Abstract
- Issue: Medicare, like other health care payers, continues to test alternative primary care–focused payment and delivery models. Evaluations of Medicare’s primary care models to date have found they do not meet statutory criteria for reducing overall spending or improving quality, both key goals for Center for Medicare and Medicaid Innovation (CMMI) demonstrations.
- Goals: Describe the characteristics of these payment models and explore options for how they could be refined to increase their likelihood of reducing Medicare spending, improving quality, or both.
- Methods: Comparison of primary care payment models, supplemented by interviews with leaders of primary care–focused provider organizations that rely on alternative payment arrangements but have generally not participated in CMMI models.
- Key Findings: The two newest primary care payment models offer more flexibility than prior approaches. Providers favor this flexibility but also suggest replacing fee-for-service with alternative payment arrangements and shifting more care to primary care providers.
- Conclusion: To increase the likelihood that models achieve overall cost savings and/or quality improvement, one option for CMMI is to test approaches that add flexibility and significantly increase the share of total health care spending devoted to primary care.
Introduction
Primary care is vital to a high-quality, high-functioning health care system.1 The U.S. health system, however, is not centered around primary care. While the United States spends about 5 percent to 7 percent on primary care as a share of total health care spending, other high-income countries devote about 14 percent.2 This difference may be one of the factors that contribute to the comparatively lower life expectancy, fewer physician visits, and more frequent hospitalizations seen in the United States, despite the U.S. spending more on health care as a share of gross domestic product than its peer nations.3
The Medicare Payment Advisory Commission (MedPAC) has determined that, like the rest of the health care system, the Medicare fee-for-service physician fee schedule undervalues primary care services relative to other services.4 And because the fee schedule is focused on discrete services, it may not be suited to supporting primary care, which includes coordination of multiple, patient-centered services for a longer period.
This situation contributes to troubling access issues for Medicare beneficiaries. MedPAC reported that in 2020, 38 percent of Medicare beneficiaries who tried to find a new primary care provider had a problem doing so.5
Federal Efforts to Increase Payment for Primary Care
The Centers for Medicare and Medicaid Services (CMS) and Congress have implemented various efforts to raise total payments for primary care providers, directly and indirectly. These efforts remain rooted in fee-for-service payment, however, and none shift the share of total spending on primary care close to the doubling or more that would be needed to match other countries’ investments in primary care.
One program, Primary Care Incentive Payment (PCIP), introduced an add-on to fee-for-service payment rates for primary care providers. Created by the Affordable Care Act and in operation from 2011 to 2015, it provided a quarterly 10 percent bonus payment on top of fee-for-service payments for certain primary care services provided by eligible primary care practitioners.6
In addition, multiple fee-for-service payment codes have been added to the Medicare physician fee schedule that allow providers to bill for care coordination and planning services, including time spent on non-face-to-face management. Although not reserved for use solely by primary care providers, these codes are more likely to support the ongoing, high-touch services that they deliver. For example, in 2020, CMS proposed a new payment add-on code for complex visits.7 While also not specific to primary care, the code is for evaluation and management visits that are part of an ongoing, longitudinal care relationship.
Medicare’s Primary Care–Focused Models
CMS is also testing new ideas to support primary care in Medicare through Center for Medicare and Medicaid Innovation (CMMI) models that are designated in the category “primary care transformation” (see Appendix A). For various reasons, none of the models completed so far have been found to meet the statutory criteria of lower spending or improved quality, and none have been expanded nationwide.8 These early primary care–focused models offered important lessons, however, and some of the promising characteristics tested continue to be incorporated and refined in new models.
CMMI is currently testing three primary care transformation models and recently completed a fourth that build on some of the lessons learned from the earlier models:
- Comprehensive Primary Care Plus (CPC+)
- Global and Professional Direct Contracting (GPDC) (GPDC will be rebranded as Accountable Care Organization Realizing Equity, Access, and Community Health Model, or ACO REACH, effective January 1, 2023)
- Independence at Home
- Primary Care First
Of the four models, the two newest ones — GPDC and Primary Care First — began in 2021 and thus do not yet have evaluation results. The two models begun earlier have initial, independent evaluation results that do not yet indicate the U.S. Secretary of Health and Human Services will be able to expand them into the Medicare program nationwide.
The four most recent primary care–focused CMMI models include some shared and unique characteristics, as detailed in Appendix B. All four include some performance-based incentive payments. All but the Independence at Home model incorporate capitation for some portion of total payments. The Primary Care First model introduces the innovation of replacing payments using the Medicare physician fee schedule with a mix of payments including flat fees for visits and a monthly capitated payment that is risk-adjusted for patient characteristics (see Appendix C for more details on payment approaches).
Insights for Refining Primary Care–Focused Models
Given the urgency of improving primary care for Medicare beneficiaries — both in light of the COVID-19 public health crisis and the recent MedPAC survey results — and the failure of CMMI models to meet their spending and quality goals, we sought input on how the current models might be refined or future models designed to increase the likelihood of success. We interviewed representatives of three health care provider organizations — ChenMed,9 Cityblock,10 and Iora Health11 — that have established innovative, primary care–focused approaches to person-centered care funded through alternative payment arrangements with Medicare Advantage, Medicaid, and commercial insurance plans.12 The care models designed by these three organizations share characteristics with CMMI’s primary care–focused demonstrations and differ from typical primary care practices in not accepting fee-for-service payment.
Although the three organizations did not participate in early primary care–focused CMMI models, Cityblock and Iora Health have recently begun to participate in the GPDC model. Leaders from the two organizations noted that their decisions to participate were based in part on the much greater flexibility that the newest models offer compared with the earlier options.
Following are the key observations and suggestions made by at least one representative of these primary care provider organizations.
- CMMI models, traditional Medicare, and health care in general would likely benefit from moving away from fee-for-service payment. Fee-for-service is not an ideal payment method for health care and may be especially ill-suited to primary care. It often presents the wrong incentives to provide more services and does not include sufficient incentives for care coordination.
- The big changes to primary care that CMMI is testing are welcome by providers. Small, incremental changes may not be worth pursuing. Models that adjust small details while retaining many of the features of the existing fee-for-service approach are not likely to achieve success.
- Shifting a greater share of total care resources from specialty care, hospitalizations, and other high-intensity care to primary care would be worth testing. When overall care for patients is structured so that primary care clinicians have smaller patient panels combined with greater autonomy and flexibility, they can have more frequent, longer visits with patients, devote time to non-face-to-face coordination, and tailor care plans to be person-centered. This may result in better chronic condition management, higher patient satisfaction, more judicious use of specialists, fewer hospitalizations, and better health outcomes.
- Primary care–focused models should encourage more providers to participate; including other groups like large health systems and managed care plans as participants may dilute the goals of the model. These models offer an opportunity to support primary care providers in delivering comprehensive, continuous care. Large health systems and similar organizations include various specialty providers and facilities that may resist shifting a greater share of total care resources to primary care. Managed care plans are already paid based on the total cost of care, so they do not need this additional support and, as payers rather than primary care providers, do not share the same mission-driven outlook.
- Accountable care organizations and other health system models whose participants include hospitals have difficulty reducing hospitalizations to achieve cost savings. Reducing hospitalizations and rehospitalizations, as well as managing postacute care, represent some of the most significant opportunities to manage beneficiaries’ care better and achieve savings. When hospitals are included as participants, they are unlikely to agree to significant shifts in care patterns that could reduce spending on hospital services.
- Risk-adjustment methodologies could be improved by incorporating social determinants of health. These factors, such as housing inequity, food insecurity, and lack of access to internet and other electronic resources, must be addressed to truly manage beneficiaries’ care.
- Models could benefit from the inclusion of Part D drug spending. Working with patients to coordinate their prescription drug utilization is integral to effectively managing Medicare beneficiaries’ care. One policy option to consider is encouraging beneficiaries to choose a Part D plan aligned with the model they are attributed to so that participating providers have greater insights into their patients’ drug regimens and evaluators are able to assess the model’s effect on total Medicare spending, including Parts A, B, and C.
- There is no effective way to incorporate cost sharing to address beneficiaries’ care choices when beneficiaries have secondary coverage that significantly reduces or eliminates cost sharing. Models in which participants are responsible for the total cost of care that enrollees incur share similarities and differences with managed care plans. Two key differences are that model enrollees retain their full rights to see any willing Medicare provider (they are not required to see providers who participate in the model), and they can keep secondary coverage, such as Medigap. Medicare Advantage plans are able to influence their enrollees’ care choice through the provider network they select, through cost-sharing requirements, and through other utilization management tools. Adding similar tools to allow models to influence enrollees’ care choices could increase the likelihood of meeting CMMI’s savings and quality goals. Any change would be challenging and might require legislation, but if this issue were addressed, it would allow a more apples-to-apples comparison of Medicare Advantage plans, the Medicare Shared Savings Program, and primary care–focused CMMI models.
- Models could be designed to encourage new enrollment of innovative organizations rather than ongoing sign-ups of incumbent participants. Given the limited success to date of CMMI models in reducing costs and improving quality, it makes sense to test some new approaches rather than focusing on renewing models for second and third iterations with a similar group of participants. A better understanding of what it would take to attract providers that have yet to participate could provide new opportunities for CMMI to shift care patterns away from the status quo.
- Primary care–focused models take several years to yield cost and quality results, but there are markers along the way that indicate progress. Some of these same markers have been observed by independent evaluators of primary care–focused models.13
- Patient engagement increases, indicating that beneficiaries trust their providers and are actively participating in their own care.
- Outcomes improve, such as diabetes management and measures of patients’ confidence in self-managing their care.
- Hospitalizations decline, signaling that patient needs are being met in a less intensive setting and that overall care is better managed.
Conclusion
Placing greater value on primary care and giving primary care providers greater flexibility to care for people outside of a fee-for-service system could improve people’s access to care, the quality of care received, and quality of life. For years, the Medicare program has sought methods to do this within the traditional fee-for-service program and within CMMI models. So far, these efforts have had little success in meeting the statutory criteria of decreased spending and/or improved quality. Primary care remains undervalued compared with specialty care, and no primary care–focused models have been expanded to the entire Medicare program.
The two most recently introduced primary care–focused models — Global and Professional Direct Contracting and Primary Care First — represent a promising step forward. (GPDC will be rebranded as Accountable Care Organization Realizing Equity, Access, and Community Health Model, or ACO REACH, effective January 1, 2023.) They offer greater flexibility to primary care providers and include far less reliance on fee-for-service payment. These model features could help participating providers advocate for their patients and tailor care to patients’ individual needs. As the representative of one organization succinctly put it, “Does care for the patients actually change? If not, the model is a waste of time.”
It may be years before there are measurable results from the primary care–focused models underway. Given that there may only be a year or two left for the models to operate by the time these results are available, one option for CMMI to consider is refining the models now to incorporate the suggestions raised by leaders from organizations that have found success focusing on primary care outside of Medicare.