized. Providers under common ownership often negotiate jointly with payers, which can result in all of those providers being paid identical prices. The current reporting format frequently requires insurers to report prices separately for each constituent provider, which can result in considerable duplication.
Duplication can also arise since reporting occurs at the plan level, but many plans offered (or administered) by a single insurer will often share the same network. At least some insurers have avoided duplication in this scenario by having the files for each plan that uses a given network link to the same underlying file of in-network prices; this approach could be required more broadly. This approach does not reduce the file size for any particular plan but does reduce the total amount of data that must be analyzed to generate estimates of prices at the market level.
To be clear, we think that it would be unwise to specify these or similar changes in legislation, as the appropriate reporting format will likely evolve as technology and the health care system evolve—and as CMS and the industry learn from experience. Nor do we think that these are the only potential strategies for reducing file sizes. But we do believe that there is room to make considerable progress in reducing the size of insurers’ machine-readable files without eliminating useful data from the files.
At the same time, policymakers should recognize that the machine-readable files will be large even with an optimal reporting approach. Given the number of separately billing providers and distinct health care services, many insurers’ files will unavoidably contain tens or hundreds of millions of distinct prices. Storing those prices and their associated metadata will often require file sizes measured in gigabytes.
Codifying the Hospital Price Transparency rules
The Hospital Price Transparency rules require hospitals to publish a machine-readable file that includes their gross charges, cash prices, and prices negotiated with third-party payers for all items and services, as well as display prices for a subset of those services in a “consumer-friendly” way. These rules, which also rely on statutory authority provided by the Affordable Care Act, were finalized by the Trump administration in 2019 and took effect at the beginning of 2021.
The bill would largely simply codify the existing Hospital Price Transparency rules, but it would also implement some additional standards aimed at making the data more useable and complete. Notably, the bill would require hospitals to begin posting information in a standardized format and require that HHS establish a process for auditing the accuracy of the posted information, neither of which exist at present. In light of the observed challenges with compliance and data quality, steps like these make sense.
A longer-term question is how to coordinate the hospital-facing price transparency requirements in the Hospital Price Transparency rules with the insurer-facing requirements in the Transparency in Coverage rules. While having multiple sources of hospital price data is useful at present given the limitations of both data sources, this type of duplicate reporting does impose additional compliance costs. Once the insurer-facing requirements are functioning better, we think it may be worth paring back the hospital-facing requirements so that they encompass only information that cannot be obtained from insurers; this would include hospitals’ gross charges and cash prices, as well as negotiated prices that are outside the scope of the insurer-facing requirements (e.g., Medicare Advantage and Medicaid managed care rates).
Clinical laboratory price transparency
Another bill would extend price transparency obligations to clinical laboratories. Laboratories would be required to publicly post their cash prices, as well as their minimum and maximum negotiated prices with commercial insurers, for all services. In our view, increasing transparency around cash prices is potentially valuable, but, as with the Hospital Transparency Rules, lawmakers may wish to consider whether requiring laboratories to disclose negotiated prices is necessary in light of the insurer-facing rules.
Building on price transparency
These proposals to increase price transparency have the potential to help reduce health care spending, but there is room to go further. One step to consider is extending the Transparency in Coverage rules, which currently apply only to individual and group market plans, to Medicare Advantage and Medicaid managed care plans. Such plans now enroll around one-quarter of the U.S. population.
Additionally, it is important to recognize that prices are only one part of the health care spending equation: utilization matters too. Indeed, even many analyses focused narrowly on prices require knowing how many people receive services from a particular provider under a particular plan. Without this information, it is impossible to calculate average prices or other summary measures of interest.
Thus, in parallel with these price transparency efforts, it would be valuable to take steps to improve access to utilization data as well. One such approach would be to take steps to expand availability of All-Payer Claims Databases (ACPDs), which collect claims data from payers operating in a particular market. APCDs can be used to create a comprehensive picture of health care delivery in a particular market.
An ideal approach would be for the federal government to step in and create a national APCD; an effort like this may be possible even under existing law, but Congressional action to direct and fund this type of effort would make a successful effort much more likely. Failing that, policymakers could take steps to make state APCD data more useful, including by restoring state APCDs’ ability to collect data from self-insured plans (which was removed by a 2016 Supreme Court decision) or supporting efforts to harmonize data collected by different state APCDs and make it easier to combine data from multiple states.
Transparency in prescription drug markets
The hearing also considered proposals aimed at increasing transparency in prescription drug markets, particularly related to pharmacy benefit managers (PBMs) and the 340B Program. We discuss each below.
Medicare PBM transparency
One bill considered by the subcommittee would seek to clarify relationships among prescription drug plan sponsors, PBMs, and pharmacies in Medicare Part D, and particularly how those relationships depend on whether the plan sponsor owns the pharmacy. (Other parts of the bill would examine Medicare Advantage and Medicare Part D plans’ relationships with providers and were discussed above.)
In detail, the bill would require the following reporting to CMS:
- Plan sponsors: Plan sponsors would be required to report on the prices paid to each in-network pharmacy for each drug. They would also report the per drug average of any payments received from pharmacies, broken down by whether the pharmacy is owned by the plan.
- PBMs: Entities providing PBM services would be required to report aggregate information for each plan they managed, including: the amount of the rebates the PBM received from manufacturers and the portion it shared with plan sponsors; the amount the PBM paid to pharmacies; the amount the plan sponsor paid to the PBM to cover claims and, separately, administrative costs; the PBM’s own administrative costs; and whether the PBM was owned by the plan in question.
- Plan-owned pharmacies: Plan-owned pharmacies would be required to report on the prices they are paid by each plan they contract with, as well as the per drug amount of any payments made to the plan. Non-plan-owned pharmacies would be permitted to report if they wished.
We believe that this information, perhaps especially the information reported by PBMs, may be useful in helping observers understand the thicket of financial relationships that exist in prescription drug markets. As such, an important strength of the bill is that it directs CMS to make this information publicly available.
However, the bill restricts that public disclosure in ways that may make it less useful than it could be. In particular, the bill requires the disclosure be structured so that it is not possible to glean drug-specific prices or to identify specific drug manufacturers, pharmacies, or PBMs. As a practical matter, that may preclude public disclosure of anything beyond aggregates and averages derived from these data, which would greatly reduce their utility. Thus, lawmakers may wish to consider allowing public disclosure of at least sponsor- and PBM-level aggregates, for which we do not see obvious risks from disclosure. At a minimum, it would be useful to clarify that researchers should be allowed to analyze the underlying entity-level reports through a process similar to the one that researchers use to access other identifiable CMS data sets as long as any published results of the research were suitably aggregated.
Commercial PBM transparency
Another bill considered by the subcommittee aims to increase the transparency of PBM services provided to group health plans by requiring PBMs to disclose certain information to plan sponsors.[7] Under the bill, PBMs would be required to report utilization, gross spending, and out-of-pocket spending at the drug level. For drug classes with at least three drugs on the formulary, they would also be required to report the plan’s net spending as well as rebates (and other price concessions) paid by manufacturers that are related to utilization under the plan. PBMs would also be required to disclose certain aggregate information, including the plan’s total gross and net spending on prescription drugs, total out-of-pocket spending by plan beneficiaries, and the total amount of rebates (and other price concessions) paid by manufacturers that are related to utilization under the plan. These reports would be required to be submitted to the Government Accountability Office (GAO), which would be tasked with studying certain features of PBM contracts, including how they treat pharmacies under common ownership.
Giving plan sponsors additional information could help them better assess the terms offered by PBMs, which could allow sponsors to select better contracts or spur greater competition among PBMs, as CBO has concluded when assessing similar prior legislation. In this respect, utilization information might be particularly useful, as it might permit self-insured employers to better forecast what their benefit costs would be under alternative PBM contracts. In principle, information on the amount of rebates being retained by the PBM could also help plan sponsors assess how much room there is to press for a better deal (whether in the form of a lower premium in a fully-insured plan or lower administrative fees under a self-insured plan), but plan sponsors would be hampered by a lack of information on PBMs’ costs and target profit margins, so we question how useful this information would be in practice. (Requiring PBMs to report information on their costs, similar to the requirements under the Medicare transparency proposal discussed above could partially fill these information gaps, but not fully.)
A weakness of this proposal is that the information disclosed would only be available to plan sponsors (and GAO). As noted above, we think that much of the value from greater transparency may be in allowing policymakers, researchers, and other private sector actors to gain a better picture of how the health care system operates. Creating some mechanism through which these reports—or summary information compiled from the reports—would be shared with a broader set of users could be valuable.
340B
The 340B program requires drug manufacturers to give large discounts on prescription drugs to hospitals and related entities that treat a relatively high share of low-income patients. Entities can sell those drugs at higher rates, keeping the difference as profits. Over time, the program has grown substantially. Between 2000 and 2020, the number of covered entities rose from 8,100 to 50,000. In 2021, these entities purchased $43.9 billion in covered drugs. Despite the large role the program now plays, there is relatively little oversight into the program’s operation and how these revenues are used.
One draft bill considered at the hearing seeks to increase transparency of this program by requiring covered entities to report certain information, including: the number of individuals who were dispensed or administered covered drugs (by payer type); the amount of charity care delivered; their profits or losses on care delivered to people with public insurance or who are uninsured; the net revenues associated with covered drugs, and how those revenues were used.
Some of this information, particularly the information about the number of patients receiving covered drugs and entities’ net revenue on these drugs, are not available from other sources and could help observers better understand the program’s operations. However, some of the new reporting requirements would mirror existing reporting requirements. In particular, information on charity care and uncompensated or undercompensated care is already reported to the federal government on Medicare costs reports. We recommend against compelling duplicate reporting of these data here. Similarly, it is unclear what can be learned from asking providers to report how they use 340B funds. Because money is fungible, providers will have broad flexibility in reporting what the 340B-derived funds notionally paid for. Thus, their answers are likely to be of little use in answering the question of ultimate interest, which is what spending would not have occurred in the absence of the 340B funds.
Other Proposals Considered at the Hearing
The subcommittee also considered several other proposals that are not directly related to site-neutral payment or transparency. We briefly discuss several of these in turn below.
Limiting cost-sharing of highly rebated drugs
Brand drugs’ “list prices” often significantly exceed the post-rebate “net prices” paid by insurers. Patients’ cost-sharing under their insurance plan is often a function of the list price, rather than the net price, which can sometimes result in the patient’s cost-sharing obligation exceeding the net price of the drug. A subcommittee discussion draft aims to provide some relief to patients taking these highly rebated drugs.
Specifically, the policy would limit an enrollee’s maximum monthly cost-sharing for highly rebated brand drugs, defined as products for which rebates and other discounts exceed 50% of the list price (in other words, drugs where the net price is less than half of the list price). Under the proposal, monthly cost sharing could not exceed 1/12th of the annual net price paid by the insurer for the drug in the prior year.
This proposal has an intuitive rationale. The fundamental purpose of cost-sharing is to deter overuse by ensuring that patients are attentive to the costs of their treatment. But that logic cannot justify charging patients more than the cost of their treatment, which would generally be expected to cause underuse.
Additionally, this proposal may lower out-of-pocket spending early in the year for patients with deductibles who take highly rebated drugs and result in more consistent monthly out-of-pocket spending. There are merits to smoothing out-of-pocket spending in this way. If this is the committee’s goal, however, it is worth considering whether changes specific to highly rebated drugs are the right response, as opposed to changing the nature of annual deductibles more generally (e.g., moving towards monthly deductibles).
Regardless, policymakers should recognize that this policy presents tradeoffs. First, there is no free lunch. To the extent that these types of policies do lower annual out-of-pocket spending for some enrollees, the higher obligations facing insurers will likely be reflected in higher cost-sharing for other services or higher premiums. Whether this tradeoff is worth making is an open question, but it is worth acknowledging.
Second, implementing this type of requirement will impose new administrative burdens on both plans and regulators—costs that will ultimately be borne by enrollees and taxpayers—and make the health care system more complex. Whether that additional cost and complexity is worth bearing depends on the magnitude of the harm done by the status quo, and we harbor doubts that harm is especially large.
We also note that policymakers may be aiming to disrupt the incentives driving gaps between list and net prices more generally. However, it is not clear how much this change would advance this goal—both because the policy only applies to a subset of drugs and because the incentives that manufacturers face when setting list prices or when negotiating net prices with purchasers would change at most marginally, even for affected drugs.
Ban on Medicaid spread pricing
PBMs often charge the entity responsible for paying enrollee claims (e.g., an insurer, an employer, a state Medicaid program) more than it pays the pharmacy for a drug, a practice known as “spread pricing.” One bill considered by the subcommittee would ban spread pricing in Medicaid.[8]
The goal of this provision appears to be to reduce what Medicaid spends on prescription drugs by reducing PBM profits, thereby saving money for state and federal governments. We caution, however, that savings from this type of provision may be smaller than the prevalence of spread pricing would suggest. If PBMs are barred from retaining the “spread,” they are likely to respond by seeking higher upfront administration fees from states to compensate for the lost revenue. Since limiting spread pricing would not change anything fundamental about the parties’ bargaining positions—including the PBM’s cost of providing its services and the value the state places on those services—we believe they would likely be largely successful in doing so.[9] That could be less true in the short run, when contract terms may be somewhat “sticky,” but we expect that any savings would fade over time. This is consistent with a recent CBO score that found some savings in the first several years that eroded rapidly over time.
Lawmakers should also note that it is possible that this type of provision could backfire. PBMs might be less motivated to negotiate low prices with pharmacies if they are not permitted to retain the spread, in which case states might both face higher administrative fees and see smaller-than-expected direct benefits from eliminating the spread. Implementing and enforcing this provision would also have some administrative costs for both states and the federal government that should be accounted for as well.
CMS analysis of effects on consolidation
The design of public health care programs can have a large effect on consolidation incentives in health care markets, as the example of ambulatory site-of-service differentials vividly illustrates. Consolidation, in turn, can result in less competition and higher costs in the commercial health care market and can reduce quality in public programs. The subcommittee considered a discussion draft that would require all rulemakings that change Medicare payment policy to seek public comment on and include projections of the effect of such changes on provider or payer consolidation. It also requires the Centers for Medicare and Medicaid Innovation’s evaluations of its payment models to assess consolidation effects.
The merits of this policy hinge on whether the increased administrative burdens on the agency are likely to be outweighed by improvements in the agency’s policy decisions. We see little harm in soliciting comments on potential consolidation effects because it imposes minimal additional costs on the agency and could conceivably highlight underappreciated incentives, although we also doubt that will often be the case in practice. Indeed, we doubt that this would represent a meaningful change from the status quo since the agency already seeks (and receives) broad comment on payment policy changes.
However, requiring the agency to devote resources to producing its own assessment of consolidation incentives in annual rulemaking or when evaluating Innovation Center models presents more substantive tradeoffs. Those resources would not be available to advance other agency objectives, and CMS is unlikely to have a comparative advantage in assessing the competitive effects of polices relative to other agencies (e.g., antitrust agencies) or outside experts who can respond to comment solicitations.
Medicare inpatient only list
The inpatient only list (IOL) specifies services that Medicare will only pay for if provided in the inpatient setting. The goal of the IOL is to ensure that beneficiaries receive care in a clinically appropriate setting. But it also raises costs. Most directly, Medicare generally pays more for services delivered in the inpatient setting. In addition, fewer providers are likely to offer these services in outpatient settings if they will not attract any Medicare beneficiaries, which can reduce competition in the commercial market. CMS made a decision to phase out the IOL in late 2020, but then reversed this decision the next year.
The subcommittee considered a discussion draft that would reduce the number of services included on the IOL. Specifically, CMS would be required to remove all musculoskeletal services from the IOL. The agency would also need to produce a report that analyzes Medicare spending, beneficiary out-of-pocket spending, and quality metrics for services included on the IOL. In conducting this report, the agency would be instructed to examine data on non-Medicare patients who receive such care in an outpatient setting. Starting in 2027, CMS would be barred from placing a service on the IOL unless the study provided “conclusive clinical evidence” that the service may not be performed safely in an outpatient setting.
A threshold question is how much this policy would change CMS’ approach to the IOL. On its face, requiring “conclusive clinical evidence” to justify a service’s inclusion on the IOL would seem to set a very high standard that would result in few services being included on the IOL. However, CMS would likely have considerable discretion to determine what constituted “conclusive clinical evidence.” Moreover, CMS currently removes services based on factors like whether HOPDs or ASCs are equipped to provide the services or whether the service is already furnished on an outpatient basis in the commercial market. In practice, this approach may already exclude many services for which suitable evidence does not exist. On balance, however, we suspect that this proposal would shrink the IOL to some degree.
We are sympathetic to the view that CMS’ incentives make it overly cautious about removing items from the IOL; adverse clinical outcomes may be highly salient, even if quite rare, while potential cost-savings may be less so. Additionally, it is not obvious that providers face financial or other incentives that would lead them to inappropriately deliver care in an outpatient setting. (A caveat is that this could depend on how CMS applied the “two midnight” rules to inpatient stays involving procedures removed from the IOL. In general, those rules say that an inpatient stay is only eligible for payment if it is expected to cross at least two midnights, but do not apply to stays involving services on the IOL.) Thus, we are cautiously supportive of encouraging CMS to be more selective in placing services on the IOL.
However, there are reasonable arguments that requiring “conclusive clinical evidence” to place an item on the IOL is an overly stringent standard. If policymakers agree, they could consider setting a weaker evidentiary standard like “the preponderance of the clinical evidence.” Alternatively, they could consider allowing CMS to retain a service on the IOL if commercial patients are not commonly receiving the service in the outpatient setting, even in cases where “conclusive clinical evidence” is not available.
[1] Even if patients treated in ASCs and HOPDs were systematically more costly, this would justify paying more for the complex cases, not paying more for all services delivered in ASCs and HOPDs.
[2] We read the legislative language as actually specifying a limit of no more than 4.1% of Part B hospital payments, rather than 4.1% of all Medicare hospital payments. But we assume that the drafters’ intention was to mirror the policy analyzed by MedPAC and, thus, that the language is intended to refer to all Medicare hospital payments.
[3] On the other hand, this evidence does at least partially alleviate concerns that greater transparency could increase prices, for example by making providers more reluctant to give certain insurers discounts that would then be demanded by other insurers.
[4] We use “commercial insurers” as a convenient, if slightly inaccurate, shorthand to the entities subject to these rules. Technically, they apply to both individual and group market insurers and group health plans; however, an insurer or third-party administrator is typically responsible for reporting on behalf of a group health plan.
[5] Because the revisions are largely drafted to supplement rather than substitute for the authorities the agencies used to promulgate the Transparency in Coverage requirements originally, it is likely that the agencies would retain the authority to require reporting of low volume codes if they wished.
[6] A similar approach is currently permitted for chargemaster contracts, but not Medicare-based contracts.
[7] Here, we use are using the term “PBM” to encompass any entity administering prescription drug coverage on behalf of a group health plan, which may also include insurers and third-party administrators.
[8] Technically, the bill applies to both PBMs and managed care plans that oversee outpatient drug coverage.
[9] The legislation limits PBMs to “reasonable” administrative fees, but it would likely be difficult to judge whether any particular fee is, in fact, “reasonable.”
Acknowledgements: We thank Richard Frank for helpful comments on a draft of this piece, and Conrad Milhaupt and Caitlin Rowley for excellent editorial assistance
Disclosures: The Brookings Institution is financed through the support of a diverse array of foundations, corporations, governments, individuals, as well as an endowment. A list of donors can be found in our annual reports published online here. The findings, interpretations, and conclusions in this report are solely those of its author(s) and are not influenced by any donation.