Original Article
Downloaded by University of Connecticut e-journal package NERL from online.liebertpub.com at 09/04/17. For personal use only.
POPULATION HEALTH MANAGEMENT Volume 00, Number 00, 2017 ª Mary Ann Liebert, Inc. DOI: 10.1089/pop.2017.0061
The 10 Conditions That Increased Vermont’s Readiness to Implement Statewide Health System Transformation David Grembowski, PhD,1 and Miriam Marcus-Smith, MHA2
Abstract
Following an arduous, 6-year policy-making process, Vermont is the first state implementing a unified, statewide all-payer integrated delivery system with value-based payment, along with aligned medical and social service reforms, for almost all residents and providers in a state. Commercial, Medicare, and Medicaid valuebased payment for most Vermonters will be administered through a new statewide accountable care organization in 2018–2022. The purpose of this article is to describe the 10 conditions that increased Vermont’s readiness to implement statewide system transformation. The authors reviewed documents, conducted internet searches of public information, interviewed key informants annually in 2014–2016, cross-validated factual and narrative interpretation, and performed content analyses to derive conditions that increased readiness and their implications for policy and practice. Four social conditions (leadership champions; a common vision; collaborative culture; social capital and collective efficacy) and 6 support conditions (money; statewide data; legal infrastructure; federal policy promoting payment reform; delivery system transformation aligned with payment reform; personnel skilled in system reform) increased Vermont’s readiness for system transformation. Vermont’s experience indicates that increasing statewide readiness for reform is slow, incremental, and exhausting to overcome the sheer inertia of large fee-based systems. The new payments may work because statewide, uniform population-based payment will affect the health care of almost all Vermonters, creating statewide, uniform provider incentives to reduce volume and making the current fee-based system less viable. The conditions for readiness and statewide system transformation may be more likely in states with regulated markets, like Vermont, than in states with highly competitive markets. Keywords: value-based payment, accountable care organization, readiness, system transformation, Vermont, State Innovation Models
V
ermont is implementing a statewide, all-payer integrated delivery system that was developed incrementally during a 6-year (2011–2016) policy-making process. Vermont’s reforms have significant policy implications because Vermont is the first state to implement a unified (or ‘‘standardized’’) statewide all-payer system with value-based payments, along with aligned medical and social services, for almost all residents and providers in a state. Vermont’s historic, statewide system transformation is made possible by federal regulations and policies. In October 2016, the State of Vermont signed the formal agreement with the Centers for Medicare & Medicaid Services (CMS) to waive federal regulations and allow other forms of Medicare and Medicaid payment.1 Commercial, Medicare, 1 2
and Medicaid value-based payment for the entire state will be administered through a new statewide accountable care organization (ACO), or Vermont All-Payer ACO Model (VAPM). The Model began on January 1, 2017, and ends on December 31, 2022, a 6-year life; the first year (2017) is devoted to planning and creating the VAPM infrastructure. By the end of Year 3, the VAPM must have a plan for the financing and delivery of Medicaid behavioral health and Medicaid home and community-based services. CMS favors spreading similar statewide payment and care delivery reforms across all major payers in other states.2 Table 1 presents the key features of Vermont’s statewide integrated delivery system. In brief, VAPM has multiple sources of revenue, mainly Medicaid, Medicare, and the health
Department of Health Services, School of Public Health, University of Washington, Seattle, Washington. Division of General Internal Medicine, School of Medicine, University of Washington, Seattle, Washington.
1
2
GREMBOWSKI AND MARCUS-SMITH
Downloaded by University of Connecticut e-journal package NERL from online.liebertpub.com at 09/04/17. For personal use only.
Table 1. Features of Vermont’s Statewide All-Payer Integrated Delivery System Feature
Description
Payers
About 300,000 Vermonters are covered by commercial health insurance. BlueCross BlueShield of Vermont has 80% of the commercial market. Medicaid and other subsidized health care covers about a third of Vermonters. Vermont has 3 ACOs, currently with more than 160,000 total patients attributable to Medicare, Medicaid, and commercial plans: OneCare Vermont: formed in 2012 by University of Vermont Medical Center and Dartmouth-Hitchcock Medical Center, consisting of all Vermont hospitals and their employed physicians, some FQHCs and independent medical offices, as well as mental health and substance abuse agencies, and skilled nursing facilities (about 100,000 patients). CHAC: formed in 2014 and consisting of Vermont’s 11 FQHCs, along with mental health and substance abuse agencies and skilled nursing facilities (about 57,000 patients). Vermont Collaborative Physicians/Accountable Care Coalition of the Green Mountains: formed in 2012 and consisting of independent primary care and specialty practices (about 9000 patients). Medicare, Medicaid, and commercial health plans pay ACOs all-inclusive population-based payments, which are similar to capitation payments, for their attributed beneficiaries. The ACOs distribute payments to health care organizations for their attributed patients. Using 2017 as a baseline, Vermont is expected to limit per capita health care expenditure growth for all major payers to 3.5% in the 5-year demonstration. Vermont also is expected to limit Medicare per capita health care expenditure growth for Medicare beneficiaries to at least 0.001–0.002 percentage points below the projected national average. Medicaid does not have a cost growth target. Payment reform is built on a primary care foundation, Blueprint for Health (Blueprint), the statewide, multipayer program that meets National Committee for Quality Assurance medical home standards, providing primary care to three fourths of Vermonters. Blueprint provider teams link patients with social and other services to address patient needs. Vermont is divided into regional Health Service Areas, each with its own hospital. Regional collaboratives are coordinating medical and social services to address the needs of the whole person in each region. Representatives from the local hospital, ACOs, Blueprint, FQHCs, behavioral health/community mental health centers, home care agencies, aging agencies, skilled nursing, housing agencies, and other health and social service agencies meet regularly, with the common goal of integrating services across agencies to address local population needs, particularly for people with mental health, substance abuse, multiple chronic conditions, and social problems. Care coordination programs also are in place to reduce fragmentation and service duplication for persons with multiple chronic conditions. State-level and ACO-level performance targets are imposed to create incentives for collaboration between the care delivery and public health systems, which in turn lead to system reforms that achieve the targets. Three categories of targets exist: (1) process milestones; (2) health delivery system; and (3) population-level health outcomes. In each category, performance targets are set for access to primary care, substance use disorder, suicides, and chronic conditions. Statewide information systems are in place or in development: All-payer claims database for Medicare, Medicaid, and 90% of commercial insurance Vermont Health Information Exchange (VHIE), a statewide clinical information system providing real-time data to support clinical care. VHIE collects and warehouses data from electronic medical records for all Vermont’s hospitals, most FQHCs, and Blueprint primary care offices. VHIE also delivers clinical information, such as laboratory reports and patient registries for chronic diseases, to providers, who can query patient records for consenting patients. The Green Mountain Care Board, an independent, 5-member body created by Act 48 in 2011, has broad regulatory authority over Vermont’s health care system.
ACOs
Value-Based Payment & Financial Incentives Cost Growth Targets
Primary Care
Delivery System Transformation
Performance Targets for Quality of Care and Health Outcomes
Information System
Vermont All-Payer ACO Model Regulation
ACOs, Accountable Care Organizations; CHAC, Community Health Accountable Care; FQHCs, Federally Qualified Health Centers.
plans in the Vermont exchange. Vermonters retain their multiple public and private health plans but receive services through a single, simplified and standardized administrative and value-based payment structure administered by a new, statewide ACO, the Vermont Care Organization (VCO), reg-
ulated by the Green Mountain Care Board (GMCB). Participation is voluntary in the provider-led VAPM. Payment will be based on the CMS Next Generation ACO Model. Although the payment details are in development, in general, the VCO will receive all-inclusive population-based payments, which
Downloaded by University of Connecticut e-journal package NERL from online.liebertpub.com at 09/04/17. For personal use only.
VERMONT’S READINESS FOR SYSTEM TRANSFORMATION
are similar to capitation payments, for their attributed beneficiaries in Medicare, Medicaid, and commercial health plans. The VCO, in turn, will distribute payments to Vermont’s 14 hospitals and Federally Qualified Health Centers, which employ more than two thirds of Vermont primary care physicians, and other health care organizations. The health plans will have common measures of performance, and each plan will retain its own benefits. Payment reform is aligned with delivery transformation to address the needs of the whole person. Reaching scale is critical for the VAPM’s success. To have a viable statewide payment model, the VCO must enroll a substantial majority of Vermonters and primary care providers. By the end of 2022, the VCO must enroll 70% of all Vermonters, 90% of Medicare beneficiaries, and 80% of primary care providers.1 Table 1 indicates that about half of Vermont’s primary care physicians participate in Vermont’s 2 ACOs, OneCare Vermont (30%) and Community Health (20%). Vermont is more likely to achieve the 80% provider target if both ACOs participate in the VAPM, which, in turn, also would increase the enrollment of Vermonters in the VAPM. To incentivize provider participation, Vermont Medicare providers are eligible to participate in the Medicare Access and CHIP Reauthorization Act’s (MACRA) advanced alternative payment model and receive a 5% bonus for their Medicare patients. A key assumption of the VAPM is that by reaching scale, the statewide population-based payments will affect most patients and providers in most medical organizations, creating common, statewide payment incentives to reduce volume while also diminishing the viability and influence of the current fee-for-service (FFS) payment in the delivery of health care. The population-based payments for most Vermonters may generate revenue streams for providers that are more predictable and dependable over time than FFS, reducing financial uncertainty and promoting sustainability. The purpose of this article is to describe the 10 conditions that increased Vermont’s readiness to launch its statewide integrated delivery system with all-payer, population-based payments. Few studies have examined the development of system readiness.3 Diffusion theory posits that innovations, such as Vermont’s VAPM, are not adopted all at once but rather are spread over time across health care organizations through a predictable process with 3 distinct stages: readiness, adoption, and implementation.3,4 System readiness is a preadoption stage during which payers, health plans, health care organizations, and providers become aware of payment and delivery reform models and learn how the models work and about their potential consequences. Based on this knowledge, health care organizations may abandon payment reform or move toward a state of readiness for reform with buy-in from most stakeholders. Greater readiness increases the likelihood that health care organizations will adopt, implement, and sustain payment reform and system tranformation. Other states may apply these findings to increase their readiness for statewide health system reform. Methods
This study of Vermont was part of a larger qualitative comparative case study of 11 value-based payment projects being carried out in 8 states funded by the Robert Wood Johnson Foundation (RWJF) and evaluated by University of
3
Washington researchers (the other states were California, Maine, Massachusetts, New Hampshire, Oregon, Pennsylvania, and Washington).3–5 The study team applied guidelines from organizational sociology and qualitative methods for conducting comparative case studies.6–10 In keeping with the principles of qualitative research, the approach to data collection and analysis was inductive, structured around a set of common questions, and allowed for open-ended responses and follow-up probes. As such, this paper neither develops a conceptual model, nor tests explicit hypotheses, but seeks to articulate lessons learned and implications for statewide health system transformation from qualitative analysis. Study protocols were reviewed and approved by the University of Washington’s Institutional Review Board. The qualitative methodology followed 6 steps. First, the evaluation team developed a semi-structured interview instrument with common questions for all sites covering context, stakeholders, objectives, approach, logic model, progress and results, facilitators and barriers, evaluation and sustainability, and lessons learned (see questionnaire Supplementary Appendix S1; Supplementary Data are available online at www.liebertpub.com/pop). The study team obtained each site’s grant application, documents from each project’s public website, and other local, publicly available information about each project. Based on this information, the team developed additional questions unique to each site. The team made a list of key informants at each site based on the site’s RWJF application records and site documents. Common data collection and analysis protocols were followed across sites. The study team then conducted 60-minute interviews with up to 15 key informants in Vermont each year for 3 years (2014–2016). Almost all interviews were in person; the remainder were by phone. In 2014 and 2015 interviews were conducted by the first and second authors, with the second author taking real-time interview notes on a laptop computer and audio-recording each interview; in 2016 interviews were conducted only by the first author, audio-recorded and transcribed, because of changes in funding and study staff. All key informants consented verbally to be interviewed and recorded at the beginning of each interview. Prior to the interview, key informants received the interview questions and information about how the interview information would be used for analysis and publication. Third, qualitative analysis was performed for each year of interviews following accepted protocols.10 Before analyzing interview data, the authors reviewed the interview notes, newspaper articles about the project, and background documents from public websites. Key informants’ responses to each interview question and site documents were compared to derive answers to each question, facilitators and barriers to implementation, and implications for policy and practice. Next, to document the answers to the interview questions, the study team produced an annual site report for all sites. For Vermont’s annual site report, the first author prepared the first draft of the site report, which then was reviewed by the second author. The 2014 and 2015 site reports also were reviewed by at least 1 other team member. Each annual report was then vetted with the lead contact for Vermont’s RWJF project, who was asked to identify any apparent errors of omission or commission and to specify any inadvertent breaches of confidentiality or revelations of proprietary or
Downloaded by University of Connecticut e-journal package NERL from online.liebertpub.com at 09/04/17. For personal use only.
4
competitively sensitive information. The team reviewed the site’s critique and revised the documents based on team consensus. Finally, the study team compared Vermont’s 3 annual site reports to derive persistent conditions promoting or inhibiting readiness for statewide transformation of Vermont’s health care system (see Supplementary Appendix S2, for list of persistent facilitators and barriers to implementation in the annual reports). In selected years, the annual reports for all sites were reviewed by the study’s Advisory Group, composed of 8 national experts in payment reform. The advisors suggested themes and insights relevant to payment reform. Their comments informed the insights in this paper. Results
The following 10 conditions are essential ingredients for increasing readiness for payment and system transformation in the face of day-to-day challenges. The conditions are interconnected and operate jointly to increase readiness. The conditions are divided into 2 categories: social and support. Social conditions: pay attention to the ‘‘Cs’’
(1) Champions Moving from FFS to value-based payment throughout the state is a journey into the unknown. Committed champions play a vital role in guiding deliberations toward decisions about what the new system will look like. In particular, the governor and the GMCB, through its role as facilitator and convener of diverse stakeholders to participate voluntarily in payment and system reform without antitrust threats, were instrumental in sustaining momentum toward payment and system reform. From 2011–2016 there was little turnover in the champions, which sustained momentum. (2) Common vision of the future payment reform A common vision of the future statewide payment and delivery system is essential for building readiness and getting stakeholders ‘‘on the same page.’’ Two competing visions of system reform were advanced: (1) initially, a single-payer universal health insurance program; or (2) an all-payer integrated delivery system. On December 17, 2014, Governor Shumlin ‘‘pulled the plug’’ on the singlepayer program, mainly because of its enormous costs that Vermont taxpayers could not afford, and attention shifted to the all-payer model.11,12 (3) Culture of collaboration and cooperation (not competition), communication, compromise and consensus-building, and civic engagement in a small, rural state Vermont is a small, rural state with about 625,000 residents and little market competition, which promotes communication and allows everyone to ‘‘be at the table’’ and adopt a statewide perspective to health system reform. As one key informant noted, ‘‘If you’re not at the table, you may be on the menu.’’ Vermont also has a communitarian culture and a history of collaboration and civic engagement among stakeholders, who are willing to cooperate and compromise for the public good. In a state where ‘‘everyone knows each other,’’ many stakeholders have worked together on health care issues for many years and have a collective knowledge of Vermont’s
GREMBOWSKI AND MARCUS-SMITH
history of health care, which facilitates communication and building consensus. Payment reform and system transformation engaged literally hundreds of Vermonters. Overall direction was provided by an 8-member Core Team, who received guidance from a 37-member Steering Committee. The Team and Committee, in turn, received guidance from at least 6 workgroups addressing payment models, quality and performance measurement, care management, and other topics. (4) Social capital and collective efficacy Social capital and collective efficacy are key ingredients of system readiness and prerequisite to achieve system reform.13–16
Social capital may be defined as a characteristic of groups, organizations, or communities – such as trust in others, norms of reciprocity, and civic participation – that can facilitate collaboration and cooperation, including the flow of information, for mutual benefit. Collective efficacy, which is related to social capital, refers to a group’s shared beliefs in their collective ability to produce desired results. The linkage of mutual trust and the shared willingness to intervene for the common good define a group’s capacity to achieve common goals. Bandura notes that the higher the perceived collective efficacy, the stronger the group’s staying power in the face of impediments and setbacks, and the greater their accomplishments.13 The 2 concepts suggest that the greater the stock of social capital and collective efficacy, the greater the likelihood of implementing statewide payment reforms. Statewide transformation from FFS to value-based payment is disruptive change that can generate uncertainty, fear, and distrust in all sectors of Vermont’s health care system. Building trust and social capital, a sense that ‘‘we work better together,’’ is time-consuming yet essential to working through the fear and uncertainty, and to sustaining willingness to collaborate and reach consensus about what payment reform will look like. Information is a part of social capital – credible evidence that the new payment model will ‘‘do no harm’’ to the financial health of providers may reduce uncertainty and improve stakeholders’ comfort levels with the payment reforms. Collective efficacy is a resource for staying on track. Over time, a ‘‘coalition of the willing’’ emerged: well-known leaders who are respected by most interest groups and have a collective sense that ‘‘we can do this.’’17 Collective efficacy is related to social capital, in that trust is the glue that holds the coalition together and fosters collaboration. Some stakeholders noted that the State Innovation Model (SIM) grant provided the resources and time to invest in trust and build the coalition of the willing. Support conditions
(5) Money Significant monetary resources are required to build readiness for statewide system transformation, particularly in a resource-scarce state. The GMCB received a grant from RWJF in 2012 to design and implement payment reform strategies. Development work under the grant eventually led to a 3-year, $45 million SIM award from the CMS’ Center
Downloaded by University of Connecticut e-journal package NERL from online.liebertpub.com at 09/04/17. For personal use only.
VERMONT’S READINESS FOR SYSTEM TRANSFORMATION
for Medicare & Medicaid Innovation to the Office of the Governor in 2013. SIM funding was well aligned with Vermont’s goals: the SIM Initiative is intended explicitly to create statewide system transformation for most care in a state. However, time remained a scarce resource. Readiness preparations were a burden for stakeholders when the numerous, ongoing, day-to-day demands for meetings and system design work crowded out time for their regular jobs. (6) Statewide data A statewide data infrastructure is required for planning the new statewide system, coordinating care to address local patient and population needs across diverse health care and social service organizations, and the monitoring and public reporting of performance for management and accountability. Data flow also contributes to social capital: looking at data together is another way to build trust among stakeholders. Most stakeholders endorsed a single, transparent, universal set of performance measures for all payers. The VAPM has its own performance measures, which over time may become the universal measures for all payers in the state. Simply collecting data is insufficient; data must be analyzed to create ‘‘actionable’’ results that inform clinical care and how to change the system to reduce costs and improve quality. GMCB oversees Vermont’s all-payer claims database, which covers 90% of commercially insured Vermonters and all Medicaid and Medicare enrollees, although Medicare data are available only after a long time lag. Vermont Information Technology Leaders, with oversight from the GMCB, is building the Vermont Health Information Exchange, which contains clinical patient records for all hospitals and most primary care offices with real-time data access to support coordination of clinical care. Vermont’s statewide data infrastructure is being used to support system transformation and accountability. The regional collaboratives are using the data infrastructure to support patient care and, at the population level, identify local needs and align services with those needs. The GMCB has developed a ‘‘dashboard’’ that profiles Vermont’s health care system and comprises 51 medical and nonmedical indicators. In the 14 regions, insurers, Blueprint for Health (Blueprint: Vermont’s statewide primary care program that meets National Committee for Quality Assurance medical home standards and has community health teams to link patients with social and other services to meet the needs of the whole person), ACO leaders, and other groups are coproducing customized performance reports for quality, cost, utilization, and other measures. Vermont’s data infrastructure has limitations. The US Supreme Court ruling in Gobeille (in this individual’s official capacity as Chair of GMCB) vs. Liberty Mutual Insurance Company essentially prohibits the state from collecting claims data from self-insured entities.18 In addition, linked claims– clinical data do not exist for most Vermont residents. Vermont also lacks regional health assessment surveys. (7) Legal infrastructure A legal infrastructure is required for payment reform. Most federal, state, and local laws and regulations support FFS reimbursement but may not offer a legal foundation for an all-payer model with population-based payment. Legislation may reduce uncertainty by defining what will happen rather than what might happen in payment reform.
5
On the heels of the federal Affordable Care Act, in 2011 the Vermont Legislature, with strong leadership from Governor Shumlin, passed Act 48, promising a ‘‘universal and unified health system’’ as a ‘‘public good’’ that also controls costs and offers quality care. The Legislature’s intention was to move Vermont toward a more highly regulated and integrated health care system but specified few details about what the new system might look like or how it would work. In June 2015, the Legislature passed Act 54 (Section 139) to explore an ‘‘all-payer model, which may be achieved through a waiver’’ from CMS. To prepare for an all-payer model, in May 2016, the Legislature passed Act 113 (H.812), which authorized the GMCB and Vermont’s Agency of Administration (or designee) to enter an agreement with CMS to waive Medicare regulations if certain criteria are met. The Act requires that all ACOs must have GMCB certification to receive payments from Medicaid or commercial insurance, and also mandates that Medicaid payments be integrated into the all-payer model. (8) Federal policy promoting payment reform Federal policies are primarily facilitators but can be barriers to statewide payment and delivery system reforms. Beginning in 2013 Vermont’s 3 ACOs participated in a statewide all-payer (commercial, Medicare, and Medicaid) shared savings program, which was based on the CMS– ACO shared savings model and served as a stepping stone to the VAPM. Vermont’s initial work on the all-payer model was followed in 2015 by 3 federal policies to move Medicare payment away from FFS to value-based payment: 1) Secretary Burwell’s announcement of tying 50% of Medicare payments to alternative payment models by the end of 2018; 2) the CMS Next Generation ACO Model, which allowed ACOs to assume higher levels of financial risk and reward (in 2015 OneCare Vermont, an ACO, accepted CMS’ invitation to enroll in the Next Generation ACO Model); and 3) Congress’ MACRA, which pays physicians extra for participating in alternative payment models and increases financial incentives in pay-for-performance/ quality reporting programs. The new federal policies legitimized and added credibility to Vermont’s all-payer model, and reinforced stakeholder views that payment reform is inevitable, creating local and statewide momentum to collaborate on value-based payment reform. However, the federal regulations and the burdensome process for obtaining the waivers generally inhibit innovation and increase the workload and time for system change. (9) Statewide delivery system transformation aligned with payment reform In the US health care system, patients with chronic conditions, disabilities, and social problems (eg, substance abuse) often receive fragmented, uncoordinated, and inefficient services that focus on siloed conditions rather than adopting a patient-centered approach that addresses the diverse needs of the whole person, which can reduce quality of care and increase costs. Low connectivity across health care organizations promotes fragmented care in the system.
Downloaded by University of Connecticut e-journal package NERL from online.liebertpub.com at 09/04/17. For personal use only.
6
To address this problem, for the past decade Blueprint’s community health teams have linked primary care patients with social and other services. Fostered by the SIM grant, each region has formed a collaborative comprising diverse health and social service agencies, which meets regularly to integrate services to address local population needs, particularly the needs of people with mental health, substance abuse, and other social problems. Collaboratives are expected to adopt a data-driven process to prioritize local needs and align services with those needs on a population level, and develop guidelines to identify persons with high service needs and how to address them in ‘‘medical neighborhoods.’’19 As part of this activity, the ACOs and Blueprint are working to coordinate their activities in all of the state’s regions. In a majority of regions, Integrated Communities Care Management Learning Collaboratives, along with a Medicaid chronic care initiative, also are advancing care coordination for people with complex needs through social and health service integration. Collaboration has challenges. Clinical providers faced a cultural switch of moving from a traditional medical model, with the physician in charge, toward team medicine responsible for managing individual and population health. Aligning delivery system transformation and payment reform was challenged by health care organizations and providers with little experience bearing financial risk and viewing a service as an expense (in population-based payment) rather than revenue (in patient-based FFS payment). (10) Personnel skilled in system reform Personnel and consultants with skills and experience are required to increase readiness. Transforming a health care system from FFS to value-based payment is complex work. Personnel originally hired to manage the current FFS system may be unprepared to innovate and engage in system transformation. Hiring new personnel and consultants with skills that align with payment reform goals may contribute to readiness. In particular, stakeholders recommend having a professional data manager(s) wellversed in public and commercial claims and clinical data from the beginning. Discussion
Vermont’s experience suggests that readiness is an essential building block in implementing statewide system transformation. The 2011–2016 process for statewide readiness was slow, incremental, and exhausting, but nonetheless successful in launching Vermont’s All-Payer Model. The study team posits that the conditions increase the likelihood of greater system readiness and overcoming the sheer inertia of large systems. However, the conditions do not guarantee readiness. For instance, New Hampshire abandoned the goal of global payments when the largest payer (covering about 50% of privately insured lives) declined to go forward with the common payment model.3 In particular, the study team hypothesizes that statewide system readiness may be more likely if a state has champions for changing the whole system of the entire state to align payment and delivery reforms,20 has a history of collaboration, and if the likelihood of building social capital and collective efficacy for payment reform are high.21 Re-
GREMBOWSKI AND MARCUS-SMITH
placing volume-based FFS carries great uncertainty and fear for payers, health care organizations, and other stakeholders, mainly because the potential financial consequences of doing so are huge. Distrust, conflict, and emotions can erupt in payment discussions, particularly among organizations that view each other as competitors, mainly over how much upside gain and downside financial risk the new payment model will have, and how revenue will be distributed across health care organizations and providers. Collaboration and trust building are time-consuming yet prerequisite to working through the inevitable conflicts among stakeholders, sustaining the willingness to weather the ups and downs, and ultimately reaching consensus. Competition is a key approach for improving value through lower prices and higher quality.22 However, if the hypothesis is correct, statewide system readiness may be less likely in states with highly competitive markets, but more likely in states with government regulation, because trust in others, cooperation and reciprocity, and collective efficacy for the public good may be lower in highly competitive markets. Vermont has a long history of stateregulated health care and belief in the beneficial role of state government. More than 20 years ago Vermont implemented regulations limiting rate variations and preventing health plans from denying coverage based on preexisting medical conditions. The GMCB has authority to control costs, such as setting payment rates for health professionals and approving hospital budgets, certificates of need, and payer premiums. In 2014 net patient revenues in Vermont’s hospitals increased 2.7%, which was much lower than the 5.1% to 9.2% annual increases in 2002 to 2013, reflecting GMCB enforcement of target growth rates for hospitals in 2013 and 2014.23 Vermont’s Medicare expenditures per capita are among the lowest across states. Grogan notes that the tension between competition versus cooperation is the most persistent contradiction in US health policy over the past several decades and a central feature of ACOs.24 Vermont’s all-payer model may not have happened without government support. CMS’s rollout of the Next Generation ACO payment model was timely and appealing to Vermont stakeholders, partly because many believed the model would spread nationwide in the future, and because Vermont’s governor and legislators were also on board. Adopting the federal model also was much easier than creating their own payment model from scratch. In general, payers want to reduce spending and providers want to increase revenue, and it may be impossible for them, working on their own, to agree and create a payment model that satisfies both interests. After 6 years most, but not all, stakeholders in Vermont supported the VAPM. Some of Vermont’s small health care organizations, such as independent physicians, have not endorsed the VAPM, questioning the finances of the allpayer model, how the model would work, and whether the large health care organizations would dominate over the small ones in the VAPM.17,25 The Vermont Legislature Joint Fiscal Office also identified the potential benefits, risks, and outstanding questions of the VAPM that remain to be answered.26,27 From 2011–2015 there was little turnover in the champions of the all-payer model. However, in 2016 some
Downloaded by University of Connecticut e-journal package NERL from online.liebertpub.com at 09/04/17. For personal use only.
VERMONT’S READINESS FOR SYSTEM TRANSFORMATION
leaders left their positions, and in 2017 Vermont has a new governor, a new speaker of the House, and new president of the Senate. Turnover may signal a loss of intelligence about local payment reform or a loss of leadership that may undermine implementation of the all-payer model. Alternatively, turnover may introduce new expertise and ideas into payment reform work and ultimately improve the reform’s chances of success. In closing, payment and health system reform has been an ongoing, seemingly never-ending yet successful process culminating in the launch of Vermont’s statewide all-payer model in 2018. Two basic questions remain. First, will the all-payer model be implemented as planned? Although the VCO is operational and has a board of directors, implementation has stalled because the VCO has not yet received the government funds that it expected for start-up and Medicaid programs and other factors.28 The second basic question is: Will it work? Will patients and providers participate in the new system? Or will patients take their health care and revenue to New Hampshire or New York? Or will the system work so well that people travel or move to Vermont for their health care? Is Vermont too small a state, or just the right size, to carry out innovative health care reform? Can Vermont’s statewide system reforms work elsewhere in the United States, given CMS’ intent to spread state-driven system transformation? Vermont may be unique, as Governor Shumlin notes:29 Only in Vermont would you find a health care community coming together from all walks, all sectors, saying together, ‘We think we have a new way to improve quality and bend the cost curve in a way that will make Vermont (health care) more affordable.’ Answers to these questions are unknown but are of the utmost importance for health policy in Vermont and for whether the policy is generalizable to the United States. Author Disclosure Statement
The authors declare that there are no conflicts of interest. The authors received the following financial support for the research, authorship, and/or publication of this article: Support for this research was provided by the Robert Wood Johnson Foundation (Grant Nos. 71209 and 73446). References
1. Green Mountain Care Board. Vermont All-Payer Accountable Care Organization Model Agreement. http://gmcboard .vermont.gov/payment-reform/APM Accessed May 28, 2017. 2. Centers for Medicare & Medicaid Services. Vermont AllPayer ACO Model. 2017. https://innovation.cms.gov/ initiatives/vermont-all-payer-aco-model Accessed October 27, 2016. 3. Conrad DA, Vaughn M, Grembowski D, Marcus-Smith M. Implementing value-based payment reform: a conceptual framework and case examples. Med Care Res Rev 2016;73: 437–457. 4. Conrad D, Grembowski D, Gibbons C, et al. A report on eight early-stage state and regional projects testing value-based payment. Health Aff ( Millwood) 2013;32: 998–1006.
7
5. Conrad DA, Grembowski D, Hernandez SE, Lau B, Marcus-Smith M. Emerging lessons from regional and state innovation in value-based payment reform: striking a balance between collaboration and disruptive innovation. Milbank Q 2014;92:568–623. 6. Ayres L, Kavanaugh K, Knafl KA. Within-case and acrosscase approaches to qualitative data analysis. Qual Health Res 2003;13:871–883. 7. Firestone WA, Herriott RE. Multisite qualitative policy research: some designs and implementation issues. In: Fetterman DM, ed. Ethnography in educational evaluation. Beverly Hills, CA: Sage Publications, 1984:63–88. 8. Pettigrew AM. Longitudinal field research on change: theory and practice. Organ Sci 1990;1:267–292. 9. Stake RE. Multiple case study analysis. New York: The Guilford Press, 2006. 10. Dey I. Qualitative data analysis. A user-friendly guide for social scientists. London: Routledge, 1993. 11. Goodnough A. In Vermont, frustrations mount over Affordable Care Act. June 4, 2015. https://www.nytimes.com/2015/ 06/05/us/in-vermont-frustrations-mount-over-affordable-careact.html Accessed July 13, 2017. 12. True M. Big expectations, few results on health care. May 20, 2015. https://vtdigger.org/2015/05/19/legislative-wrapbig-expectations-few-results-on-health-care Accessed June 1, 2015. 13. Bandura A. Exercise of human agency through collective efficacy. Curr Dir Psychol Sci 2000;9:75–78. 14. Sampson RJ. Neighborhoods and violent crime: a multilevel study of collective efficacy. Science 1997;277: 918–924. 15. Sampson RJ. How do communities undergird or undermine human development? Relevant contexts and social mechanisms. In Booth A, Crouter AC, eds. Does it take a village? Community effects on children, adolescents, and families. Mahwah, NJ: Lawrence Erlbaum Associates, Publishers, 2001:3–30. 16. Putnam RD. Bowling alone: the collapse and revival of the American community. New York: Simon & Schuster, 2000. 17. Davis HE. The Vermont ACO landscape. October 22, 2015. https://vtdigger.org/2015/10/22/the-vermont-aco-landscape Accessed October 28, 2016. 18. Bland SE, Crowley JS, Gostin LO. Strategies for health system innovation after Gobeille vs. Liberty Mutual Insurance Company. JAMA 2016;316:581–582. 19. Taylor EF, Lake T, Nysenbaum J, Peterson G, Meyers D. Coordinating care in the medical neighborhood: critical components and available mechanisms. https://pcmh.ahrq .gov/sites/default/files/attachments/Coordinating%20Care% 20in%20the%20Medical%20Neighborhood.pdf Accessed June 20, 2015. 20. Koshy R, Conrad DA, Grembowski D. Lessons from Washington State’s medical home payment pilot: what it will take to change American healthcare. Popul Health Manag 2015;18:237–245. 21. Weiner BJ. A theory of organizational readiness for change. Implement Sci 2009;4:67. 22. Baicker K, Levy H. Coordination versus competition in health care reform. N Engl J Med 2013;369:789–791. 23. Green Mountain Care Board. GMCB Annual Report— January 15, 2015. http://gmcboard.vermont.gov/documents/ publications/annual-reports/annualrpt011515 Accessed April 23, 2015.
Downloaded by University of Connecticut e-journal package NERL from online.liebertpub.com at 09/04/17. For personal use only.
8
24. Grogan CM. Encouraging competition and cooperation: the Affordable Care Act’s contradiction? J Health Polit Policy Law 2015;40:633–638. 25. Mansfield E. How single-payer health care became a singleprovider plan. October 26, 2016. https://vtdigger.org/2016/ 10/26/single-payer-health-care-became-single-provider-plan Accessed October 27, 2015. 26. Vermont Legislative Joint Fiscal Office. All-payer model— potential benefits, risks and outstanding questions. www .leg.state.vt.us/jfo/healthcare/APM_summary_sheet.pdf Accessed October 29, 2016. 27. Mansfield E. Joint Fiscal Office points to unanswered questions on all-payer model. October 26, 2016. https:// vtdigger.org/2016/10/26/joint-fiscal-office-points-unansweredquestions-payer-model Accessed October 27, 2016. 28. Mansfield E. Vermont Care Organization to rework vision, priorities. May 16, 2017. https://vtdigger.org/2017/05/16/
GREMBOWSKI AND MARCUS-SMITH
vermont-care-organization-rework-vision-priorities Accessed May 26, 2017. 29. Goswami NP. Shumlin signs all-payer health care plan. October 28, 2016. http://101-btweb.newscyclecloud.com/ article/20161028/NEWS03/161029621/section/NEWS04 Accessed July 13, 2017.
Address correspondence to: David Grembowski, PhD Department of Health Services School of Public Health University of Washington Box 357660 Seattle, WA 98195-7660 E-mail: [email protected]
Downloaded by University of Connecticut e-journal package NERL from online.liebertpub.com at 09/04/17. For personal use only.
POPULATION HEALTH MANAGEMENT Volume 00, Number 00, 2017 ª Mary Ann Liebert, Inc. DOI: 10.1089/pop.2017.0061
The 10 Conditions That Increased Vermont’s Readiness to Implement Statewide Health System Transformation David Grembowski, PhD,1 and Miriam Marcus-Smith, MHA2
Abstract
Following an arduous, 6-year policy-making process, Vermont is the first state implementing a unified, statewide all-payer integrated delivery system with value-based payment, along with aligned medical and social service reforms, for almost all residents and providers in a state. Commercial, Medicare, and Medicaid valuebased payment for most Vermonters will be administered through a new statewide accountable care organization in 2018–2022. The purpose of this article is to describe the 10 conditions that increased Vermont’s readiness to implement statewide system transformation. The authors reviewed documents, conducted internet searches of public information, interviewed key informants annually in 2014–2016, cross-validated factual and narrative interpretation, and performed content analyses to derive conditions that increased readiness and their implications for policy and practice. Four social conditions (leadership champions; a common vision; collaborative culture; social capital and collective efficacy) and 6 support conditions (money; statewide data; legal infrastructure; federal policy promoting payment reform; delivery system transformation aligned with payment reform; personnel skilled in system reform) increased Vermont’s readiness for system transformation. Vermont’s experience indicates that increasing statewide readiness for reform is slow, incremental, and exhausting to overcome the sheer inertia of large fee-based systems. The new payments may work because statewide, uniform population-based payment will affect the health care of almost all Vermonters, creating statewide, uniform provider incentives to reduce volume and making the current fee-based system less viable. The conditions for readiness and statewide system transformation may be more likely in states with regulated markets, like Vermont, than in states with highly competitive markets. Keywords: value-based payment, accountable care organization, readiness, system transformation, Vermont, State Innovation Models
V
ermont is implementing a statewide, all-payer integrated delivery system that was developed incrementally during a 6-year (2011–2016) policy-making process. Vermont’s reforms have significant policy implications because Vermont is the first state to implement a unified (or ‘‘standardized’’) statewide all-payer system with value-based payments, along with aligned medical and social services, for almost all residents and providers in a state. Vermont’s historic, statewide system transformation is made possible by federal regulations and policies. In October 2016, the State of Vermont signed the formal agreement with the Centers for Medicare & Medicaid Services (CMS) to waive federal regulations and allow other forms of Medicare and Medicaid payment.1 Commercial, Medicare, 1 2
and Medicaid value-based payment for the entire state will be administered through a new statewide accountable care organization (ACO), or Vermont All-Payer ACO Model (VAPM). The Model began on January 1, 2017, and ends on December 31, 2022, a 6-year life; the first year (2017) is devoted to planning and creating the VAPM infrastructure. By the end of Year 3, the VAPM must have a plan for the financing and delivery of Medicaid behavioral health and Medicaid home and community-based services. CMS favors spreading similar statewide payment and care delivery reforms across all major payers in other states.2 Table 1 presents the key features of Vermont’s statewide integrated delivery system. In brief, VAPM has multiple sources of revenue, mainly Medicaid, Medicare, and the health
Department of Health Services, School of Public Health, University of Washington, Seattle, Washington. Division of General Internal Medicine, School of Medicine, University of Washington, Seattle, Washington.
1
2
GREMBOWSKI AND MARCUS-SMITH
Downloaded by University of Connecticut e-journal package NERL from online.liebertpub.com at 09/04/17. For personal use only.
Table 1. Features of Vermont’s Statewide All-Payer Integrated Delivery System Feature
Description
Payers
About 300,000 Vermonters are covered by commercial health insurance. BlueCross BlueShield of Vermont has 80% of the commercial market. Medicaid and other subsidized health care covers about a third of Vermonters. Vermont has 3 ACOs, currently with more than 160,000 total patients attributable to Medicare, Medicaid, and commercial plans: OneCare Vermont: formed in 2012 by University of Vermont Medical Center and Dartmouth-Hitchcock Medical Center, consisting of all Vermont hospitals and their employed physicians, some FQHCs and independent medical offices, as well as mental health and substance abuse agencies, and skilled nursing facilities (about 100,000 patients). CHAC: formed in 2014 and consisting of Vermont’s 11 FQHCs, along with mental health and substance abuse agencies and skilled nursing facilities (about 57,000 patients). Vermont Collaborative Physicians/Accountable Care Coalition of the Green Mountains: formed in 2012 and consisting of independent primary care and specialty practices (about 9000 patients). Medicare, Medicaid, and commercial health plans pay ACOs all-inclusive population-based payments, which are similar to capitation payments, for their attributed beneficiaries. The ACOs distribute payments to health care organizations for their attributed patients. Using 2017 as a baseline, Vermont is expected to limit per capita health care expenditure growth for all major payers to 3.5% in the 5-year demonstration. Vermont also is expected to limit Medicare per capita health care expenditure growth for Medicare beneficiaries to at least 0.001–0.002 percentage points below the projected national average. Medicaid does not have a cost growth target. Payment reform is built on a primary care foundation, Blueprint for Health (Blueprint), the statewide, multipayer program that meets National Committee for Quality Assurance medical home standards, providing primary care to three fourths of Vermonters. Blueprint provider teams link patients with social and other services to address patient needs. Vermont is divided into regional Health Service Areas, each with its own hospital. Regional collaboratives are coordinating medical and social services to address the needs of the whole person in each region. Representatives from the local hospital, ACOs, Blueprint, FQHCs, behavioral health/community mental health centers, home care agencies, aging agencies, skilled nursing, housing agencies, and other health and social service agencies meet regularly, with the common goal of integrating services across agencies to address local population needs, particularly for people with mental health, substance abuse, multiple chronic conditions, and social problems. Care coordination programs also are in place to reduce fragmentation and service duplication for persons with multiple chronic conditions. State-level and ACO-level performance targets are imposed to create incentives for collaboration between the care delivery and public health systems, which in turn lead to system reforms that achieve the targets. Three categories of targets exist: (1) process milestones; (2) health delivery system; and (3) population-level health outcomes. In each category, performance targets are set for access to primary care, substance use disorder, suicides, and chronic conditions. Statewide information systems are in place or in development: All-payer claims database for Medicare, Medicaid, and 90% of commercial insurance Vermont Health Information Exchange (VHIE), a statewide clinical information system providing real-time data to support clinical care. VHIE collects and warehouses data from electronic medical records for all Vermont’s hospitals, most FQHCs, and Blueprint primary care offices. VHIE also delivers clinical information, such as laboratory reports and patient registries for chronic diseases, to providers, who can query patient records for consenting patients. The Green Mountain Care Board, an independent, 5-member body created by Act 48 in 2011, has broad regulatory authority over Vermont’s health care system.
ACOs
Value-Based Payment & Financial Incentives Cost Growth Targets
Primary Care
Delivery System Transformation
Performance Targets for Quality of Care and Health Outcomes
Information System
Vermont All-Payer ACO Model Regulation
ACOs, Accountable Care Organizations; CHAC, Community Health Accountable Care; FQHCs, Federally Qualified Health Centers.
plans in the Vermont exchange. Vermonters retain their multiple public and private health plans but receive services through a single, simplified and standardized administrative and value-based payment structure administered by a new, statewide ACO, the Vermont Care Organization (VCO), reg-
ulated by the Green Mountain Care Board (GMCB). Participation is voluntary in the provider-led VAPM. Payment will be based on the CMS Next Generation ACO Model. Although the payment details are in development, in general, the VCO will receive all-inclusive population-based payments, which
Downloaded by University of Connecticut e-journal package NERL from online.liebertpub.com at 09/04/17. For personal use only.
VERMONT’S READINESS FOR SYSTEM TRANSFORMATION
are similar to capitation payments, for their attributed beneficiaries in Medicare, Medicaid, and commercial health plans. The VCO, in turn, will distribute payments to Vermont’s 14 hospitals and Federally Qualified Health Centers, which employ more than two thirds of Vermont primary care physicians, and other health care organizations. The health plans will have common measures of performance, and each plan will retain its own benefits. Payment reform is aligned with delivery transformation to address the needs of the whole person. Reaching scale is critical for the VAPM’s success. To have a viable statewide payment model, the VCO must enroll a substantial majority of Vermonters and primary care providers. By the end of 2022, the VCO must enroll 70% of all Vermonters, 90% of Medicare beneficiaries, and 80% of primary care providers.1 Table 1 indicates that about half of Vermont’s primary care physicians participate in Vermont’s 2 ACOs, OneCare Vermont (30%) and Community Health (20%). Vermont is more likely to achieve the 80% provider target if both ACOs participate in the VAPM, which, in turn, also would increase the enrollment of Vermonters in the VAPM. To incentivize provider participation, Vermont Medicare providers are eligible to participate in the Medicare Access and CHIP Reauthorization Act’s (MACRA) advanced alternative payment model and receive a 5% bonus for their Medicare patients. A key assumption of the VAPM is that by reaching scale, the statewide population-based payments will affect most patients and providers in most medical organizations, creating common, statewide payment incentives to reduce volume while also diminishing the viability and influence of the current fee-for-service (FFS) payment in the delivery of health care. The population-based payments for most Vermonters may generate revenue streams for providers that are more predictable and dependable over time than FFS, reducing financial uncertainty and promoting sustainability. The purpose of this article is to describe the 10 conditions that increased Vermont’s readiness to launch its statewide integrated delivery system with all-payer, population-based payments. Few studies have examined the development of system readiness.3 Diffusion theory posits that innovations, such as Vermont’s VAPM, are not adopted all at once but rather are spread over time across health care organizations through a predictable process with 3 distinct stages: readiness, adoption, and implementation.3,4 System readiness is a preadoption stage during which payers, health plans, health care organizations, and providers become aware of payment and delivery reform models and learn how the models work and about their potential consequences. Based on this knowledge, health care organizations may abandon payment reform or move toward a state of readiness for reform with buy-in from most stakeholders. Greater readiness increases the likelihood that health care organizations will adopt, implement, and sustain payment reform and system tranformation. Other states may apply these findings to increase their readiness for statewide health system reform. Methods
This study of Vermont was part of a larger qualitative comparative case study of 11 value-based payment projects being carried out in 8 states funded by the Robert Wood Johnson Foundation (RWJF) and evaluated by University of
3
Washington researchers (the other states were California, Maine, Massachusetts, New Hampshire, Oregon, Pennsylvania, and Washington).3–5 The study team applied guidelines from organizational sociology and qualitative methods for conducting comparative case studies.6–10 In keeping with the principles of qualitative research, the approach to data collection and analysis was inductive, structured around a set of common questions, and allowed for open-ended responses and follow-up probes. As such, this paper neither develops a conceptual model, nor tests explicit hypotheses, but seeks to articulate lessons learned and implications for statewide health system transformation from qualitative analysis. Study protocols were reviewed and approved by the University of Washington’s Institutional Review Board. The qualitative methodology followed 6 steps. First, the evaluation team developed a semi-structured interview instrument with common questions for all sites covering context, stakeholders, objectives, approach, logic model, progress and results, facilitators and barriers, evaluation and sustainability, and lessons learned (see questionnaire Supplementary Appendix S1; Supplementary Data are available online at www.liebertpub.com/pop). The study team obtained each site’s grant application, documents from each project’s public website, and other local, publicly available information about each project. Based on this information, the team developed additional questions unique to each site. The team made a list of key informants at each site based on the site’s RWJF application records and site documents. Common data collection and analysis protocols were followed across sites. The study team then conducted 60-minute interviews with up to 15 key informants in Vermont each year for 3 years (2014–2016). Almost all interviews were in person; the remainder were by phone. In 2014 and 2015 interviews were conducted by the first and second authors, with the second author taking real-time interview notes on a laptop computer and audio-recording each interview; in 2016 interviews were conducted only by the first author, audio-recorded and transcribed, because of changes in funding and study staff. All key informants consented verbally to be interviewed and recorded at the beginning of each interview. Prior to the interview, key informants received the interview questions and information about how the interview information would be used for analysis and publication. Third, qualitative analysis was performed for each year of interviews following accepted protocols.10 Before analyzing interview data, the authors reviewed the interview notes, newspaper articles about the project, and background documents from public websites. Key informants’ responses to each interview question and site documents were compared to derive answers to each question, facilitators and barriers to implementation, and implications for policy and practice. Next, to document the answers to the interview questions, the study team produced an annual site report for all sites. For Vermont’s annual site report, the first author prepared the first draft of the site report, which then was reviewed by the second author. The 2014 and 2015 site reports also were reviewed by at least 1 other team member. Each annual report was then vetted with the lead contact for Vermont’s RWJF project, who was asked to identify any apparent errors of omission or commission and to specify any inadvertent breaches of confidentiality or revelations of proprietary or
Downloaded by University of Connecticut e-journal package NERL from online.liebertpub.com at 09/04/17. For personal use only.
4
competitively sensitive information. The team reviewed the site’s critique and revised the documents based on team consensus. Finally, the study team compared Vermont’s 3 annual site reports to derive persistent conditions promoting or inhibiting readiness for statewide transformation of Vermont’s health care system (see Supplementary Appendix S2, for list of persistent facilitators and barriers to implementation in the annual reports). In selected years, the annual reports for all sites were reviewed by the study’s Advisory Group, composed of 8 national experts in payment reform. The advisors suggested themes and insights relevant to payment reform. Their comments informed the insights in this paper. Results
The following 10 conditions are essential ingredients for increasing readiness for payment and system transformation in the face of day-to-day challenges. The conditions are interconnected and operate jointly to increase readiness. The conditions are divided into 2 categories: social and support. Social conditions: pay attention to the ‘‘Cs’’
(1) Champions Moving from FFS to value-based payment throughout the state is a journey into the unknown. Committed champions play a vital role in guiding deliberations toward decisions about what the new system will look like. In particular, the governor and the GMCB, through its role as facilitator and convener of diverse stakeholders to participate voluntarily in payment and system reform without antitrust threats, were instrumental in sustaining momentum toward payment and system reform. From 2011–2016 there was little turnover in the champions, which sustained momentum. (2) Common vision of the future payment reform A common vision of the future statewide payment and delivery system is essential for building readiness and getting stakeholders ‘‘on the same page.’’ Two competing visions of system reform were advanced: (1) initially, a single-payer universal health insurance program; or (2) an all-payer integrated delivery system. On December 17, 2014, Governor Shumlin ‘‘pulled the plug’’ on the singlepayer program, mainly because of its enormous costs that Vermont taxpayers could not afford, and attention shifted to the all-payer model.11,12 (3) Culture of collaboration and cooperation (not competition), communication, compromise and consensus-building, and civic engagement in a small, rural state Vermont is a small, rural state with about 625,000 residents and little market competition, which promotes communication and allows everyone to ‘‘be at the table’’ and adopt a statewide perspective to health system reform. As one key informant noted, ‘‘If you’re not at the table, you may be on the menu.’’ Vermont also has a communitarian culture and a history of collaboration and civic engagement among stakeholders, who are willing to cooperate and compromise for the public good. In a state where ‘‘everyone knows each other,’’ many stakeholders have worked together on health care issues for many years and have a collective knowledge of Vermont’s
GREMBOWSKI AND MARCUS-SMITH
history of health care, which facilitates communication and building consensus. Payment reform and system transformation engaged literally hundreds of Vermonters. Overall direction was provided by an 8-member Core Team, who received guidance from a 37-member Steering Committee. The Team and Committee, in turn, received guidance from at least 6 workgroups addressing payment models, quality and performance measurement, care management, and other topics. (4) Social capital and collective efficacy Social capital and collective efficacy are key ingredients of system readiness and prerequisite to achieve system reform.13–16
Social capital may be defined as a characteristic of groups, organizations, or communities – such as trust in others, norms of reciprocity, and civic participation – that can facilitate collaboration and cooperation, including the flow of information, for mutual benefit. Collective efficacy, which is related to social capital, refers to a group’s shared beliefs in their collective ability to produce desired results. The linkage of mutual trust and the shared willingness to intervene for the common good define a group’s capacity to achieve common goals. Bandura notes that the higher the perceived collective efficacy, the stronger the group’s staying power in the face of impediments and setbacks, and the greater their accomplishments.13 The 2 concepts suggest that the greater the stock of social capital and collective efficacy, the greater the likelihood of implementing statewide payment reforms. Statewide transformation from FFS to value-based payment is disruptive change that can generate uncertainty, fear, and distrust in all sectors of Vermont’s health care system. Building trust and social capital, a sense that ‘‘we work better together,’’ is time-consuming yet essential to working through the fear and uncertainty, and to sustaining willingness to collaborate and reach consensus about what payment reform will look like. Information is a part of social capital – credible evidence that the new payment model will ‘‘do no harm’’ to the financial health of providers may reduce uncertainty and improve stakeholders’ comfort levels with the payment reforms. Collective efficacy is a resource for staying on track. Over time, a ‘‘coalition of the willing’’ emerged: well-known leaders who are respected by most interest groups and have a collective sense that ‘‘we can do this.’’17 Collective efficacy is related to social capital, in that trust is the glue that holds the coalition together and fosters collaboration. Some stakeholders noted that the State Innovation Model (SIM) grant provided the resources and time to invest in trust and build the coalition of the willing. Support conditions
(5) Money Significant monetary resources are required to build readiness for statewide system transformation, particularly in a resource-scarce state. The GMCB received a grant from RWJF in 2012 to design and implement payment reform strategies. Development work under the grant eventually led to a 3-year, $45 million SIM award from the CMS’ Center
Downloaded by University of Connecticut e-journal package NERL from online.liebertpub.com at 09/04/17. For personal use only.
VERMONT’S READINESS FOR SYSTEM TRANSFORMATION
for Medicare & Medicaid Innovation to the Office of the Governor in 2013. SIM funding was well aligned with Vermont’s goals: the SIM Initiative is intended explicitly to create statewide system transformation for most care in a state. However, time remained a scarce resource. Readiness preparations were a burden for stakeholders when the numerous, ongoing, day-to-day demands for meetings and system design work crowded out time for their regular jobs. (6) Statewide data A statewide data infrastructure is required for planning the new statewide system, coordinating care to address local patient and population needs across diverse health care and social service organizations, and the monitoring and public reporting of performance for management and accountability. Data flow also contributes to social capital: looking at data together is another way to build trust among stakeholders. Most stakeholders endorsed a single, transparent, universal set of performance measures for all payers. The VAPM has its own performance measures, which over time may become the universal measures for all payers in the state. Simply collecting data is insufficient; data must be analyzed to create ‘‘actionable’’ results that inform clinical care and how to change the system to reduce costs and improve quality. GMCB oversees Vermont’s all-payer claims database, which covers 90% of commercially insured Vermonters and all Medicaid and Medicare enrollees, although Medicare data are available only after a long time lag. Vermont Information Technology Leaders, with oversight from the GMCB, is building the Vermont Health Information Exchange, which contains clinical patient records for all hospitals and most primary care offices with real-time data access to support coordination of clinical care. Vermont’s statewide data infrastructure is being used to support system transformation and accountability. The regional collaboratives are using the data infrastructure to support patient care and, at the population level, identify local needs and align services with those needs. The GMCB has developed a ‘‘dashboard’’ that profiles Vermont’s health care system and comprises 51 medical and nonmedical indicators. In the 14 regions, insurers, Blueprint for Health (Blueprint: Vermont’s statewide primary care program that meets National Committee for Quality Assurance medical home standards and has community health teams to link patients with social and other services to meet the needs of the whole person), ACO leaders, and other groups are coproducing customized performance reports for quality, cost, utilization, and other measures. Vermont’s data infrastructure has limitations. The US Supreme Court ruling in Gobeille (in this individual’s official capacity as Chair of GMCB) vs. Liberty Mutual Insurance Company essentially prohibits the state from collecting claims data from self-insured entities.18 In addition, linked claims– clinical data do not exist for most Vermont residents. Vermont also lacks regional health assessment surveys. (7) Legal infrastructure A legal infrastructure is required for payment reform. Most federal, state, and local laws and regulations support FFS reimbursement but may not offer a legal foundation for an all-payer model with population-based payment. Legislation may reduce uncertainty by defining what will happen rather than what might happen in payment reform.
5
On the heels of the federal Affordable Care Act, in 2011 the Vermont Legislature, with strong leadership from Governor Shumlin, passed Act 48, promising a ‘‘universal and unified health system’’ as a ‘‘public good’’ that also controls costs and offers quality care. The Legislature’s intention was to move Vermont toward a more highly regulated and integrated health care system but specified few details about what the new system might look like or how it would work. In June 2015, the Legislature passed Act 54 (Section 139) to explore an ‘‘all-payer model, which may be achieved through a waiver’’ from CMS. To prepare for an all-payer model, in May 2016, the Legislature passed Act 113 (H.812), which authorized the GMCB and Vermont’s Agency of Administration (or designee) to enter an agreement with CMS to waive Medicare regulations if certain criteria are met. The Act requires that all ACOs must have GMCB certification to receive payments from Medicaid or commercial insurance, and also mandates that Medicaid payments be integrated into the all-payer model. (8) Federal policy promoting payment reform Federal policies are primarily facilitators but can be barriers to statewide payment and delivery system reforms. Beginning in 2013 Vermont’s 3 ACOs participated in a statewide all-payer (commercial, Medicare, and Medicaid) shared savings program, which was based on the CMS– ACO shared savings model and served as a stepping stone to the VAPM. Vermont’s initial work on the all-payer model was followed in 2015 by 3 federal policies to move Medicare payment away from FFS to value-based payment: 1) Secretary Burwell’s announcement of tying 50% of Medicare payments to alternative payment models by the end of 2018; 2) the CMS Next Generation ACO Model, which allowed ACOs to assume higher levels of financial risk and reward (in 2015 OneCare Vermont, an ACO, accepted CMS’ invitation to enroll in the Next Generation ACO Model); and 3) Congress’ MACRA, which pays physicians extra for participating in alternative payment models and increases financial incentives in pay-for-performance/ quality reporting programs. The new federal policies legitimized and added credibility to Vermont’s all-payer model, and reinforced stakeholder views that payment reform is inevitable, creating local and statewide momentum to collaborate on value-based payment reform. However, the federal regulations and the burdensome process for obtaining the waivers generally inhibit innovation and increase the workload and time for system change. (9) Statewide delivery system transformation aligned with payment reform In the US health care system, patients with chronic conditions, disabilities, and social problems (eg, substance abuse) often receive fragmented, uncoordinated, and inefficient services that focus on siloed conditions rather than adopting a patient-centered approach that addresses the diverse needs of the whole person, which can reduce quality of care and increase costs. Low connectivity across health care organizations promotes fragmented care in the system.
Downloaded by University of Connecticut e-journal package NERL from online.liebertpub.com at 09/04/17. For personal use only.
6
To address this problem, for the past decade Blueprint’s community health teams have linked primary care patients with social and other services. Fostered by the SIM grant, each region has formed a collaborative comprising diverse health and social service agencies, which meets regularly to integrate services to address local population needs, particularly the needs of people with mental health, substance abuse, and other social problems. Collaboratives are expected to adopt a data-driven process to prioritize local needs and align services with those needs on a population level, and develop guidelines to identify persons with high service needs and how to address them in ‘‘medical neighborhoods.’’19 As part of this activity, the ACOs and Blueprint are working to coordinate their activities in all of the state’s regions. In a majority of regions, Integrated Communities Care Management Learning Collaboratives, along with a Medicaid chronic care initiative, also are advancing care coordination for people with complex needs through social and health service integration. Collaboration has challenges. Clinical providers faced a cultural switch of moving from a traditional medical model, with the physician in charge, toward team medicine responsible for managing individual and population health. Aligning delivery system transformation and payment reform was challenged by health care organizations and providers with little experience bearing financial risk and viewing a service as an expense (in population-based payment) rather than revenue (in patient-based FFS payment). (10) Personnel skilled in system reform Personnel and consultants with skills and experience are required to increase readiness. Transforming a health care system from FFS to value-based payment is complex work. Personnel originally hired to manage the current FFS system may be unprepared to innovate and engage in system transformation. Hiring new personnel and consultants with skills that align with payment reform goals may contribute to readiness. In particular, stakeholders recommend having a professional data manager(s) wellversed in public and commercial claims and clinical data from the beginning. Discussion
Vermont’s experience suggests that readiness is an essential building block in implementing statewide system transformation. The 2011–2016 process for statewide readiness was slow, incremental, and exhausting, but nonetheless successful in launching Vermont’s All-Payer Model. The study team posits that the conditions increase the likelihood of greater system readiness and overcoming the sheer inertia of large systems. However, the conditions do not guarantee readiness. For instance, New Hampshire abandoned the goal of global payments when the largest payer (covering about 50% of privately insured lives) declined to go forward with the common payment model.3 In particular, the study team hypothesizes that statewide system readiness may be more likely if a state has champions for changing the whole system of the entire state to align payment and delivery reforms,20 has a history of collaboration, and if the likelihood of building social capital and collective efficacy for payment reform are high.21 Re-
GREMBOWSKI AND MARCUS-SMITH
placing volume-based FFS carries great uncertainty and fear for payers, health care organizations, and other stakeholders, mainly because the potential financial consequences of doing so are huge. Distrust, conflict, and emotions can erupt in payment discussions, particularly among organizations that view each other as competitors, mainly over how much upside gain and downside financial risk the new payment model will have, and how revenue will be distributed across health care organizations and providers. Collaboration and trust building are time-consuming yet prerequisite to working through the inevitable conflicts among stakeholders, sustaining the willingness to weather the ups and downs, and ultimately reaching consensus. Competition is a key approach for improving value through lower prices and higher quality.22 However, if the hypothesis is correct, statewide system readiness may be less likely in states with highly competitive markets, but more likely in states with government regulation, because trust in others, cooperation and reciprocity, and collective efficacy for the public good may be lower in highly competitive markets. Vermont has a long history of stateregulated health care and belief in the beneficial role of state government. More than 20 years ago Vermont implemented regulations limiting rate variations and preventing health plans from denying coverage based on preexisting medical conditions. The GMCB has authority to control costs, such as setting payment rates for health professionals and approving hospital budgets, certificates of need, and payer premiums. In 2014 net patient revenues in Vermont’s hospitals increased 2.7%, which was much lower than the 5.1% to 9.2% annual increases in 2002 to 2013, reflecting GMCB enforcement of target growth rates for hospitals in 2013 and 2014.23 Vermont’s Medicare expenditures per capita are among the lowest across states. Grogan notes that the tension between competition versus cooperation is the most persistent contradiction in US health policy over the past several decades and a central feature of ACOs.24 Vermont’s all-payer model may not have happened without government support. CMS’s rollout of the Next Generation ACO payment model was timely and appealing to Vermont stakeholders, partly because many believed the model would spread nationwide in the future, and because Vermont’s governor and legislators were also on board. Adopting the federal model also was much easier than creating their own payment model from scratch. In general, payers want to reduce spending and providers want to increase revenue, and it may be impossible for them, working on their own, to agree and create a payment model that satisfies both interests. After 6 years most, but not all, stakeholders in Vermont supported the VAPM. Some of Vermont’s small health care organizations, such as independent physicians, have not endorsed the VAPM, questioning the finances of the allpayer model, how the model would work, and whether the large health care organizations would dominate over the small ones in the VAPM.17,25 The Vermont Legislature Joint Fiscal Office also identified the potential benefits, risks, and outstanding questions of the VAPM that remain to be answered.26,27 From 2011–2015 there was little turnover in the champions of the all-payer model. However, in 2016 some
Downloaded by University of Connecticut e-journal package NERL from online.liebertpub.com at 09/04/17. For personal use only.
VERMONT’S READINESS FOR SYSTEM TRANSFORMATION
leaders left their positions, and in 2017 Vermont has a new governor, a new speaker of the House, and new president of the Senate. Turnover may signal a loss of intelligence about local payment reform or a loss of leadership that may undermine implementation of the all-payer model. Alternatively, turnover may introduce new expertise and ideas into payment reform work and ultimately improve the reform’s chances of success. In closing, payment and health system reform has been an ongoing, seemingly never-ending yet successful process culminating in the launch of Vermont’s statewide all-payer model in 2018. Two basic questions remain. First, will the all-payer model be implemented as planned? Although the VCO is operational and has a board of directors, implementation has stalled because the VCO has not yet received the government funds that it expected for start-up and Medicaid programs and other factors.28 The second basic question is: Will it work? Will patients and providers participate in the new system? Or will patients take their health care and revenue to New Hampshire or New York? Or will the system work so well that people travel or move to Vermont for their health care? Is Vermont too small a state, or just the right size, to carry out innovative health care reform? Can Vermont’s statewide system reforms work elsewhere in the United States, given CMS’ intent to spread state-driven system transformation? Vermont may be unique, as Governor Shumlin notes:29 Only in Vermont would you find a health care community coming together from all walks, all sectors, saying together, ‘We think we have a new way to improve quality and bend the cost curve in a way that will make Vermont (health care) more affordable.’ Answers to these questions are unknown but are of the utmost importance for health policy in Vermont and for whether the policy is generalizable to the United States. Author Disclosure Statement
The authors declare that there are no conflicts of interest. The authors received the following financial support for the research, authorship, and/or publication of this article: Support for this research was provided by the Robert Wood Johnson Foundation (Grant Nos. 71209 and 73446). References
1. Green Mountain Care Board. Vermont All-Payer Accountable Care Organization Model Agreement. http://gmcboard .vermont.gov/payment-reform/APM Accessed May 28, 2017. 2. Centers for Medicare & Medicaid Services. Vermont AllPayer ACO Model. 2017. https://innovation.cms.gov/ initiatives/vermont-all-payer-aco-model Accessed October 27, 2016. 3. Conrad DA, Vaughn M, Grembowski D, Marcus-Smith M. Implementing value-based payment reform: a conceptual framework and case examples. Med Care Res Rev 2016;73: 437–457. 4. Conrad D, Grembowski D, Gibbons C, et al. A report on eight early-stage state and regional projects testing value-based payment. Health Aff ( Millwood) 2013;32: 998–1006.
7
5. Conrad DA, Grembowski D, Hernandez SE, Lau B, Marcus-Smith M. Emerging lessons from regional and state innovation in value-based payment reform: striking a balance between collaboration and disruptive innovation. Milbank Q 2014;92:568–623. 6. Ayres L, Kavanaugh K, Knafl KA. Within-case and acrosscase approaches to qualitative data analysis. Qual Health Res 2003;13:871–883. 7. Firestone WA, Herriott RE. Multisite qualitative policy research: some designs and implementation issues. In: Fetterman DM, ed. Ethnography in educational evaluation. Beverly Hills, CA: Sage Publications, 1984:63–88. 8. Pettigrew AM. Longitudinal field research on change: theory and practice. Organ Sci 1990;1:267–292. 9. Stake RE. Multiple case study analysis. New York: The Guilford Press, 2006. 10. Dey I. Qualitative data analysis. A user-friendly guide for social scientists. London: Routledge, 1993. 11. Goodnough A. In Vermont, frustrations mount over Affordable Care Act. June 4, 2015. https://www.nytimes.com/2015/ 06/05/us/in-vermont-frustrations-mount-over-affordable-careact.html Accessed July 13, 2017. 12. True M. Big expectations, few results on health care. May 20, 2015. https://vtdigger.org/2015/05/19/legislative-wrapbig-expectations-few-results-on-health-care Accessed June 1, 2015. 13. Bandura A. Exercise of human agency through collective efficacy. Curr Dir Psychol Sci 2000;9:75–78. 14. Sampson RJ. Neighborhoods and violent crime: a multilevel study of collective efficacy. Science 1997;277: 918–924. 15. Sampson RJ. How do communities undergird or undermine human development? Relevant contexts and social mechanisms. In Booth A, Crouter AC, eds. Does it take a village? Community effects on children, adolescents, and families. Mahwah, NJ: Lawrence Erlbaum Associates, Publishers, 2001:3–30. 16. Putnam RD. Bowling alone: the collapse and revival of the American community. New York: Simon & Schuster, 2000. 17. Davis HE. The Vermont ACO landscape. October 22, 2015. https://vtdigger.org/2015/10/22/the-vermont-aco-landscape Accessed October 28, 2016. 18. Bland SE, Crowley JS, Gostin LO. Strategies for health system innovation after Gobeille vs. Liberty Mutual Insurance Company. JAMA 2016;316:581–582. 19. Taylor EF, Lake T, Nysenbaum J, Peterson G, Meyers D. Coordinating care in the medical neighborhood: critical components and available mechanisms. https://pcmh.ahrq .gov/sites/default/files/attachments/Coordinating%20Care% 20in%20the%20Medical%20Neighborhood.pdf Accessed June 20, 2015. 20. Koshy R, Conrad DA, Grembowski D. Lessons from Washington State’s medical home payment pilot: what it will take to change American healthcare. Popul Health Manag 2015;18:237–245. 21. Weiner BJ. A theory of organizational readiness for change. Implement Sci 2009;4:67. 22. Baicker K, Levy H. Coordination versus competition in health care reform. N Engl J Med 2013;369:789–791. 23. Green Mountain Care Board. GMCB Annual Report— January 15, 2015. http://gmcboard.vermont.gov/documents/ publications/annual-reports/annualrpt011515 Accessed April 23, 2015.
Downloaded by University of Connecticut e-journal package NERL from online.liebertpub.com at 09/04/17. For personal use only.
8
24. Grogan CM. Encouraging competition and cooperation: the Affordable Care Act’s contradiction? J Health Polit Policy Law 2015;40:633–638. 25. Mansfield E. How single-payer health care became a singleprovider plan. October 26, 2016. https://vtdigger.org/2016/ 10/26/single-payer-health-care-became-single-provider-plan Accessed October 27, 2015. 26. Vermont Legislative Joint Fiscal Office. All-payer model— potential benefits, risks and outstanding questions. www .leg.state.vt.us/jfo/healthcare/APM_summary_sheet.pdf Accessed October 29, 2016. 27. Mansfield E. Joint Fiscal Office points to unanswered questions on all-payer model. October 26, 2016. https:// vtdigger.org/2016/10/26/joint-fiscal-office-points-unansweredquestions-payer-model Accessed October 27, 2016. 28. Mansfield E. Vermont Care Organization to rework vision, priorities. May 16, 2017. https://vtdigger.org/2017/05/16/
GREMBOWSKI AND MARCUS-SMITH
vermont-care-organization-rework-vision-priorities Accessed May 26, 2017. 29. Goswami NP. Shumlin signs all-payer health care plan. October 28, 2016. http://101-btweb.newscyclecloud.com/ article/20161028/NEWS03/161029621/section/NEWS04 Accessed July 13, 2017.
Address correspondence to: David Grembowski, PhD Department of Health Services School of Public Health University of Washington Box 357660 Seattle, WA 98195-7660 E-mail: [email protected]
