Over the last ten years, policymakers of both parties have identified health care payment reform as a critical national priority. Payment incentives offered by Medicare and other large payers create the environment in which providers must choose between being rewarded for performing more services and procedures, as in fee-for-service medicine or, for efficiently managing the cost and quality of care received by their patients in risk-sharing alternative payment models.
By reforming payment systems, we can change the way we deliver care to unlock a world in which health care can be both better and more affordable. But this will not happen on its own. The policy goals of payment reform can only be achieved by maximizing two objectives that are in tension: participation in risk-based alternative payment models and performance in those models. These goals are in tension because Medicare incentivizes participation in those models through financial rewards based on program performance. For example, modifications to the way that the Centers for Medicare and Medicaid Services (CMS) sets financial benchmarks for accountable care organizations (ACOs) can make the program more or less attractive to new program entrants. At the same time, those policy changes also affect the incentives of the ACOs within the program to perform well and reduce spending, which serves to benefit Medicare’s long run fiscal health.
Other incentives available to encourage participation—such as expanding the base payment differential between payment models—have been used with success since the passage of The Medicare Access and CHIP Reauthorization Act (MACRA) in 2015. But those incentives are eroding, as we will describe, and Congress must increase the incentive to join higher-risk models if we are to realize the benefits of widespread adoption of new payment models.
This post focuses on participation in alternative payment models as a beneficial outcome and consensus policy objective of the last decade. Some have made arguments on behalf of fee-for-service buttressed by pay-for-performance models like the Merit-Based Incentive Payment System in Medicare. We believe pay-for-performance has been largely ineffective and does not show sufficient promise for the type of practice transformation possible under Advanced Alternative Payment Models (A-APMs), as discussed here and in many other papers.
Growth Of Alternative Payment Models
We’ve spent the past 10 years going from theory to proof of concept. Starting from an almost purely fee-for-service universe in 2010, adoption of alternative payment models increased dramatically by 2020, catalyzed by policies enacted within the Affordable Care Act, decisions by major state and commercial purchasers, and the support of the medical professions. The Medicare Shared Savings Program (MSSP), in particular, is now a mature program serving more than 10 million Medicare beneficiaries. It has been supplemented by frequent updates and quick evolution informed by ACO models tested at the Center for Medicare & Medicaid Innovation (CMMI), including the Pioneer ACO model and the Next Generation ACO model. In addition, hundreds of thousands of providers have participated in other CMMI models, including bundled payments and primary care medical home models.
According to the Health Care Payment Reform Learning and Action Network, in 2018, 35.8 percent of payments to health care providers, covering about 77 percent of the insured population, flowed through payment models that included some accountability for cost and quality of care. However, most of those were in payment models that are “upside only,” that is, do not include loss sharing yet; only 14.5 percent of payments flowed through models featuring downside risk in 2018, up from 12.5 percent in 2017. In addition, in 2020 and 2021, growth in participation in MSSP ACOs slowed and then declined, partly reflecting recent changes CMS made that reduced the attractiveness of the program to some providers relative to staying in fee-for-service.
Incentives For Providers To Join Alternative Payment Models
Broadly speaking, physicians serving patients in original Medicare have two options for Medicare payment platforms over the next ten years. The default option is to remain in fee-for-service, face low-to-zero updates indefinitely, and be subject to adjustments under the Merit-based Incentive Payment System, roughly related to clinical quality and information exchange.
The second option is to join an Advanced Alternative Payment Model (A-APM) such as a risk-bearing ACO. Congress has defined A-APMs as those models with greater than “nominal” financial risk. This option has low—but somewhat higher than fee-for-service—fee schedule adjustments; a 5 percent bonus available just through 2022; and the opportunity to earn shared savings. This option also, however, could require the provider to pay Medicare back for certain higher-than-expected costs for patient care, also called shared losses.
Policymakers from both parties as well as most payers and patient groups prefer the second option, that providers should bear some responsibility for the cost and quality for the care delivered to patients they manage. But the economic underpinnings encouraging providers to join A-APMs are eroding because of sunsetting provisions and CMS policy decisions that disincentivized participation in A-APMs.
Exhibit 1: Summary of key payment differences for providers in A-APMs vs those remaining in fee-for-service.
Source: authors’ analysis. **The theoretical maximum bonus payment in any year is 27 percent but CMS projects that the maximum bonus paid in 2023 will be 5.3%; the structure of the MIPS incentive makes it unlikely that the maximum bonus will ever exceed 10%.
Payment Incentives Drive The Decision To Stay In MIPS Or To Join Advanced APMs
Passed in 2015, MACRA created the two payment options described above. Couched in terms of choices for providers, Congress intended to create incentives for providers to adopt A-APMs that also allowed for providers not in A-APMs to be scored according to performance on cost, quality, usage of electronic medical records, and structural improvements. We briefly summarize the components of payment for practices staying in MIPS, the incentives to join A-APMs, and the conflict between the payment options arising in the coming years.
Physician Fee Schedule
Medicare physician fee schedule updates will occur differentially for providers in A-APMs and all other providers. Congress replaced the Sustainable Growth Rate mechanism that predated MACRA and instead provided very low annual updates to the fee schedule for the following decade. Over the years 2016 -2019 the updates were 0.5 percent for all providers; over the years 2019-2025 the fee schedule is frozen with 0 percent updates; and for 2026 and beyond, Congress provided for a 0.25 percent update for providers not in A-APMs and a 0.75 percent update for providers in A-APMs. This payment differential continues indefinitely, providing a 5 percent payment wedge between A-APMs and non-participants by 2035, for example.
MIPS went into effect in 2017, with a slow phase-in of “teeth” planned by both Congress and CMS. MIPS is structured as a predominantly budget-neutral payment adjustment to the physician fee schedule, with incentive payments limited to penalties collected (plus an additional $500 million available through 2024 for exceptional performance on MIPS). Both payments and bonuses are judged on the same linear scale, scored from 0 to 100, with thresholds set each year for minimum performance and exceptional performance. The MIPS score is the sum of scores in each of the four (or fewer) domains: clinical quality, electronic medical record-based information exchange, cost, and practice improvement. In 2020, the performance threshold was 45 points; the maximum penalty applies to those scoring at one-quarter of that amount (11.25); practices with 45 points or more avoid penalties entirely; scores between 45 and the exceptional performance threshold of 85 points receive a positive adjustment, and scores above 85 receive the exceptional performance adjustment on a linear scale.
CMS projects that for the 2021 performance year (2023 payment year), 7 percent of those submitting data (60,463) and all of those not submitting data (27,115) will receive a penalty. Those not submitting data will receive the maximum 9 percent penalty.
In all, 87,578 may receive penalties in total out of 891,000 eligible clinicians, or about 9.8 percent; the rest will likely receive positive adjustment between 0 and 5.3 percent.
The most significant and the most variable incentive to join A-APMs is usually the payment arrangement of the model itself. In Shared Savings Program ACOs, for example, new entrants can share 40-75 percent of savings (i.e., reduced total costs) against a benchmark. The sharing rate increases alongside the amount of risk the participant accepts: 40 percent for the non-risk early years of the program, 50 percent for lower risk-bearing years (with losses capped at the lower of 4 percent of a benchmark or 8 percent of fee-for-service revenue), 75 percent for the “Enhanced” track with losses capped at 15 percent of expenditures for the ACO’s attributed beneficiaries. These incentives, among the most important determinants of success within an ACO as well as the incentive to join it, have been modified frequently in recent years. Rulemaking in 2018 that introduced the Pathways to Success framework and its faster path to risk may have seriously and inadvertently slashed the incentive to join ACOs for many high cost providers even as it increases the apparent savings of ACOs participating in the program.
In practice, providers on the margin of participation analyze the combination of fee schedule adjustments against potential gains and losses in the A-APMs to determine whether to participate. The fear of downside risk can dominate the attention of smaller providers, even in the face of significant upside from shared savings. For example, an independent practice may fixate on the 8 percent of their practice revenue that could be lost in a risk-bearing track of the MSSP. A 10,000-life ACO with an $11,000 benchmark that doesn’t qualify for the 8 percent revenue protection could face a loss of $4.4 million in the worst case. The sticker shock of these possibilities and the unknown nature of the risk scares many providers away from joining A-APMs, and limited push from the fee schedule exacerbates the problem.
Providers also believe that avoiding significant MIPS penalties is far easier and more reliable than successfully achieving performance in a risk-bearing ACO with more stringent cost/quality tests, meaning that the cost required to remain in FFS/MIPS is also lower.
Impediments to Advancing Delivery System Reform
The murky incentives to join A-APMs are beginning to manifest in declining participation in Medicare ACOs and slower than desired growth in risk-bearing ACOs. Given the way that the financial incentives to join A-APMs have been dampened, fewer providers will transition to adopt them. Many participants in Medicare ACOs this year must choose whether to transition to the new Pathways tracks that require adoption of downside risk in year 3; for many providers, the non-risk path in MIPS will be more attractive. Progress on delivery system reform is threatened by the overlapping problems discussed in this post: the expiring A-APM bonus; the possibility that MIPS total payments could exceed A-APM payments; the small wedge between the A-APM and FFS updates beginning in 2026; the assortment of quality measurement options for MIPS providers compared to ACO quality measurement; and the coverage of only Medicare Part B spending in MIPS compared to both Parts A + B in the ACO frameworks.
The Next MACRA Legislation
Legislation will be required to address these issues because many of these problems arise from the MACRA statute. Responding to COVID-19 has dominated the agenda in 2021 in Congress and throughout the Biden administration, but the urgency of improving the MACRA incentives is significant, especially given the stress on the Medicare Trust Funds.
While challenges persist, the good news is that the changes that would have the most impactful are relatively straightforward. Payment reform legislation should aim to reward participants in higher-risk models with substantially higher rewards, and we should make it attractive to “level up” and join these models. At the same time, Congress and CMS should work to ensure that these higher-risk payment models deliver more savings and quality improvements over time. Simplifying the payment incentives so that providers understand their path to the greatest payment increases is also an important objective that supports both more participation and more success in risk-bearing models.
Specifically, Congress should:
- Extend and ideally increase the 5 percent A-APM bonus. This can be budget neutral if it is paired with reductions in fee increases for providers outside A-APMs or if the bonus is included in the benchmark of A-APMs.
- Make MIPS less attractive than A-APMs by reducing the number of measures in MIPS (as CMS has already begun attempting through the “MIPS Value Pathway”) and by capping the MIPS bonus at 1-2 percent to eliminate disincentives to join A-APMs, making clear that base payments will be larger in risk-bearing models than MIPS.
- Work with CMS to prioritize which models deserve priority for A-APM incentives and clarify how providers qualify for those incentives. This includes addressing technical issues such as determining how to handle specialist participation in A-APMs and what threshold of a practice’s business should be in an A-APM to “count” for any rewards; it also includes ensuring benchmarks don’t discourage participation of rural providers, or higher-cost providers.
Congress has an opportunity to harness the significant energy behind payment reform and channel it into clear and strong payment incentives for providers to improve the quality of care they deliver and manage total costs of care. Inaction in the coming years could lead to a reversal of the progress made over the past decade.
The authors thank Louise Yinug for editing assistance. The authors are employed at organizations that provide services to entities described in the blog post. Tim Gronniger and Joe Lucas are employed at Caravan Health. Rahul Rajkumar is an employee at Optum Care Solutions and Duke University Health System and has a relevant equity stake in PicassoMD, Advantia Health. Scott Heiser is an employee of Cityblock Health. Misha Segal is an employee of Berkeley Research Group.