This article first appeared in healthaffairs.org. It is co-authored with Amol S. Navathe, Ezekiel J. Emanuel, Sherry Glied, and Farzad Mostashari.
The federal government is facing a fiscal crisis with impending insolvency of the Medicare Trust Fund in 2024, a long-standing challenge accelerated by the COVID-19 Public Health Emergency. Health care continues to be unaffordable for many Americans with persistent health inequities, for those of low socioeconomic status, and for those in minority communities. Addressing the short- and long-run problems of the Medicare Trust Fund and health care affordability will require improving the efficiency of health care delivery, achieved by reducing payment rates and decreasing service volume (i.e. overuse or redundant care) while maintaining or improving quality.
The Center For Medicare And Medicaid Innovation (CMMI) is one mechanism to improve efficiency. The first 10 years of CMMI saw numerous models of payment reform tested: primary care vs. specialty care vs. all care, voluntary vs. mandatory participation, upside only vs. 2-sided risk models, Medicare only vs. multi-payer efforts, state innovation models, etc. While single model evaluations of these new models have shown only modest and occasional savings, it is likely that the combined effect of payment changes across models catalyzed greater impacts than these studies suggest — including negative per capita spending in Medicare. Reductions in rates may achieve short-term savings, but will require complementary value-based efforts to produce more health and bend the long-term trajectory.
The COVID-19 Public Health Emergency (PHE) has once again laid bare stark disparities for low income and minority communities. But it has also highlighted the dramatic scope and scale of shifts—such as more telemedicine and home care— that the national health delivery system can achieve. We believe that as CMMI embarks upon its second decade, it must consolidate and institutionalize these shifts and strategically test and scale new payment models to achieve:
- Permanent changes in Medicare payment policy based on targeted CMMI models that generate financial savings as certified by CMS Actuary. Few CMMI demonstrations to date generated successes that have become payment policies, and yet CMMI created too many models for clinicians and health delivery organizations to reasonably manage. CMMI should test a smaller number of models that can be scaled nationally.
- Enhanced health care equity through the Medicare and Medicaid programs. There is emerging evidence that value-based payment reforms hinder access to care (e.g., access to elective surgeries) and worsen disparities for low socioeconomic and minority populations. To date models seem to emphasize cost and quality accountability in ways that may inadvertently lead providers to adopt practices that harm low income and/or minority communities. The path forward must include a twin focus on lower costs and improved health care equity – these goals must be integrated so value-based payments address drivers of health as a pathway to better health care value.
1. Implement A Cohesive Strategy Of Coordinated Alternative Payment Models (APMs)
The three types of APMs to date – advanced primary care, episode-based specialty, and population-based (e.g. ACOs or Direct Contracting)– should be aligned under a common framework to create synergies and enhancements across participating provider groups. This should include articulating the end-state delivery and payment models to clarify how models are deployed to engage provider groups across the spectrum of capability and readiness for APMs – including support for independent physician groups. This will also enable focusing in on ‘best in class’ models that will simplify the payment reform landscape for delivery organizations. Importantly, each model should have a path to mandatory participation, Indeed, mandatory models have received bipartisan support.
A Cohesive Strategy Would Create A “Marketplace” For Health Services Across The Continuum Of Care And Across APMs
One example of a cohesive strategy would be to include larger delivery organizations with total cost accountability in population-based models that utilize episode-based specialty care (i.e., care that is intrinsically episodic in nature). Episode-based model participants would earn savings by beating the market median benchmark; population-based model participants would also earn savings by beating the market median benchmark for all care. Making episode costs transparent to population-based model participants will stimulate competition on episode costs among specialists and facilities and provide a path to coordinated financial incentives.
Service-line level innovation has proven to be more tractable and quicker for delivery systems, yet population-based payments are required to 1) account for substantial chronic and preventive care that is not episodic and 2) offset incentives that could lead to preferential risk-selection or gaming like behaviors. Advanced primary care models should be used with two objectives. The first objective is to support capacity building, infrastructure investments, and experience for provider groups without experience in managing populations as they organize into broader population management organizations. The second objective is to pay the linchpin drivers of access and health care value in primary care in a fixed, capitated mechanism that confers greater flexibility even within population-based models to mitigate resource allocation and empower primary care service lines within larger organizations by increasing their revenue streams.
Thus, advanced primary care models should both evolve into population-based models but also persist as a payment mechanism within them. This approach to coordinating across the three APM types should create a framework under which participants will be incentivized to direct patients to the most efficient delivery organizations. It also mitigates conflicts and competition between population- and episode-based models, while creating a marketplace like approach for referrals into episodic and primary care. This marketplace will incentivize population-based model participants to selectively clinically integrate with or refer to efficient facilities and specialists, offsetting hospital market power and introducing competition in regional markets in spite of price regulation.
The linchpin will continue to be accountable care organization (ACO)-like models that have capacity to manage populations – to some extent mimicking the role of Medicare Advantage plans within traditional Medicare. These will be supplemented with specialty (part B) models such as evidence-based bundled payment models, new home-based APMs, and targeted flagship initiatives (e.g., Zero CHF readmissions, diabetes reversal, or 100% blood pressure control).
In the immediate term, CMMI will be able to build upon existing models rather than disruptively cancelling active or planned demonstrations. The basic structure of population-based, episode-based, and advanced primary care models provides a foundation on which to aggressively innovate, coordinate, and re-align incentives. (Exhibit 1) Over time, CMMI should consolidate the approach around the strategic plan.
Exhibit 1: Coordinated Alternative Payment Models
Source: Authors’ analysis
Program Designs Would Use Pragmatic Health Care Market-Based Incentives Employing Behavioral Economic Principles
An important yet simple insight from past CMMI models is that designs that rely on delivery organizations to cut their own revenue in the hopes that earning back a portion of a dollar saved are flawed. “Race-to-the-bottom” designs that use historic benchmarks also create conflicting incentives for participation and reduced investment in infrastructure and capabilities.
CMMI must push forward with pragmatic regional health care market-based benchmarking approaches. These approaches will either phase-in incentives or utilize a voluntary period to transition to full mandatory participation Benchmark Incentives should be ‘reverse engineered’ from financial solvency and affordability goals, allowing for longer-term stability but accounting for often dynamic secular trends. And they should aim to enroll all current non-participants in APMs. Aligning the benchmark incentives with the APM framework above will create greater transparency and a stronger “marketplace” for health care services where prices are regulated. This will also more easily enable accurate monitoring of program impact to more rapidly make assessments for the CMS Actuary. Staggered rollouts across markets and phasing in the APMs over a few years could be a particularly useful mechanism to increase participation.
Incentive designs should be paired with powerful non-financial interventions from behavior economics, such as using peer comparisons that activate professional norms and changing default/status quo patterns of care.
APMs Would Preserve And Incentivize Telehealth, Home Care, And Other Preference-Oriented Favorable Trends From The PHE
Several aspects of care such as telehealth, remote monitoring, and increased home health care can provide lasting benefits. While it would be pragmatic to target flexibilities toward existing or aspirational APM mechanisms for value, greater leniency among APM participants will mitigate concerns around fraud and abuse.
2. Declare Reducing Health Care Inequities As A Twin Goal Alongside Improving Value
While equity is not specifically included in the CMMI statute, it is authorized implicitly as an essential element of quality. Value-based payments are necessary for affordable, high quality health care for all Americans. Yet, like fee-for-service, value-based payments have the potential to maintain or exacerbate existing health care disparities, which has been shown in population-based and episode-based models to date. To most effectively protect beneficiaries in vulnerable groups, CMMI must pursue health care equity as a twin goal alongside improving health care value.
Require Measures And Incentives That Directly Reward Improvements In Equity
CMMI should be a national leader in setting standards to monitor and incentivize improvements in health care equity. This will require investments in better data (e.g., on race/ethnicity), measurement, and risk-adjustment methods. Programs should require participants to uniformly screen for and document drivers of health — using the Accountable Health Communities Screening tool or others. One concrete short-term step should be to make documenting existing drivers of health-related ICD-10 Z-codes (Z55-Z65) reimbursable including “Inadequate Material Resource” codes that address food, housing, and income. Screening for and documenting health-related social needs will provide needed data to inform actuarial analyses, risk adjustment, rate-setting, investment of public and private funding (philanthropy and community benefit dollars), and strategies to address racial inequities. To further align across CMS, CMMI should build drivers of health measures into MIPS and all APMs. Providers should be eligible for financial bonuses for quality and cost savings only if they achieve improvements in equity along standardized measures. While CMMI would provide technical assistance, it would also require implementation plans for participating providers to articulate intervention strategies. This could include community health worker programs or integration with social services such as SNAP as a core mechanism to bridge directly into the community.
As an example, hip and knee replacement surgery bundles have been tested by CMMI in voluntary and mandatory reforms. Yet, Black beneficiaries are less likely to receive hip/knee replacement surgery. Measure sets should include metrics around equity in community inclusion in bundled payment programs (to ensure generalizability at scaling), access to surgery, and health outcomes such as functional status.
Numerous states are now utilizing measures that actively monitor health disparities in Medicaid programs. For example, Michigan is publishing 13 measures by race and ethnicity publicly while also assertively pursuing reductions in disparities for low birth weight infants, while California and others are reporting measures for seniors, disabled individuals, and by race/ethnicity for managed care organizations. These measures can also be incentivized directly under alternative payment models. Oregon, for example, has used this approach to improve emergency department utilization among individuals experiencing mental illness within its Coordinated Care Organizations (similar to Medicare’s ACOs). And, as North Carolina has done, CMMI should support the development of statewide coordinated care networks to connect community resources to patients with unmet social needs.
Test And Scale Models Specifically Tailored To Communities Facing Inequities
Many communities and hospitals face challenges that cannot be addressed through “standard” APMs and programmatic incentives. Rather, they require a tailored approach. An excellent example is the Pennsylvania Rural Health model, which attempts to address access and sustainability issues for rural hospitals. Disease groups with high proportions of racial minorities like end-stage renal disease or dual eligible beneficiaries may require similar approaches that could vary by geography.
Launch Several Flagship Pursuits Of Equity
A flagship pursuit will draw appropriate attention, advocacy, energy, and resources. This could focus on vexing equity challenges in diseases concentrated in vulnerable communities. A prime example would be a focus on end-stage renal disease (ESRD) care with goals such as doubling kidney transplant rates or halving progression of chronic kidney disease for Blacks.
3. Re-Evaluate The Fee-Schedule
The professional fee schedule used by Medicare does not reflect the value of health services and is heavily biased toward procedures. Many procedures are overvalued while primary care services are undervalued. While upcoming changes shift in the right direction, systematic approaches to aligning the fee schedule with value could generate substantial savings. Changes to the fee schedule could also provide greater impetus for APM participation. A concrete step would be to re-price the top 250 HCPCS codes and use actual times in electronic medical records (EMRs) to price procedures. Changes to Medicare rates, professional and facility, could achieve short-term savings.
4. Convene More Robust And Expansive Multi-Payer Efforts To Enhance Practice Pattern Changes That Support Health Care Affordability
Medicare has an outsize impact on health care prices and practices that extend into the private insurance sector. The market power dynamics that exist between delivery organizations and health insurers hinder the adoption of APMs and unit price competition. Further, there is evidence of positive spillovers of benefits created by Medicare payment reforms on non-Medicare patients. This places CMMI in a unique position to catalyze broad health sector changes around health care affordability, equity, and value (while also avoiding antitrust issues).
Place Health Care Delivery Organizations And The Practitioner At The Center Of Multi-Payer Efforts
Because of the volume and variety of payment reform initiatives across federal, state, and private payers, the operational complexity for the practitioner has ballooned. CMMI can serve as an alignment vehicle by using its convening power to develop standards that could be easily adapted to local and payer specific settings to reduce this burden and unlock tremendous value for patients, delivery organizations, and payers.
Convene The Health Insurance Sector To Align On Payment Reform Standards
While CMMI has advanced some multi-payer initiatives in primary care, cancer care, and Medicare Advantage, these have functioned like pilot initiatives rather than impact at scale. To date, each health insurer implements value-based payment initiatives with its own set of design features such as quality metrics, financial incentive formulas, and data reporting. This exacerbates the issues around market power and makes participation dramatically more costly and onerous for delivery organizations. CMMI could convene and publish standardized designs for multiple models that are not so unique to Medicare populations that they are inapplicable outside of the elderly. These would include equity metrics as part of standards.
Enhance Efforts To Collaborate With States And Medicaid/CHIP
A central aspect of multi-payer efforts would include engagement with states. Section 1115 demonstrations and waiver authorities in section 1915 of the Social Security Act have allowed many states to innovate within Medicaid. Since many states primarily deliver Medicaid via managed care organizations (MCOs), this creates additional opportunity to create alignment across public and private payers. Yet again, state efforts have generally complicated delivery reform at the individual practitioner level. Instead, state waiver programs should align with a strategic framework for delivery reform, including requirements for MCOs.
5. Aggressively Pursue Administrative Simplification To Reduce Associated Costs And Clinician Burnout In All Models
CMMI can make important contributions and build engagement from front-line clinicians by testing administrative simplification programs. These could include fundamental activities such as standardized credentialing and claims processing, but also provocative solutions around automatic prior-authorization, point-of-care incentives instead of prior-authorization, or new IT enhancements that eliminate coding requirements.
A comprehensive approach to health sector improvements will require coordination with other agencies on other domains such as health information technology and interoperability (ONC), transparency (Congress), community resources (HRSA), and others. While the proposed recommendations are not exhaustive, these, in particular, should be feasible within CMMI’s statute – and thus, could be implemented directly by Center’s leadership.
We thank the following individuals for thoughtful discussions and/or feedback: Mark McClellan, Mandy Cohen, Elizabeth Fowler, Aneesh Chopra, Michael Chernew, Valinda Rutledge, Patrick Conway.