The Affordable Care Act of 2010 (ACA) and subsequent executive actions by the Obama administration accelerated efforts to bring value-based payment to health care. Under the banner of value-based purchasing, federal officials have touted the “triple aim”—improving the quality of care and health outcomes while bending the cost curve.
In the case of Medicaid and value-based payment, Delivery System Reform Incentive Payment (DSRIP) waivers have loomed large. When we reviewed DSRIP developments in 2015, eight states had obtained Section 1115 waivers from the Centers for Medicare and Medicaid Services (CMS) to operate DSRIP or DSRIP-like initiatives. By 2018, the number had grown to 13 with the implementation timeframes for most of these waivers extending into the early 2020s. Five DSRIP initiatives cluster in the Northeast (Massachusetts, New Hampshire, New Jersey, New York, Rhode Island) and another four in the West (Arizona, California, Oregon, Washington). Three are from the south-central region (Kansas, New Mexico, Texas) and one from the Deep South (Alabama).
DSRIP waivers provide Medicaid funds to designated organizations (primarily hospitals) that achieve performance goals related to infrastructure development, the quality of care, and health outcomes. Given the significance to Medicaid of this decade-long reform initiative, it is important to again take stock of DSRIP. Three questions seem especially pertinent. What types of organizational networks currently operate under the DSRIP rubric? What have these initiatives accomplished? What developments and challenges await DSRIP over the next few years?
Variations on the DSRIP Theme
States pursued the early DSRIP waivers largely to preserve hospital subsidies derived from Medicaid upper payment limits. These federally set limits often exceeded what states paid for services. The gap between the limits and state payment created funding pools to subsidize hospitals. When the movement toward managed care closed off this financial spigot, CMS agreed to replace these funds with new money anchored in value-based payment.
More recently, DSRIP funding has become delinked from this supplemental funding. Instead, a state negotiates the amount of DSRIP funds with CMS, presumably demonstrating that the federal investment will be offset by enough savings to meet the budget neutrality requirement of Section 1115 waivers. Beyond this, current DSRIP initiatives differ from the 2015 cluster in three particularly noteworthy ways.
Current DSRIP Initiatives Are Less Hospital-Oriented, More Focused On Broader Provider Partnerships
Seventy-five percent of the DSRIP waivers in 2015 were hospital oriented. This model assumed that performance payments to these facilities would prompt them to form collaborative relationships with community providers, but these connections were not formally specified in waiver documents. Only New York and Texas went appreciably beyond the hospital-oriented model to create regional networks of various kinds of providers to form DSRIPs.
By 2018 fewer than 40 percent of the DSRIPs (California, Kansas, New Jersey, New Mexico, and Oregon) were hospital oriented. DSRIPs in the eight other states stressed partnerships between hospitals and other kinds of providers to achieve quality, health and cost reduction goals. Such DSRIP newcomers as New Hampshire, Rhode Island, and Washington have in varying degrees embraced the importance of recruiting primary care providers, behavioral health specialists, nursing facilities and home- and community-based service providers to collaborative networks.
Current DSRIP Initiatives Place A Greater Emphasis On Behavioral Health
To a greater degree than Medicare and commercially covered populations, Medicaid enrollees have two defining characteristics that contribute to their health problems and drive up program costs through otherwise avoidable hospital use and in other ways. First, they are disproportionately disadvantaged in terms of the social determinants of health such as income, education, diet, housing and transportation. Second, they more readily suffer from mental health and substance use disorders.
Several older DSRIPs, such as New Jersey, New York, and Texas, recognized the need to kindle collaboration between hospitals and behavioral health specialists. In New York, each performing provider system implemented at least one clinical improvement program focused on behavioral health. In other states, however, a provider-driven, highly decentralized approach to project selection allowed hospitals to pick from a long menu of pay-for-performance foci and diluted the focus on behavioral health. In New Jersey, few hospitals selected behavioral health even though it was on the menu of options.
Among the current cohort of DSRIP initiatives, three (Arizona, New Hampshire, Rhode Island) give priority to enhanced services for those with mental health and substance use disorders. Arizona’s DSRIP directs lump sum payments to Medicaid managed care organizations, which then offer financial incentives to physical and behavioral health providers to integrate their services, including for enrollees transitioning from criminal justice facilities. New Hampshire’s DSRIP establishes regional Integrated Delivery Networks that target individuals at risk for, or already diagnosed, with mental health and substance use disorders. These networks must include behavioral health specialists and community-based organizations.
The Rhode Island DSRIP seeks to galvanize the development of two types of Medicaid Accountable Care Organizations (ACOs)—comprehensive and specialized—both of which seek to enhance care coordination for those with behavioral health problems. Comprehensive entities will be responsible for services delivered to the general Medicaid population, while specialized ACOs will target beneficiaries with particular (often severe) substance use and related behavioral health disorders.
While not as exclusively focused on behavioral health, Massachusetts and Washington have also evinced appreciable commitment to these services. In Massachusetts, 30 percent of DSRIP funds go to Medicaid ACOs and managed care organizations to bolster services for enrollees with complex behavioral health conditions, many of whom also need long-term services and supports. Washington’s DSRIP lists the integration of physical and behavioral health purchasing as one of five objectives.
Current DSRIP Waivers Seek To Spread Value-Based Purchasing To A Larger Portion Of Medicaid Services
The diffusion of alternative payment models has become a DSRIP performance metric in its own right. All DSRIP states rely on commercial managed care organizations (MCOs) to serve Medicaid enrollees. Managed care penetration ranged from an estimated 61 percent in Alabama to a high of 92 percent in New Jersey and Oregon. The diffusion of value-based purchasing depends on whether state Medicaid officials can implant the practice in their contracts with MCOs and whether these organizations effectively implement the value-based payment agreements.
In their negotiations with CMS, at least three states (California, New York, Washington) have pledged to meet specific managed care targets for alternative payment models or face penalties for failing to do so. New York must move 80 to 90 percent of managed care payments to value-based models to avoid reduced DSRIP funding (only 30 percent were value-based at the start of the waiver’s fourth year). Other states have less stringent DSRIP provisions to promote the adoption of alternative payment models by MCOs. For instance, New Hampshire officials have agreed to collaborate with stakeholders to develop a roadmap for moving 50 percent of Medicaid managed care payments to value-based models.
Massachusetts and Rhode Island have also sought to foster value-based payment by intertwining DSRIP with the creation of Medicaid ACOs. Massachusetts requires all DSRIP providers to participate in one of three types of ACOs, two of which feature ACOs contracting with MCOs. So, too, Rhode Island has established integrated provider organizations called Accountable Entities; these organizations enter into agreements with Medicaid MCOs to improve care quality and patient experience while generating program savings.
DSRIP Accomplishments
To date, little evidence exists that DSRIP waivers have significantly improved quality and health outcomes or reduced spending on health care services. A recent comprehensive report found that DSRIP waivers have forged an infrastructure for collecting data about provider performance, “but it is too early to present significant findings related to outcomes.”
Another report, focused on DSRIP waivers in California, New Jersey and Texas, reached a similar conclusion. This report examined the impact of these waivers on the use of hospital emergency department use and hospitalizations for ambulatory-care sensitive conditions. The evidence for all three states was “mixed” and it was difficult to attribute any improvements to DSRIP.
At least two factors complicate the assessment of DSRIP’s impact. First, these waivers are not designed as controlled research studies, so it is hard to disentangle their effects on outcomes from other factors; this is particularly true because the states have generally launched multiple initiatives to reduce hospital use. Second, the service integration that DSRIP waivers seek involves overcoming silos that have separated hospital and community-based health and social services for decades.
Even in states that have adopted partnership models, challenges associated with distributing DSRIP funds to community-based organizations have surfaced. One difficulty has been isolating the contributions of each partner to the overall effort in order to reward them. Another more significant barrier has been onerous reporting requirements associated with DSRIP payment. Many community-based social service providers tend to rely on grants and do not have the experience or the expertise to meet the reporting requirements necessary to receive payments from hospitals or the state.
Despite these limitations, the DSRIP waivers have yielded gains. They have helped stoke state progress in creating an infrastructure for Medicaid value-based payment. DSRIP has increased the amount of outcome data routinely collected from providers and generally enhanced state and hospital capacity to track the quality of services Medicaid enrollees receive. The program has also encouraged the diffusion of value-based payments to Medicaid MCOs. These developments may well be an important longer-term legacy of the DSRIP waivers.
Whither DSRIP?
The Medicaid and CHIP Payment and Access Commission recently asserted that CMS “views DSRIP funding as a one-time investment, and does not plan to renew DSRIP demonstrations.” Instead, CMS has prioritized the diffusion of value-based purchasing to Medicaid managed care. It remains to be seen whether the DSRIP waivers extending into the 2020s will in fact be phased out. Much will depend on executive branch (including presidential) priorities at the time and whether DSRIP states strongly lobby to preserve their waivers. The results of the current cohort of DSRIP waivers suggest that renewing them in some form may buttress rather than impede efforts to promote other value-based approaches, such as Medicaid ACOs or managed care provider payments more fully rooted in value-based payment.
Independent of whether DSRIP waivers ultimately get renewed, they are a significant part of the Medicaid reform story over the next four years and deserve research attention. A comparative analysis of not just the outputs and outcomes of the existing waivers but their implementation dynamics (getting inside the “black box”) could well yield valuable lessons. Consider, for instance, the selection of performance metrics and the standards applied to them. If mental health and substance use disorders are a major focus of several DSRIPs, at least some quality measures should focus specifically on these domains.
A comparative analysis of Medicaid ACOs in four states suggests, however, that it will not be easy to introduce these measures. In all four states, officials pressed to have quality metrics that at least partly gauged mental health and substance use disorder services. Health care providers, however, strongly and effectively resisted the use of such metrics. This was less because they worried about being held responsible for social and individual factors beyond their influence. Instead, resistance more readily sprang from a strong sense among providers that they were subject to performance metric overload that imposed substantial reporting costs on them.
Independent analyses suggest that provider concerns with overload are far from baseless. They underscore the substantial reporting costs of value-based payment systems and how the proliferation of metrics and “quality measure misalignment” have placed a substantial “administrative burden” on providers, at times exposing them to conflicting performance incentives. Providers thus prefer that diverse payers employ the same metrics. Serious questions persist, however, as to whether a one-size-fits-all approach would adequately motivate DSRIP providers to meet the behavioral health needs of Medicaid enrollees. The processes leading to the selection of performance metrics in the various DSRIP initiatives and the likely incentive effects of these metrics are among myriad implementation details that deserve assessment.