With the expansion of coverage as a result of federal health care reform, safety net providers are confronting a challenge to care for the underserved while also competing as a provider of choice for the newly insured. Safety net institutions may be able to achieve these goals by pursuing greater delivery system integration. We interviewed safety net hospital and community health center (CHCs) leaders in five US cities to determine what strategies these organizations are employing to promote care integration in the safety net. Although there is some experimentation with payment reform and health information exchange, safety net providers identify significant policy and structural barriers to integrating service delivery. The enhanced Medicaid payments for CHCs and the federal requirement that CHCs retain independent boards discourage these organizations from integrating with other safety net providers. Current policies are not mobilizing safety net providers to pursue integration as a way to deliver more efficient and effective care. Medicaid and other policies at the federal and state level could be revised to overcome known fragmentation in the health care safety net. This includes addressing the conflicts in financing and governance arrangements that are encouraging providers to resist integration to preserve their independence.
Editor's note: The two essays in this issue's Report on Health Reform Implementation section emerged from a workshop, generously funded by the Robert Wood Johnson Foundation, that was held in Chicago, Illinois, in January 2014. The purpose of the workshop was threefold: first, to increase communication and learning between state-level policy practitioners and health policy researchers; second, to address key ACA implementation issues that states are currently grappling with; and third, in response to these issues, to identify useful policy instruments and strategies for dissemination across the states. With these goals in mind, we asked several policy practitioners in different states to submit questions on current implementation challenges that might benefit from the insights of a policy researcher. We then identified researchers with significant expertise in applicable areas to respond to a small selection of these important questions. Howard Kahn's question on how safety net clinics and hospitals will evolve in the wake of recent health care reform and the response by Julia Murphy et al. is an example of the work that came out of this productive process. This is the third and final set of essays to be published from the January 2014 workshop. We welcome any feedback on the process or the issues.
—Colleen M. Grogan
Safety net providers have been characterized as providers of last resort who care for all patients, regardless of their ability to pay for their care. Safety net providers include the safety net hospitals and community health centers (CHCs) that provide a disproportionate amount of hospital and community care to Medicaid and uninsured patients. The Affordable Care Act (ACA) is dramatically reducing the number of uninsured patients, as well as some of the funding targeted for safety net providers to cover the costs associated with caring for the uninsured. If the newly insured continue to seek care from safety net facilities, this will provide a new source of revenue. However, there is no guarantee that these patients will continue to receive their care from safety net providers, despite the latter's considerable experience providing tailored services that accommodate the language and cultural diversity of this population. As a result, the safety net must evolve and adapt in the face of increased competition while continuing to serve the remaining uninsured.
The lessons from the best-performing health care systems suggest that integrating service delivery is an essential reform in bringing a variety of care improvements into alignment to maximize their potential benefits. Service integration can promote efficiency by reducing waste and duplication in services (Pourat et al. 2012. Integration also facilitates access and care coordination, ultimately leading to improved quality of care and customer service. Although safety net providers serve similar populations, delivery of care has historically been fragmented. Safety net hospitals and CHCs often operate independently of each other, resulting in little or no coordination of care and inefficient use of resources (Cunningham, Felland, and Stark 2012; Shortell and Weinberger 2012).
Policy makers have several levers available to them to strengthen integration and achieve high-value care, including payment reform, expanded use of health information technology, the removal of regulatory barriers, patient engagement and consumer activism, and performance measurement and disclosure (Enthoven 2009; Shortell and McCurdy 2010; Kizer 2013a; Goodwin et al. 2012. These critical change levers are most effective at achieving clinical integration across providers when combined and aligned (Goodwin et al. 2012. Many of these change levers were promoted in the ACA, but to a greater degree in Medicare than the Medicaid program, the dominant payer for safety net providers.
We conducted in-depth qualitative interviews with safety net executives in five cities considered among the leaders in pursuing safety net integration, to explore their successes and challenges in delivery system integration. The safety net sites were selected based on input from national experts and published materials (Mohan et al. 2013; Gabow, Eisert, and Wright 2003; Perez et al. 2013. All the sites are highly urbanized, provide a substantial proportion of care to disadvantaged populations, and receive a significant proportion of their income from Medicaid. We examine the degree to which policy levers are being used to facilitate their integration efforts. Our findings offer insights for safety net providers interested in delivery system integration as a way to transform from the provider of last resort to the provider of choice, for all their patients.
Study Data and Methods
The five safety net study sites were Boston, Minneapolis, Denver, Los Angeles, and San Francisco. Except for Los Angeles, which operates three large, general acute care hospitals, each of these cities has one safety net acute care hospital. The Community Clinic Consortia, which are member organizations for freestanding community clinics and CHCs, were included in the study for all the corresponding regions except for Minnesota, which did not agree to participate.
For each study site, data collection included primary source materials, state legislation, and public documents. Data on the characteristics of each of the public hospital systems were provided by America's Essential Hospitals (Zaman, Cummings, and Lacox 2012). We conducted site visits and semistructured interviews with the senior leadership team from January to December 2013. This included the chief executive, chief medical officer, and the executive with lead responsibility for strategy of each public hospital and the chief executive of the CHC consortium. Potential subjects were contacted via e-mail to enlist their participation in a one-hour interview to understand the constructive steps that their organizations were taking toward integrating care and payment. The subsequent interviews were conducted in person in all but three cases; the remaining three were conducted by phone.
The interviews followed a set protocol of open-ended questions that relate to the key characteristics of a high-performing integrated delivery system. The questions were created using the themes in the Safety Net Accountable Care Organization (ACO) Readiness Assessment Tool (Shortell and Weinberger 2012). The questions sought to explore the extent to which safety net providers were applying strategies to bring about clinical integration across primary and acute care in relation to policy levers identified in the literature (see table 1). Example questions include “How is the way you are receiving payment related to integrating care?” and “How has your governance structure evolved over time?”
All interviews were recorded and transcribed. Interview content was analyzed relative to the policy levers. For each policy lever, prior expectations were developed to assess the extent to which providers were adopting the types of strategies that would bring about integration if that policy lever was being applied across the safety net. The transcripts were also analyzed for emergent themes. A second investigator completed a thematic analysis of the data and confirmed the key themes. Disagreements were resolved by consensus. We present the summarized themes as well as quotes from respondents that illustrate these themes.
The study design and procedures were reviewed and approved by the University of California, San Francisco, Committee on Human Research.
The five case study sites offered a spectrum in terms of integration status and the extent to which integration is being actively pursued between safety net hospitals and their neighboring CHCs. The spectrum ranged from not integrated and not seeking to formalize partnerships to fully integrated with a limited set of CHCs in their region. Table 2 provides information on each hospital site and its associated community clinics.
Although the interviews were designed to explore the degree to which different policy levers facilitated clinical integration, the results reflect the relative weight that the interviewees gave to each strategy. Payment reform garnered the most discussion, followed by health information technology, removal of regulatory barriers, activating the consumer, and performance measurement and disclosure.
The majority of the health system leadership viewed payment reform as a powerful strategy for incentivizing providers to integrate their delivery systems. However, interviews revealed only a few examples where CHCs and the hospital were operating under a shared budget with joint responsibility for the cost of care for a defined patient population. In those sites experimenting with new payment models, this accounted for a small share of the hospitals' revenue, and the hospitals were protecting themselves against the risk of financial losses. For example, Hennepin County Medical Center in Minneapolis created a small program called Hennepin Health on a per-member, per-month rate with one of its neighboring CHCs, the county, and a health plan. The hospital reported that this program accounted for less than 10 percent of its total revenue, and the majority of the savings were being allocated back to the hospital rather than the other partners. Most of the case study sites operated in a predominantly fee-for-service environment and continued to focus on systems that optimize revenue in this payment model.
As long as we're living with some component of our care in a fee-for-service world, then the incentives are all messed up for driving up unnecessary use of technology and unnecessary procedures, and I think, in order to get a control on our costs nationwide, we need to have a much more capitated system. (Hospital leader, Denver)
The sites that expressed interest in extending shared savings through capitation to more patient groups were those that had ownership of a health plan. Caring for patients on a capitated rate was seen as a mechanism for driving down costs and keeping the institution financially afloat:
We actually own and operate our own health plan. … That's been critically important for our financial success in that we really feel confident that it is that managed care nature of a capitated health plan that has been really important for us to drive down costs. (Hospital leader, Denver)
However, in those sites that took on financial responsibility for a narrowly defined patient population, this sometimes led to a two-tier care model. Patients were flagged when they entered the hospital and were treated differently as a result:
We have our own health plan, so that's part of that integration question, so for those folks that are in our health plan we are highly incentivized to keep them out of the hospital and keep them well, because they represent, sort of, a cost to us.
We had it flagged in [our electronic health record system]; you walked through the door, we knew who you were, we knew what sort of, you know, imperatives were associated with your care. (Hospital leader, Minneapolis)
Leaders in Los Angeles particularly expressed apprehension about payment reforms because they feared that they would lead to the development of an inequitable system for their insured versus remaining uninsured population.
I don't want to have multiple tiers of care. I want to have one-tiered care. I don't want our providers or frontline staff to have to differentiate. (Hospital leader, Los Angeles)
The only site that expressed a sense of urgency in pursuing a more formal contractual arrangement with its neighboring CHCs was San Francisco General Hospital (SFGH). Starting in 2011, the state of California began to transition Medicaid Seniors and Persons with Disabilities who were previously covered under fee-for-service into managed care plans. SFGH contracts with the San Francisco Health Plan to provide services to this high-need population in a fully capitated arrangement that places SFGH financially at risk for service use at outside providers. However, SFGH accepted this responsibility without having a system in place to keep patients in-network to manage their care. As a result, SFGH leaders reported “hemorrhaging money” from patients receiving their care outside their provider network.
Interviewees reported that the most significant payment barrier to integration was the misalignment in how safety net hospitals and community clinics that have Federally Qualified Health Center (FQHC) status are paid. Community health centers that receive FQHC or FQHC “look-alike” status from the Centers for Medicare and Medicaid Services receive an enhanced reimbursement rate for patient visits using a prospective payment system methodology. This requirement applies to fee-for-service and managed care contractual arrangements. A number of the hospital leaders believed that clinics were reluctant to enter financial risk-sharing arrangements with the hospital because of the protected financial arrangement under the prospective payment system:
[FQHCs] receive disproportionate payments and it is really on a fee-for-service basis, which can be a disincentive for their contractual arrangement with us. (Hospital leader, Boston)
Community Clinic Consortium leaders concurred that the more lucrative payment arrangement for FQHCs and look-alike clinics diminishes their motivation to pursue exclusive partnerships with hospitals. Community providers that do not have access to this enhanced payment have been more interested in integrating with hospital systems:
What's funny is that the lack of resources increases the interest in the dialogue, because resource scarcity is a great motivator. It's the ones that have got the least resources that have had to be the most creative in integrating to make things happen. (Community consortium leader, Los Angeles)
Health Information Technology (HIT)
Most sites had allocated considerable resources for extending HIT within the hospital systems and the individual community clinics in their region but noted limited resources to implement information exchange across providers. Having access to the same patient data across providers was viewed as crucial for providing coordinated patient care, as well as for managing costs and ensuring accurate billing:
If we have a common platform for everything: inpatient, outpatient, ORs, non-ORs, health centers, that's the only way we'll really be able to understand where all the care is getting delivered and what the spend is. (Hospital leader, Boston)
Many of the interviewees viewed shared HIT and data transparency as a critical precursor to developing more formal partnerships. In particular, evidence that patients were being referred within network was reported as key to building the necessary trust to bring providers together. Safety net hospital leaders were skeptical of CHCs' commitment to refer their Medicaid patients to the specialty clinics associated with the hospital. Without referral data it was assumed that CHCs were not referring those patients to specialty clinics associated with the safety net hospital, and as a result these hospitals were losing revenue.
The current IT infrastructure was viewed as being better for financial rather than clinical integration. Even in the one site where shared HIT is fully embedded, interviewees report that they did not have the right kinds of data to manage populations across providers:
We took the path that most medical centers took, but it's the wrong one for population management. (Hospital leader, Denver)
Although respondents reported that they did not have the right heath information technology infrastructure to facilitate a population health focus, they regard the developing ability to track financial data across different safety net providers as an important precursor to service delivery integration and payment reform.
Removal of Regulatory Barriers
Safety net leaders perceived governance requirements at FQHCs to be a barrier to integration. FQHCs are required by federal law to be governed by an independent board drawn predominantly from the communities they serve. Therefore safety net hospitals and FQHCs must retain separate governance structures, which makes it difficult to align agendas and to build the cohesiveness to operate as one organization:
I think they're independent, so it's not a true network that we can rely on where trust is secure, so it's bidirectional tension, because we have an agenda as an independent corporation as do they. (Hospital leader, Boston)
As one Community Clinic Consortium leader noted, because the clinics' resources are tied to these federal regulations, they place a strong emphasis on ensuring that they are operating within the parameters of these rules and regulations:
The FQHCs have access to resources the other groups don't have. They also have a common base. They have a common association they can turn to. They live, eat, and breathe these regulations because their funding goes by it. (Community consortium leader, San Francisco)
While CHCs and public hospitals care for similar populations, the competition for limited resources has contributed to mistrust between the two types of organizations. It takes a great deal of effort to overcome the regulatory barriers to integration, and that effort first requires an environment of trust and cooperation. If motivated to overcome it, then it may not be insurmountable.
Denver Health has been operating as a fully integrated delivery system with the FQHCs in its region by running two separate boards of directors. Leaders at this site indicated that it requires significant investment and creativity to construct governance arrangements that meet the regulatory requirements and allow them to function in an integrated fashion. Boston Medical Center reported that after nearly two years of negotiations it was close to reaching a governance arrangement with the center's affiliated CHCs.
Securing independence from the local government was also regarded as an important step in allowing the hospital board the flexibility to pursue a number of strategies, including integration. The sites that have made the most progress in integrating care across the safety net have done so after separating the hospital board from government operational control.
Activating the Consumer
Some safety net leaders expressed a belief that more needs to be done to get their patients to take responsibility for their own care; however, none spoke of this as a strategy to bring about integration across safety net providers. In general, safety net hospital leaders reported limited experience or need to use strategies to activate consumers who typically use them as a provider of last resort. If anything, hospital leaders tended to regard consumer choice as something of a threat that could have an untoward effect of encouraging their patients to seek care elsewhere.
Community clinic leaders had a somewhat different view of patient activation and choice than safety net hospital leaders. They viewed consumer choice for their patients as a priority even if it undermines a dedicated commitment to working with the safety net hospital in their community.
Patients choose which provider they want. They're allowed to change providers any time they want. The patient makes their choice, as you know, more on the provider than they do on the plan. (Community consortium leader, San Francisco)
In contrast to hospital leaders, who viewed limited networks of providers as a critical ingredient for clinical integration, community clinics were reluctant to enter into an exclusive relationship that results in a closed network of providers. They viewed their ability to provide patients with choices for referrals as an important factor in making their own primary care site more attractive to patients. However, community clinics also perceived that integration may be an important way to attract patients:
Consumers want to see a higher level of integration; they don't understand why it's taking so long or why it's so hard. They see Kaiser as the standard model — even if they are uninsured, they are familiar with it. (Community consortium leader, San Francisco)
Performance Measurement and Disclosure
The interviews provided little evidence of data being assembled to help patients understand the quality and cost of care in the safety net. Most of the interviewees regarded performance measurement and disclosure as an important mechanism for improving quality. However, even at those sites that were most able to produce data to support their quality improvement activities, interviewees believed that it was not the right data for facilitating provider integration:
I don't think we have built great metrics that capitalize on integration. Most of the work that we have done is really around management of primary care services. (Hospital leader, Denver)
In general, interviews revealed a gap in the availability of timely measures to incentivize clinical integration. Interviewees viewed the absence of a reliable and sophisticated performance measurement feedback loop as a barrier to clinical integration as well as payment reform.
Alternative Strategies to Integration
As an alternative to, and perhaps as a distraction from, the hard work of integration, safety net providers were seeking to capture private insurance reimbursement by opening clinics in the communities where their employees and their families live, as well as in locations where traditional safety net populations that are reaching Medicare eligibility reside. One hospital leader reported:
We would like to dilute out Medicaid. We're about 40% Medicaid right now. Quite frankly, we'd like to grow Medicare. (Hospital leader, Minneapolis)
A second hospital leader illustrated the strategy that some hospitals were adopting to capture new privately insured patients and to diversify payer mix:
The other investment we would be making would be a new mission component because right now those clinics are designed for seeing the indigent patients and the uninsured patients and some Medicare, but in the new system what we are going to be looking for is going after a larger proportion of our employees and dependents and the city's employees and dependents. To do that the clinics have to be located not where the indigent population lives but where the employees and their dependent population live. (Hospital leader, Denver)
Leaders in the five communities that we surveyed described substantial challenges in the integration of clinical services between their safety net hospitals and their CHCs. Safety net leadership viewed payment reform as the most powerful lever for promoting integrated care delivery models. While there is experimentation with new payment models at the local level, the scale of these programs has been small, and it has not led to broader system transformation. Interviewees also reported that there is an absence of reliable and sophisticated clinical integration performance measures, which acts as a barrier to clinical integration and payment reform. Safety net hospitals appeared to lag behind CHCs in recognizing and responding to the importance of consumer choice. The only area where there was consistently a shared strategy was in health information technology, where electronic data sharing between safety net hospitals and community-based clinics was being used in limited cases as a mechanism to facilitate care coordination, track patient referrals, or ensure accurate billing.
Differences in policy and funding, including reforms of the ACA, exacerbate the divisions between different types of providers in the safety net. The ACA provided $11 billion in new funding to the Health Centers program. By contrast, safety net hospitals are confronted with increasing financial challenges as a result of ongoing cuts in federal and local government funding, changes in Medicaid reimbursement rates that are often below those of other payers, and continued uncertainty over the long-term future of the disproportionate share hospital (DSH) program. This misalignment in the current financial arrangements presents a challenge to full system integration. Although safety net hospitals often provide primary and community care through their own affiliated clinics, the grant-supported CHCs are being positioned by the ACA to expand primary care capacity potentially in competition with safety net hospital systems.
Changes in Medicaid, the dominant payer in the health care safety net, could encourage safety net providers to more aggressively pursue integration. Each state administers its own Medicaid program, which introduces greater variability and focus on the issue of safety net integration. However, there is an opportunity for states to incentivize safety net hospitals and community-based clinics to provide integrated care through their Medicaid programs. For example, states could consider extending the enhanced Medicaid primary care reimbursement that is supported with federal funds through 2014 with the stipulation that primary care practitioners would need to furnish services as part of an integrated system of care. This could be done in combination with restructuring the state's DSH programs to align with primary care enhanced reimbursements, including linking the allocation of DSH payments to metrics that will accelerate integration across existing primary and specialist care providers for Medicaid patients. States should also consider using their unique position to leverage their health care purchasing power for Medicaid members and state employees, to support new ACO payment and contracting models (Purington et al. 2011. For example, Colorado has established Regional Care Collaborative Organizations, and Hennepin Healthcare has now entered into the Minnesota Medicaid ACO demonstration project (Perez et al. 2013. Through payment reform and the adoption of statewide performance reporting requirements, this approach has the potential to incentivize safety net providers to come together to control costs and improve health outcomes.
The federal government also has a role to play in safety net integration, both in terms of giving states the flexibility they need to test payment strategies that encourage care integration and in reconsidering its policies for FQHCs. Although the prospective payment system helps FQHCs to expand services to Medicaid patients while continuing to provide care to the uninsured, their per diem rate is based on encounters, which continues to incentivize visits over value. This financial arrangement gives FQHCs the protection to stay fixed to their own agenda, which potentially prevents the safety net as a whole from achieving greater benefits through integration. In addition, the requirement for FQHCs to retain independent boards creates an additional governance barrier that needs to be overcome in what are already complex negotiations and can often lead to unnecessary duplication and redundancy.
Given the substantial barriers that currently exist for administrative and financial integration between safety net hospitals and community-based clinics, it may be unrealistic to expect them to form an integrated care delivery system, but it is still possible for them to find ways to coordinate their independent activities with one another. Without altering their governance structure, they could work toward achieving functional care coordination by having a common vision, shared goals, information management tools and infrastructure, policies and procedures for coordinating care, methods of accountability and performance management, aligned financial risk and rewards for clinical outcomes, and a population health focus (Kizer 2013b). Health information exchanges are an example of how separate organizations can cooperate to meet their individual needs while also supporting improvements in care integration. These exchanges provide the capability to electronically move clinical information among disparate health care information systems across organizations within a region, community, or hospital system.
This study included only a small number of sites, which may not be generalizable to all safety net providers. However, given that these health care providers are considered on the leading edge of care integration in the safety net, most likely their challenges exist among other providers as well.
Coverage among those who have relied on the safety net is changing more rapidly than the safety net delivery system itself. Early indications following the full implementation of the ACA are that at least some safety net providers are seeing a sudden increase in patients with health insurance coverage (Galewitz 2014). Whether this is short-lived or sustained over time will depend partly on whether safety net providers can navigate the substantial barriers they face in achieving delivery system integration so that they can become not only more efficient and effective but also more attractive to patients.