Introduction and summary
Although the national health care debate over the past year and a half has been dominated by efforts to roll back the Affordable Care Act and Medicaid, policymakers in many states have continued to take the lead on practical reforms to make health care costs more sustainable. States have a variety of tools to reduce health care costs without shifting costs to patients or making harmful changes to their Medicaid programs that would adversely affect beneficiaries.
In several states, innovative payment and delivery system reforms have been implemented for a long enough period of time to have generated a track record of results that can be evaluated. This report will review the results to date of major payment and delivery system reforms in Maryland, Massachusetts, Oregon, and Arkansas. Over the past few years, these states have launched ambitious reforms in an attempt to reduce health care costs while maintaining or improving health care quality, in stark contrast to recent waiver proposals by some states to roll back Medicaid eligibility and benefits through policies such as work requirements and lock-out periods.1 While the latter efforts merely shift costs to low-income individuals and families, the state reforms reviewed in this report attempt to tackle the actual underlying costs of health care.
Thus far, the results are promising. Maryland’s ambitious evolution of its all-payer rate setting system into a system of global budgets for hospitals has generated more than $500 million in Medicare savings, achieving its savings goals ahead of schedule. Massachusetts has established an annual cost growth benchmark that has seen mixed results, but the state has recently proposed additional reforms to address costs. Oregon has implemented a Medicaid Coordinated Care Organizations model that has successfully met its cost growth targets each year and generated $2.2 billion in state and federal savings. And Arkansas’ multipayer bundled payments for certain episodes of care and patient-centered medical home models have reduced costs and resulted in positive impacts on some quality and outcomes metrics.
These reforms take different approaches—from directly regulating hospital budgets to bundling payments for certain procedures—but they all share a focus on driving down actual costs and a commitment to maintaining or improving the quality of health care rather than shifting costs to patients. The experiences of these states demonstrate the value of state experimentation and indicate that a variety of approaches can work. Thus, states that are currently exploring approaches to addressing health care costs have several effective policy options to consider.
Maryland: All-payer global budgets for hospitals
Maryland’s all-payer rate setting system for hospitals, in which the state establishes universal payment rates for hospital services, has been in place since the 1970s. Under the system, a state government commission sets the payment rates used by health insurers to pay for hospital care, instead of those insurers negotiating their own rates individually with hospitals. This includes not only private insurers and Medicaid but also Medicare, whose participation is enabled by a waiver from the federal government.2
In 2014, the state launched a substantial overhaul of the system, evolving its management of hospital payment through the addition of global budgets for hospitals in order to better manage costs and quality.3 Under the new global budget system, the state now establishes overall budgets for hospitals rather than just setting the price of each individual service. Prior to the launch of the new system, the state piloted global budgets in several rural hospitals.4 Maryland’s process for developing the global budget levels is complex and incudes population adjustments to account for demographic changes and market shift adjustments to account for trends in patients moving between different hospitals.5
A major rationale for this evolution was to address the incentives for volume of a fee-for-service system. Although Maryland’s system had been successful in bringing down prices “from an average cost per hospital admission that was 25 percent above the US average to the middle of the pack,” the volume of hospital services had in recent years increased more significantly than in other states.6 Researchers and policymakers believed that hospitals were responding to lower price growth by working to make up the difference in volume, which was possible because the state was still paying for each admission.7
Initially, the rate setting system successfully avoided incentivizing excess volume through the inclusion of an adjustment for hospital volume, in which excess volume above a certain baseline was paid for at a lower rate. However, this adjustment was eliminated in 2001 as pressure from hospitals grew on the rate-setting commission, after which the marked increase in volume began.8 As the Urban Institute explains:
During the period 2001–08 admissions in Maryland grew at 2.4 percent per year, compared to only 0.8 percent in Maryland in the previous decade and approximately 1 percent per year nationally during those years.9
Although the state reimplemented the volume adjustment in 2008, the impact of the intervening years along with other policy changes had put Maryland at risk of potentially failing to live up to the terms of its federal waiver by 2013, spurring policymakers to pursue a broader evolution of the system.10
Despite Maryland’s own decision to pursue global budgets, this history indicates that Maryland’s earlier iteration of all-payer rate setting, rather than being an outdated or ineffective method, still holds significant promise as a model for state or federal reforms if the appropriate adjustments are implemented to avoid incentivizing excess volume.
As a result of the state’s increase in volume, Maryland’s per-beneficiary Medicare costs and rate of hospital admissions were high relative to other states.11 Under the new global budget system, Maryland aimed to hold annual hospital spending growth to the same level as the long-term growth rate of the state’s economy, using a 10-year projection of 3.58 percent.12
In addition, policymakers hoped that the new system would incentivize hospitals to invest more heavily in population health and community health initiatives, while also improving the state’s performance on preventable health care utilization measures such as hospital readmissions and ambulatory care sensitive admissions.
The new system was negotiated with the federal Center for Medicare & Medicaid Innovation (CMMI) as an innovation model, meaning that Maryland is accountable to the federal government to meet certain quality and performance goals, as well as to produce a specified level of Medicare savings.13 Under the agreement, Maryland has committed to producing $330 million in Medicare savings over five years by holding Medicare per-capita hospital cost growth to 0.5 percent below national Medicare per-capita hospital cost growth annually.14 If Maryland were to fail to achieve these objectives over five years, then the state would revert to the standard Medicare payment system, rather than reverting to the previous iteration of the state’s all-payer rate setting system.15
Maryland’s transition to the new system took place ahead of schedule. Under the CMMI agreement, Maryland was supposed to transition 80 percent of hospital revenue away from fee-for-service payments by the end of 2018. Yet in practice, by mid-2014, the state had already transitioned 95 percent of hospital revenue away from fee-for-service.16 According to stakeholders, Maryland hospitals largely concluded that it would be a better business strategy to avoid a period of uncertainty and make a quick transition away from fee for service to set themselves up for success under the new global budget system.17 This is a dynamic that might not necessarily repeat itself in other states, given Maryland’s long history of hospital rate setting.
In addition, Maryland achieved the five-year Medicare savings target it had promised the CMMI well ahead of schedule. According to the state Health Services Cost Review Commission (HSCRC), by the end of 2016, the new model had already generated $586 million in Medicare savings from reduced hospital spending growth, significantly exceeding the 2018 target of $330 million.18 On a per-capita basis, Maryland’s all-payer annual hospital spending growth has been held to an average of 1.53 percent, beating the 3.58 percent target rate.19
Similarly, an April 2018 evaluation by RTI International and funded by the Centers for Medicare & Medicaid Services estimated that Maryland’s global budgets had reduced Medicare hospital spending by $554 million, while resulting in total Medicare savings of $679 million over the first three years.20 Importantly, the evaluation found no evidence that Maryland’s global budgets were resulting in cost shifting from hospitals to other parts of the state’s health care system.21 However, the RTI International evaluation did not find evidence of reduced hospital spending for Maryland patients covered by commercial insurance, though the commercial data only covered the first two years of the global budgets.22
Finally, the RTI evaluation found mixed results on utilization, but some metrics did see notable progress.23 For example, potentially preventable hospital admissions dropped by 9.4 percent among Maryland Medicare beneficiaries relative to the comparison group, while overall inpatient hospital admissions declined somewhat for Maryland Medicare beneficiaries and commercial patients relative to the comparison group.24
In addition to financial savings, the HSCRC’s data indicate that Maryland has also made some improvements on quality of care and utilization. For example, potentially preventable complications have dropped by 44 percent.25 Although Maryland’s hospital readmissions rate has exceeded the national average in recent years, the state has reduced the degree to which it exceeds the national average by 79 percent and is on track to meet its goal of bringing it in line with or below the national average by 2018.26 Under the new model, hospitals have increasingly begun to focus on proactive support for patients, including adding case managers to emergency rooms and working to connect discharged patients to home-based or post-acute care settings.27
Two studies smaller than the RTI International evaluation have also been published recently, both of which raise questions about how much impact Maryland’s reforms have had to date on the utilization of hospital services.
An April 2018 study in Health Affairs examined data from seven of the eight rural hospitals that implemented global budgets in 2010 before the rest of the state; the data covered the years 2007 to 2013.28 Comparing Medicare beneficiaries using these hospitals to those using other Maryland hospitals—excluding hospitals in major urban areas—the study did not find that the global budgets were associated with changes in hospital utilization.29 Due to timing issues, this study did not include one participating rural hospital that has been cited by supporters of Maryland’s reforms as an example of making significant changes to care delivery.30
A January 2018 study in JAMA Internal Medicine sought to assess whether changes to the delivery system explained why Maryland hospitals were meeting key cost goals. The study compared Maryland Medicare beneficiaries receiving hospital care to beneficiaries in matched out-of-state hospitals over the first two years of the global budget program.31 Using two different methodologies, the study found certain positive changes in Maryland’s hospital utilization, but because the changes were not consistent across both methodologies, the authors concluded that any recent changes in utilization were not clearly attributable to Maryland’s global budgets and may be the result of pre-reform trends.32 A response article in JAMA Internal Medicine pointed out several limitations to the analysis, including the overarching question of “how quickly new payment models should be expected to transform the delivery system.”33 The response article noted that Maryland hospitals were still experimenting and adapting to the new model, and suggested that 5 to 10 years was a more realistic time frame in which to judge the impact on the delivery of care.34
Collectively, these studies indicate that Maryland’s global budgets are working to successfully control Medicare hospital spending. They also suggest that challenges remain in improving the delivery of hospital care and achieving further progress on costs and outcomes.
In December 2016, Maryland proposed its progression plan to begin transforming the rest of the state’s delivery system.35 The proposed progression plan would start in January 2019 and would expand Maryland’s cost growth commitments beyond hospital cost growth to the total cost of care for Medicare beneficiaries. Implementation would initially focus on Medicare beneficiaries who have chronic conditions or are dually eligible for Medicare and Medicaid, including by establishing accountable care organizations to coordinate care more effectively for these beneficiaries.
Under the progression plan, Maryland would continue to meet hospital cost growth targets. In addition, Maryland and the Centers for Medicare & Medicaid Services would expand this framework beyond hospitals by negotiating cost growth targets for Maryland’s total cost of care for Medicare beneficiaries, including Medicaid costs for dual-eligible beneficiaries.
One important component of the progression plan is a Primary Care Program that Maryland had proposed separately as a version of the CMMI’s Comprehensive Primary Care Plus model.36 Recognizing that hospital-focused community health initiatives to tackle chronic diseases are not enough, Maryland’s Primary Care Program is designed to improve primary care coordination, particularly for the treatment and management of chronic conditions.
Maryland is still in the process of negotiating with the federal government over both the progression plan and the Primary Care Program, and has not yet announced federal approval for either initiative.37
Massachusetts: Cost growth benchmark
In 2012, Massachusetts established an annual state health care cost growth target—in other words, a goal for the rate at which the state’s health care costs would increase. Under this benchmark, the state would attempt to hold health care cost growth to the rate of growth in the potential gross state product for the first five years, then under that rate for the following five years.38
For the purposes of meeting the benchmark, health care cost growth is calculated as the change in total health care expenditures per capita. The benchmark has no formal enforcement mechanisms except for the ability to place providers on performance improvement plans; rather, it is meant to spur progress by creating a sense of accountability among policymakers and health industry stakeholders.
Massachusetts has experienced mixed results in staying under the benchmark. The state successfully stayed under the benchmark in 2013 and 2016 but failed to stay under it in 2014 and 2015. Growth in these years is listed below (the benchmark in these years was 3.6 percent):39
- 2013: 2.4 percent
- 2014: 4.2 percent
- 2015: 4.8 percent
- 2016: 2.8 percent40
Notably, in both 2014 and 2015, when the state failed to meet the benchmark, prescription drug spending saw double digit growth, while physician and hospital services spending growth generally fell below the benchmark.41 In addition, 2014 saw a one-time jump of about 19 percent for state Medicaid spending as the Affordable Care Act’s Medicaid expansion was implemented, after which Medicaid spending growth dropped back down to normal levels in 2015.42 In 2016, the fastest-growing sector of health spending was prescription drugs, followed by hospital outpatient spending.43
For context, however, the state’s health care spending growth has improved relative to the national average, having been “similar to or below national spending growth since 2012, after exceeding national growth rates from 2002 to 2008.”44 For 2018, the benchmark has dropped to 3.1 percent; the state did have flexibility to increase it back up to 3.6 percent but decided to let the scheduled decrease take effect.45
Despite mixed success in meeting the benchmark, supporters believe that having a target performs an important accountability purpose. The benchmark forces an annual evaluation of health care cost growth and may contribute to pressure to adopt additional reforms. However, the lack of hard enforcement mechanisms means that the state has few concrete ways to force spending growth to meet the target.
The program does allow the state to place health care providers on a performance improvement plan if their costs are deemed to be excessive.46 Providers who have been placed on a performance improvement plan will be listed publicly online, in the hopes of this public shaming encouraging improvement. Some researchers have argued that this mechanism, along with the overall existence of the benchmark, has an impact on provider price negotiations, since “both payers and providers are aware that they will be subject to a performance-improvement plan through the [Massachusetts Health Policy Commission] if their high spending could potentially jeopardize the Commonwealth’s ability to meet the benchmark.”47
However, the state has not yet publicly announced the implementation of any performance improvement plans, so it is unclear how much effect the threat of this has had on providers’ behavior thus far. The Massachusetts Health Policy Commission passed regulations last year outlining the terms of the performance improvement plan process more clearly, indicating the commission’s interest in beginning to take a more aggressive role in using this authority to control costs.48
Alternate payment models
In March 2018, the state launched a major Medicaid accountable care organizations (ACOs) program that will result in more than 850,000 Medicaid beneficiaries being covered by an ACO, in which teams of health care providers will be responsible for the health care costs and outcomes of their patients.49 The federal government granted approval for the program through a Section 1115 Medicaid waiver, which will provide $1.8 billion in federal funding over five years to help launch and implement the ACOs.50
Importantly, these ACOs will integrate long-term services and supports and behavioral health care. In addition, the ACOs will have some flexible funding to invest in social determinants of health, including factors, such as housing and transportation, that can have significant ramifications for people’s health but are not traditionally covered by health insurance.
Proposed Medicaid waiver
In July 2017, the state proposed a Section 1115 Medicaid waiver amendment that, among other policy changes, would take significant action to reduce Medicaid prescription drug spending.51 Under the proposed waiver, Massachusetts would be permitted to establish a closed formulary, or list of covered drugs, for its Medicaid program in order to negotiate lower drug prices.52 A major factor that the state would consider when deciding to include or exclude a drug on the formulary would be the evidence of the drug’s clinical effectiveness relative to other drugs. However, it remains unclear whether the Centers for Medicare & Medicaid Services will approve this proposal.53
Unfortunately, the proposed waiver also includes a problematic proposal to shift Medicaid beneficiaries earning greater than 100 percent of the federal poverty level off Medicaid and onto marketplace coverage. If this is implemented without appropriate safeguards and wraparound benefits, it would result in these beneficiaries receiving fewer benefits and likely paying higher costs.
Oregon: Coordinated care model
In 2012, Oregon launched a program designed to improve the coordination of care in Medicaid. The model revolved around coordinated care organizations (CCOs), which are a form of accountable care organizations. The Oregon Health Authority describes a CCO as “a network of all types of health care providers (physical health care, addictions and mental health care) who have agreed to work together in their local communities to serve people who receive health care coverage under the Oregon Health Plan (Medicaid).”54
Oregon’s CCOs coordinate different types of care—including integrating mental health care, dental care, and substance abuse treatment—and are held accountable for care quality and population health outcomes.55 And importantly, the CCOs have some flexibility to invest in nonmedical services and supports that can improve social determinants of health.
To pursue these reforms, Oregon modified its existing Section 1115 Medicaid waiver. Under the terms of the waiver, Oregon would receive $1.9 billion in federal funding over five years to implement the CCO model, but the state had to reduce Medicaid per-beneficiary spending growth by 2 percentage points under trend by the end of the second year while meeting quality metrics, or face penalties.56 Since the baseline growth rate was 5.4 percent, this meant that Oregon needed to hold growth to 3.4 percent or less.57
Oregon’s Medicaid CCOs cover about 1 million people and are organized geographically.58 The state has 15 CCOs as of 2018, many of which originated as managed care organizations under the state’s Medicaid program and evolved to become part of the new model.59
Thus far, Oregon has been able to meet the 3.4 percent annual growth rate target for Medicaid spending each year.60 Compared with Massachusetts’ struggles in meeting its cost growth benchmark, Oregon’s task has been more straightforward. Since the target is solely for Medicaid cost growth and the CCOs are paid by the state prospectively through global budgets, the state government has a direct policy lever to hold annual Medicaid per-beneficiary spending growth to 3.4 percent or less.
In 2015, Oregon released an evaluation of the model’s performance to date.61 The evaluation included a survey of beneficiaries that found improvements over fee-for-service Medicaid in health care access, care quality, utilization of primary care, care coordination assistance related to social determinants of health, and self-reported health outcomes. CCOs did not perform significantly differently than fee-for-service Medicaid with respect to the rate of preventive screenings or emergency department utilization.
According to the Oregon Health Authority, the CCOs achieved $2.2 billion in federal and state savings from 2012 to 2017.62 In addition, the state continually tracks and reports certain metrics related to CCO performance, many of which have seen improvement. For example, avoidable emergency department utilization dropped from 14.2 member months per 1,000 member months in 2011 to 6.9 member months per 1,000 member months in 2016, while the percentage of young children who receive developmental screenings increased from 21 percent in 2011 to more than 62 percent in 2016.63
A 2017 study in JAMA Internal Medicine compared Oregon’s CCOs to Colorado’s more traditional Medicaid ACOs.64 The study found that spending on certain services in both states declined from 2010 through 2014 by roughly the same amount. However, Oregon did see greater improvements in access to care and quality of care.
Furthermore, a 2017 study in Health Affairs compared Oregon’s CCOs with Washington state’s Medicaid program from 2011 to 2014 and found that relative to Washington, Oregon’s approach produced savings of about 7 percentage points across five different service areas that were studied.65 In particular, Oregon saw promising reductions in hospital utilization and spending; concerningly, however, Oregon also had reductions in primary care office visits.
Finally, a March 2018 study in Health Affairs indicated that CCOs have made progress in beginning to reduce disparities in health care access and utilization by some measures for black and American Indian/Alaska Native Medicaid beneficiaries in Oregon.66
2017 waiver renewal
In January 2017, the federal government approved a renewal of Oregon’s waiver for five years. The renewed waiver continues the major elements of the 2012 model, while providing some additional flexibility to promote primary care and to pay for nontraditional services that improve social determinants of health, such as housing.67 Previously, this nontraditional spending was considered to be administrative rather than health-related spending for the purposes of calculating CCOs’ medical loss ratio, or the required ratio of spending on health versus profits and administrative costs.68 The waiver renewal clarifies that these types of flexible services that address social determinants of health can be considered “health-related” for the purposes of the medical loss ratio, removing an obstacle to scaling up these types of investments.69
However, Oregon did not receive the $1.25 billion in federal funding that it had requested, primarily for the purpose of continuing to expand the ability of CCOs to invest in social determinants of health.70 Currently, the scale of such investments in social determinants of health remains somewhat limited and varies by CCO. Improving upon this remains a focus of Oregon policymakers, as demonstrated by proposed state legislation that, among other CCO transparency and reform provisions, would require CCOs to invest a certain portion of their income or reserve funds in programs to improve health disparities or social determinants of health.71 Gov. Kate Brown (D) also recently identified social determinants of health as a key priority for the CCO model going forward and requested policy recommendations from the Oregon Health Policy Board to make improvements on this issue.72
Arkansas: Multipayer bundled payments
In 2012, Arkansas launched the Arkansas Health Care Payment Improvement Initiative (AHCPII), an ambitious statewide payment reform model that included most of the state’s insurance payers, including Medicaid, private insurers, and large employers such as Walmart Inc.73
The AHCPII includes a large-scale bundled payment model that pays based on a patient’s entire episode of care for a given condition rather than for each individual service. By shifting away from a system where providers are paid more for providing a greater number of services, bundled payments aim both to save money and to encourage providers to focus on the quality of care rather than the quantity of care. As of 2017, the state was paying bundles for more than a dozen different episodes of care, mostly for conditions requiring surgeries and other hospital-based care.74 When initially implementing the AHCPII, Arkansas received $42 million in federal funding support through a CMMI State Innovation Model Test grant.75
In addition to one of the country’s most ambitious bundled payment reforms, the AHCPII also overhauled primary care through a patient-centered medical home (PCMH) model, intended to provide a greater focus on the management of chronic conditions.76
Data from the state indicates that these efforts have reduced health care spending while improving some measures of health care outcomes.
In 2015, the PCMH model resulted in $54.4 million in Medicaid savings, which worked out to $35 million in net savings after factoring in care coordination payments to providers and shared savings to providers.77 Meanwhile, the episodes-of-care model has also resulted in savings. For example, the hip and knee replacement bundled payment reduced Medicaid costs for this episode by 4 percent from 2014 to 2015, while the bundled payment for chronic obstructive pulmonary disease reduced per-episode Medicaid costs by 8 percent from 2014 to 2015.78
The PCMH and episodes-of-care models have also made a positive impact on many health care outcomes and quality metrics. The state’s Medicaid data show that rates of hospitalization and emergency department utilization declined in both 2014 and 2015; in 2015 among Arkansas Medicaid beneficiaries, the rates of hospitalization and emergency department visits declined by 16.5 percent and 5.6 percent, respectively.79 The state’s bundled payment for perinatal care has reduced Arkansas’ Medicaid cesarean section rate from 39 percent to 32 percent, while also driving down average costs per episode from $3,508 to $3,413.
Several episodes saw increases in the rates of follow-up visits to ensure patient recovery, an indication of how provider behavior may be changing under the bundled payments—since provider reimbursement for complications related to a previous treatment generally come out of the original bundled payment rather than representing a new, additional payment for providers. For example, episodes for chronic obstructive pulmonary disease and asthma saw the rate of physician follow-up visits for Medicaid beneficiaries increase by 87 percent from 2014 to 2016 and by 15 percent from 2014 to 2015.80
Private payers have also reported positive results. For instance, Arkansas Blue Cross and Blue Shield saw tonsillectomy episode costs decline by 5 percent from 2014 to 2015, while the length of stay for congestive heart failure patients dropped by 17 percent from 2014 to 2015.81
However, the improvements are not quite universal. Some of the state’s quality metrics, including infant and child wellness visits, declined slightly from 2014 to 2015, which the state has flagged as an issue to monitor closely going forward.82
Finally, in addition to the state payer data, a National Bureau of Economic Research paper compared Arkansas with nearby states and found that the state’s perinatal episode of care reduced spending by 3.8 percent relative to other states.83 The researchers concluded that this reduction was primarily due to reductions in inpatient facility prices rather than changes in utilization, and the quality measures studied remained unchanged except for improvements for one measure concerning screening rates.
Other states have already taken inspiration from Arkansas’ approach. For example, Tennessee has implemented a similar bundled payment model for certain episodes of care, seeing positive results on spending and quality. From 2014 to 2015, Tennessee’s first three bundled payments—for perinatal care, asthma, and total joint replacement—generated $11.1 million in savings.84
Unfortunately, however, Arkansas recently took a major step backward by requesting and winning approval for a waiver to implement work requirements and other harmful policies in Medicaid that will make it harder for many Arkansans to keep their Medicaid coverage and access health care.85
Reducing health care costs is hard. However, innovative reforms underway in several states provide promising models for other states to adopt or adapt. Maryland’s global budgets for hospitals, Arkansas’ bundled payments, and Oregon’s coordinated care organizations have all produced savings. While Massachusetts has seen more inconsistent results, its approach still carries potential, and the state is moving forward with additional reforms.
Encouragingly, the initial results from these state efforts indicate that several different types of reforms can work. As a result, state policymakers have a variety of effective options to make progress on reducing health care costs in ways that maintain or improve the quality of care.
As some states begin to implement waivers that would reduce Medicaid eligibility, benefits, and affordability under the guise of cutting costs, the examples of these pioneering states demonstrate a better path forward: reining in health care spending growth through reforms that protect patients’ care instead of leaving them worse off.
About the author
Thomas Huelskoetter is the policy analyst for Health Policy at the Center for American Progress. Prior to joining the Center in 2013, he taught high school English in France and interned with the Center on Budget and Policy Priorities. He graduated from Kenyon College in 2012 with a bachelor’s degree in political science.
This publication was made possible in part by a grant from the Peter G. Peterson Foundation. The statements made and the views expressed are solely the responsibility of the Center for American Progress.