Access the audio version of this article
Over the past few decades, value-based payment (VBP) has been touted as the magic cure-all for the U.S. healthcare system. With the highest spending per capita for healthcare of any developed country, yet ranking well below average for key health indicators like life expectancy, avoidable mortality, chronic disease morbidity, and infant mortality, the U.S. healthcare system unquestionably needs reform. However, after over a decade of experimentation with VBP models, a question has arisen: Can VBP achieve the goals of reducing the cost of healthcare while improving health outcomes, or will VBP instead increase health disparities and harm vulnerable populations?
The underlying theory of VBP is to transform the traditional fee-for-service (FFS) model by realigning payment incentives. Under the FFS model, providers bill for each health service performed without any regard to health outcomes. Because of the lack of accountability for outcomes, providers are incentivized to overuse unnecessary or overly expensive services, and underuse less costly services like preventative care. VBP models seek to realign these incentives by paying for healthcare based on quality and outcomes, rather than per service provided.
Over the last decade, experimentation with VBP models has generally utilized one or more of the following methods: bundled payments, capitation/global payment, pay-for-performance, and shared savings. While each method has its own nuances, the general idea is that if providers are able to (1) keep costs below set benchmarks and/or (2) health outcomes at or above set targets, then they will either receive a bonus or share in the savings they help create for government programs like Medicare and Medicaid (upside risk). However, in more advanced payment models, providers will also receive financial penalties or share in the losses to the government programs if their costs exceed the benchmarks and/or the outcomes do not meet their targets (downside risk).
Persuading providers to participate in VBP models is one of the largest barriers to system-wide transformation. The cost of installing the new infrastructure and technology needed to track service quality, health outcomes, and cost benchmarks is a large deterrent to participation. In addition, convincing providers to participate in models with downside risk, which is the end goal in order for VBP to be successful, is a tough sell. As a result, VBP models have had to provide generous incentives, bonuses, and upside-risk-only models in order to initially draw in providers. However, these generous financial incentives that have been necessary to attract and maintain participation have prevented the majority of VBP models from producing net savings for government programs.
The strategy of using upside-risk-only models as an on-ramp to draw in providers and eventually move them into more advanced models with downside risk has additional drawbacks in models with voluntary participation. Providers are able to game the system by entering a VBP model with downside risk when they believe the financial risk is in their favor, and then exit the model when it is no longer in their financial best interest. This generally leaves government programs paying generous shared savings bonuses to providers and then being left with all of the losses when providers perform poorly in terms of cost savings and outcomes.
The solution to these issues may seem obvious: mandatory participation in order for providers to receive government payments. However, VBP models are still experimental. Lack of concrete data on which methods and models meet the overarching VBP goals, lack of knowledge on best practices and strategies for providers, the cost of infrastructure development in order to track cost and quality metrics, and evidence that VBP models may not advance health equity make mandatory participation an end goal that is still out of reach. In particular, after over a decade of extensive experimentation, one lesson learned is that current VBP methods may decrease health equity.
Health equity exists when all people have “the opportunity to ‘attain their full health potential’ and no one is ‘disadvantaged from achieving this potential because of their social position or other socially determined circumstance.’” In order to achieve health equity, social risk factors for poor health outcomes must be accounted for and eventually eliminated.
Social risk factors, also known as social determinants of health, are “the conditions in the environments where people are born, live, learn, work, play, worship, and age that affect a wide range of health, functioning, and quality-of-life outcomes and risks.” Social risk factors include, but are not limited to, racism, sexism, homophobia, ableism, xenophobia/nationalism, education disparities, income inequality, housing insecurity, transportation access, health systems and services access, social isolation, food insecurity, unemployment, and public safety concerns. In the United States, health disparities exist as a result of these factors in areas such as overall life expectancy, infant and maternal mortality, heart disease, diabetes, hypertension, chronic illness, disability, cancer, mental illness, and substance use.
Because social risk factors are key determinants of health outcomes, VBP reform needs to take them into account in order to achieve the goals of reducing costs and improving health outcomes. However, VBP models have done a poor job at accounting for them so far. The vast majority of VBP models were not designed with either health equity or social risk factors in mind, and social risk information has not been routinely collected. Even when social risk information has been collected, the lack of standardization in terminology and data location has made it difficult to fully assess the effects of models on many vulnerable populations.
The Center for Medicare and Medicaid Innovation (CMMI), the federal agency tasked with creating and testing VBP models, has analyzed the limited data that is available on the effect of social risk factors on Medicare VBP models. CMMI’s two findings show that VBP programs may actually worsen health equity instead of improve it: (1) beneficiaries with social risk factors had worse outcomes on many quality measures and (2) providers who disproportionately served beneficiaries with social risk factors performed worse and received higher penalties through downside risk models. There is evidence that these results are not simply another reflection of the effect social risk factors can have on health outcomes under any type of payment model, but rather evidence that VBP models may actually be regressive and widen health inequality even further.
In what has been termed a “Reverse Robin Hood” effect, VBP models with mandatory participation have further disadvantaged providers who care for a disproportionately high number of patients with social risk factors. Safety-net systems and systems that care for a higher proportion of racial and ethnic minority patients have fared worse and been disproportionately penalized through downside risk adjustments in mandatory VBP models. Providers in these same models who do not care for high-risk-factor populations, on the other hand, have fared well and shared in the upside-risk savings they generated. As a result, these VBP models have redirected resources away from “already under-resourced safety-net systems to well-resourced ones serving less vulnerable populations.”
For example, under Merit-based Incentive Payment System (MIPS), which is a VBP program created under the Medicare Access and CHIP Reauthorization Act (MACRA) in 2015, “providers in safety-net systems and those serving the highest proportion of dual eligible Medicare/Medicaid patients performed worse and had more frequent financial penalties. In addition, “[p]roviders in systems serving the highest proportion of racial and ethnic minority patients also fared worse.” Similarly, all of Medicare’s hospital VBP models (the Hospital Readmissions Reduction Program, the Hospital Value-Based Purchasing Program, and the Hospital-Acquired Condition Reduction Program) have disproportionately penalized and transferred resources away from safety-net hospitals.
These results have an adverse effect on voluntary VBP models as well. Because VBP models do not typically account for social risk factors and the results have shown adverse effects on providers serving these populations, providers who serve social risk populations have no incentive to join a system that will only penalize them. If VBP is capable of lowering costs and improving quality of care, then social risk populations will be excluded from these benefits if their providers do not participate in VBP models.
VBP has been shown to exacerbate healthcare disparities even outside of VBP models by incentivizing cherry-picking within VBP models. Providers who are entering or already participating in a VBP model are disincentivized to care for patients with social risk factors within the VBP model in order to optimize their chances of performing well on benchmarks and receiving bonuses or shared savings. The incentive for providers to choose only healthier patients, which typically excludes large portions of social-risk-factor populations, leaves even fewer providers outside of VBP models to care for vulnerable populations. Reducing the number of providers available to care for social-risk-factor populations is likely to widen disparities even outside of VBP models.
To summarize, while VBP models may realign incentives for providers to focus more on quality and outcomes rather than volume of services, these results show a need to realign incentives around health equity as well. Both mandatory and voluntary VBP models widen health inequalities for social-risk-factor populations. Mandatory VBP models direct resources away from social-risk-factor populations because providers are either penalized for serving them or choose to cherry-pick healthier patients and not serve those populations at all in order to avoid penalties. Voluntary VBP models also direct resources away because providers serving social-risk-factor populations will either not participate in VBP models and continue to be unaccountable for their costs and outcomes or, like in mandatory models, will choose to serve only healthier patient populations and leave social-risk -factor populations with fewer providers to care for them outside the VBP models. However, these issues have been recognized by CMMI, and VBP models could potentially be altered to address them.
The Centers for Medicare & Medicaid Services (CMS) released a “Strategy Refresh” in 2021 that made health equity one of the five main strategic objectives for the next decade to achieve their vision of “equitable outcomes through high-quality, affordable, and person-centered care.” CMMI will begin to consider equity at every stage of model development, testing, expansion, and termination. All new models will include patients from historically underserved populations and will require participants to report the demographic data of their beneficiaries. However, as social-risk-factor data has not been routinely collected up to this point, these next steps are largely centered on data collection and analysis.
Many solutions have been suggested to fix VBP’s health equity problem. The most obvious solution would be to adjust the benchmarks downward to account for social risk factors, however, that is also the most controversial solution. Proponents of this solution argue that social risk factors are outside of the provider’s control and accounting for them will allow providers to participate in VBP models without being punished for caring for vulnerable populations. These patients will have worse outcomes even if the provider delivers the same quality of care received by patients without social risk factors, so benchmarks should adjust for the differences in outcomes so that providers are not penalized. On other hand, critics argue that what may appear as differences by social groups may be genuinely attributed to quality differences and not the social factors themselves. In this case, adjusting for the social factor would obscure genuine disparities and make it more difficult to hold those providing lower-quality care accountable. The concern is that altering the benchmarks for social-risk-factor populations “implicitly accepts a lower standard for vulnerable patients” and could further entrench healthcare inequalities by reducing incentives for improvement.
The Department of Health and Human Services (HHS) has sided with the critics and does not plan to adjust benchmarks to account for social risk factors. Because the overarching goal is to achieve health equity and not simply to level the playing field for providers, HHS will allow benchmarks to be adjusted for resource use and patient experience, but not for process and outcome measures. Providers should be required to strive for the same high outcomes for all patients regardless of their social risk factors.
HHS recognizes that this approach would continue to punish providers for serving social-risk-factor populations without additional changes. To prevent this, HHS directs CMS to provide additional resources to these providers to assist them in caring for their social-risk-factor populations. For example, additional supports may include sharing best practices through learning networks and encouraging medical providers to build links with social service providers. Additional resources to meet beneficiaries’ social needs may be made available through alternative payment models, supplemental benefits that address social needs, or additional payments. Targeted payments to support providers’ efforts to address social risk factors may also be made through VBP incentive payments.
Once implemented, this approach could prevent resources from being redirected away from social-risk-factor populations within VBP models and remove the incentives for providers to cherry-pick patients.
Another similar solution to is to adjust penalties based on peer groups. Instead of comparing providers serving social-risk-factor populations to those that do not and adjusting the benchmarks according to their performance to avoid penalties, providers would be placed into peer groups with similar patient mixes and only compared to each other for calculating penalties. This method would incentivize providers to improve care without holding them accountable for factors beyond their control. However, HHS has declined to adjust benchmarks downward for social risk factors at all, which means this strategy has also been rejected for the same reasons.
Other solutions include adding quality measures related to health equity and integrating health services and social services in order to target the underlying social risk factors. Creating new benchmarks related to improving health equity in VBP models could be a valuable tool for combatting the underlying social risk factors. However, incentivizing providers to reduce health outcome disparities may not be effective if the effects of social risk factors are truly beyond their control. This strategy could only work if effective health equity measures are developed and providers can use that data to screen and address the unmet social needs of their patients. In addition, integration between health services and social services can improve health outcomes and research utilization. Therefore, tying social services into VBP models could help providers meet the new health equity benchmarks.
Pursuant to their strategy refresh, CMMI has begun to create new VBP models and make changes to current models in order to target health equity. For example, CMMI’s accountable care organization (ACO) Realizing Equity, Access, and Community Health (REACH) model began on January 1, 2023, with health equity as a top priority. The ACO REACH model requires participants to collect data on social risk factors, establish an “equity strategy,” and adopt initiatives to reduce health disparities.
The model follows HHS’s new strategy to alter benchmarks for resource use, but not outcomes, and provide additional resources to providers serving social-risk-factor populations. In what is labeled a “health equity benchmark,” spending benchmarks are increased for providers caring for disproportionately large social-risk-factor populations in order to reduce disincentives to serve those populations. This allows providers to give the additional care these populations need without worrying about the extra cost. In addition, providers receive the additional financial resources they are allotted through the benchmark upfront so that they can develop the new infrastructure and resources needed in order to redesign care delivery. In a way, these upfront payments are reversing the “Reverse Robin Hood” by “shift[ing] upfront financial resources from organizations caring for less marginalized patients to those caring for more.”
CMMI has also begun to make alterations to current models to improve their impact on health equity. The Medicare Shared Savings Program (MSSP) is Medicare’s primary ACO program and contains the vast majority of Traditional Medicare ACO beneficiaries. In 2023, providers will begin to receive “advance investment payments” that increase based on the size of their social risk population. CMS is also increasing the cost benchmarks for ACOs serving “high cost of care populations” in order to incentivize them to join the MSSP and increase the likelihood of shared savings. Both these changes to the MSSP and the creation of models like ACO REACH put CMMI’s strategy in action, but it remains to be seen whether they can solve VBP’s health equity problems.
In order for VBP to achieve its goals of reducing costs and improving outcomes, it must take health equity into account to avoid increasing health disparities. In fact, populations with social risk factors may benefit the most from VBP reform because “improved care coordination and provider cooperation will be of the highest utility to the most complex beneficiaries with the most care needs.” However, up to this point, VBP models have had a negative effect on health equity by penalizing providers serving social-risk-factor populations, disincentivizing providers to care for them, and redirecting resources to providers who are not serving social-risk-factor populations.
Finding a solution to this problem is an urgent matter for VBP reform because models have generally not generated cost savings unless providers take on substantial downside risk. Providers are unlikely to take on such risk unless it is mandatory. However, mandatory VBP models have produced the “Reverse Robin Hood” effect and widened health disparities.
CMS has recognized VBP’s health equity problem and made it a top priority going forward. Their strategy for improving health equity revolves around continuing to require the same high-quality outcomes for these populations, but increasing spending benchmarks and upfront financial resources to providers serving them. While these changes may improve health equity through VBP, it is important to note that these measures are targeting the downstream effects of social risk factors. Health equity will likely remain an issue in the United States until upstream problems creating social risk factors are eliminated.
