3
Health Care System Investments in Population Health Improvement
The first panel of the day focused on two principal mechanisms that could provide financial support to health systems for improving population health. One mechanism is derived from the Internal Revenue Service (IRS) community benefit requirement, which calls on not-for-profit hospitals to provide a benefit to their communities commensurate with their tax exemption. A second mechanism involves redesigning the payment system and using health system financing to increase the service breadth of health systems as well as their connections with other actors.
Hospitals and health systems can be an important source of funding for population health programs, largely as a result of the community benefit provisions in the federal tax code, which were redefined by the Affordable Care Act (ACA). Four speakers discribed several issues regarding the funding that the health care system can provide for population health. Kevin Barnett, senior investigator at the Public Health Institute, provided an overview of the subject and spoke about opportunities to spread and scale the types of investments being undertaken by different health systems in order to direct them in a more strategic direction that will likely be more effective. He also discussed some of the policy tools that are available to promote such investments. Reverend Gary Gunderson, vice president of Faith and Health Ministries at Wake Forest Baptist Medical Center, and Teresa Cutts, associate professor in the Department of Social Sciences and Health Policy at Wake Forest School of Medicine, then gave a joint presentation on the lessons learned from both Stakeholder Health, a coalition of mission-driven health systems, and a community-driven program in Memphis, while Valerie Agostino, senior-
vice president of health and housing operations with Mercy Housing, discussed the role that public housing can play as a partner in improving population health. A discussion, moderated by workshop planning committee member Debbie Chang, vice president of policy and prevention at Nemours, followed the presentations.
OPPORTUNITIES, CHALLENGES, AND PRIORITIES
Over the course of his two-decade career studying the charitable obligations of tax-exempt hospitals, Kevin Barnett has seen a steady movement from what he characterized as random acts of kindness to a more strategic approach to investing in population health. Although many hospitals have what he views as outstanding practices regarding their investments in population health, many, if not most, hospitals are still early on the learning and action curve, he said. “Our challenge is: How do we bring those hospitals along in the context of the reforms and the kinds of transformation that we want to see?”
Help addressing this challenge comes from data that are now available from the IRS, thanks to the Form 990 Schedule H reporting requirements on community benefits. Barnett said that while these data are important for federal policy considerations, there is emerging evidence that it is local and regional stakeholders who will find the data particularly valuable because they allow them to determine what hospitals are doing in terms of providing community benefits and advancing the public’s health. The data may also help researchers better understand the relationship between reimbursement shortfalls and community benefit efforts.
Barnett’s work has focused on the relationship between the location of hospitals and the community benefits that they provide. He noted that research going back to the 1980s has clearly shown that a hospital’s location is a major determinant of the funds that are available for community benefit purposes. Hospitals in more affluent areas tend to be better off than those in less affluent areas in terms of patient volumes and also in terms of the percentage of services that they provide that are less than optimally reimbursed. This means that the hospitals in the poorest neighborhoods have less money to invest in the community after covering uncompensated care. This creates an inequitable distribution of community benefit resources that can be used to improve community health. “That is a significant issue as it relates to the social determinants of health,” Barnett said.
One important issue that hospitals need to address is how they define “community.” The IRS encourages hospitals to use their geographic service area as the starting point for creating that definition. “We know
through a variety of sources, though, that there may be inconsistencies in the way that a hospital defines ‘community’ for community benefit purposes,” Barnett said, adding that this is particularly true when there are geographic concentrations of health disparities that are not in a hospital’s immediate backyard. In his experience engaging with representatives of hospitals, they believe that concentrations of disparities that are not in their immediate vicinity are outside their sphere of responsibility.
One of the tools that Barnett and his colleagues have developed to help local health departments and critical access hospitals—particularly those in rural areas where resources are limited—is what he calls the “vulnerable populations footprint map.”1 These maps can show which hospitals are located more closely to more affluent areas and more distant from where the concentrations of poverty are high. In one study completed for the Centers for Disease Control and Prevention, Barnett and his colleagues examined the community needs assessments for hospitals located in 15 randomly selected sites that had sub-county areas where at least 40 percent of the population was under the federal poverty level and 40 percent of the population had not completed high school. Their analysis of how these hospitals defined the community and its needs showed that less than one-quarter of the 44 hospitals studied identified the areas of concentrated poverty and health disparities in geographic terms, while one-third identified health disparities using racial or ethnic terms instead of location. “We cannot say at this juncture whether or not that was inadvertent or there was intent behind it,” Barnett said, “but it highlights the need to begin to focus these efforts and to be more thoughtful about how the subsequent implementation strategies are designed to address these issues.”
Another aspect of the IRS reporting rules for Form 990 Schedule H is that they require hospitals to consider input from community stakeholders when developing the community health needs assessment. The reporting rules do not, however, specify how community stakeholders should be engaged in setting priorities and how hospitals should determine where those priorities fit into planning or implementation processes. In fact, Barnett said that he and his colleagues have found that the priority-setting processes are generally poorly designed and implemented. What often happens, he explained, is that priorities are framed so broadly that it allows for the perpetuation of existing programs and leads to a lack of focus on the geographic concentration of health disparities.
There are a number of public policy tools that could be used to address some of these shortcomings. One example that Barnett suggested would be payment-in-lieu-of-taxes programs as exemplified by
_________________
1See http://assessment.communitycommons.org/Footprint (accessed July 24, 2014).
the Pennsylvania Community Benefit Law, which allows hospitals to get a tax credit equal to three times their cash contribution to a general fund as an alternative to investing in community benefit. Barnett said this type of program may be problematic because the money rarely gets allocated in a targeted manner toward health and prevention. An example of a different approach is Massachusetts’s Determination of Needs program, which requires hospitals, when they construct new buildings, to invest a portion of money in prevention-related activities. Similarly, a growing number of areas are requiring that hospitals sign Community Benefit Agreements that include an obligation to address disparities in geographic regions that extend beyond a hospital’s immediate neighborhood. The recently created Los Angeles Wellness Trust, for example, requires hospitals to contribute 1 percent of their general operating revenues into a prevention trust fund.
Barnett said that a problem with all of these programs is that they are essentially check-writing exercises that leave the hospitals on the sidelines. “Our challenge and our opportunity going forward is to make sure that hospitals are part of the process and part of the transformation,” he said. Hospitals need to be at the table, working as partners with the community to build the necessary population health capacity that can move the agenda forward. Bringing hospitals to the table as partners will require that hospitals and health systems move from a compliance orientation to one that focuses on transformation. Barnett described guidelines developed by Sara Rosenbaum and her colleagues at George Washington University that describe how different actors define community, thus helping to characterize the differences between these two orientations. For example, IRS regulations define community as a hospital’s service area, and hospitals are required to identify underserved populations and develop programs to address disparities at the service area level. A broader, transformation-focused orientation would identify geographic concentrations of health inequities that fall within a larger region so that hospitals focus their resources where those needs are greatest. Barnett said that the dialogue with hospitals has to be about more than how they comply with what they are required to do at the federal level. It should also be about how they can work together to ensure the economic survival of the hospital particularly as health budgets undergo change in the future.
In a brief comment on the issue of data pooling and data sharing, Barnett said that a recent study in California assessing the roles and contributions of community health workers and promotores (lay health advisors drawn from the Latino community) found that almost none of the community health centers could readily identify how these workers contribute
to advancing the Triple Aim.2 “The core reason was that [the centers] lack analytical capacity and did not have access to hospital utilization data,” Barnett said. As a result, the community health centers could not begin to understand the total cost of care and what their return on investment was. “This is a very specific area where we need to be thinking about alignment,” Barnett said, and he noted that the same issue is a concern in the broader area of health care workforce development. “We have each of our hospitals in any particular region looking at what their immediate needs are in terms of their frontline workforce,” he said, “but there is almost no funding available for regional approaches and for an infrastructure that will support an approach that optimally leverages the resources of the hospitals and the clinics and the other providers in those particular areas.”
The fact that most hospitals still look at community benefit as a compliance issue gets in the way of how they think about focusing their resources, Barnett said. Another obstacle is that often local leaders of health systems do not have a good idea of what population health is. One approach to overcoming this barrier to progress is to get the leaders of local hospitals to talk to one another about how to collaborate and co-invest in a specific neighborhood or set of neighborhoods. In the past, such conversations were unlikely because cooperation was bad business, given the large number of uninsured in these underserved neighborhoods—a situation that may change as more people enroll in insurance programs or are covered by Medicaid under the ACA. Unfortunately, Barnett said, the results from his community benefit study showed that most hospitals focus their efforts on clinical care, rather than social, economic, or physical environment factors.
Barnett listed three priorities for moving forward. First, he said, it will be important for the population health practice and research community to broadly disseminate the growing volume of exemplary practices and the tools to support local accountability and engagement across sectors and institutions. Barnett credited the Trust for America’s Health with taking the lead in this area. Second, it will be critical to frame the problem in a way that appeals to hospitals. “If we want to work together with hospitals, we cannot start the conversation with ‘How do we get into your pockets?’” Barnett said. “It has to be about how can we work together to solve this issue, to find a way to help hospitals leverage the limited resources that they have and to build an ethic of shared ownership in the community.” Finally, he said, the field needs to focus on place. “There is no excuse any longer for the kind of fundamental disinvestment we have had in our low-income, predominately minority communities,” he stated.
_________________
2The Institute for Healthcare Improvement introduced the concept of the Triple Aim in 2006, and it has since been adopted by many health care organizations and also adapted for use in the activities of the Centers for Medicare & Medicaid Services as the Three-Part Aim.
In concluding his presentation, Barnett cited two examples of health systems—Dignity Health and Catholic Health East Trinity Health—that have created what are in essence community health divisions that make very low interest rate loans of $1 million to $2 million that can be used to address the front-end risk that often prevents banks from investing in particular communities. “We are looking at ways to get other health systems to emulate these kinds of investment strategies,” Barnett said. The key issue here will be to make sure that such strategies become integrated into a balanced portfolio of investments made by the broad spectrum of stakeholders.3
HEALTH CARE SYSTEMS AS PARTNERS IN THE TRANSFORMATION OF COMMUNITY HEALTH
Moderator Debbie Chang asked the audience to consider three questions while listening to the next two panelists: How can effective investments be spread and scaled? What tools are already available to do this work? Who else needs to be at the table? Chang described how Nemours, as a children’s health system and operating foundation, expanded its view of “its” population to the entire state of Delaware and how that shift has transformed every dimension of its work. Others around the nation are engaging in similar work, including the three panelists.
Stakeholder Health is a group of 43 mission-driven health care systems, including 36 nonprofit health systems that understand their mission to include population health improvement. For the past 2 years, Gary Gunderson said, senior staff members from a number of those systems, known collectively as the Health Systems Learning Group, have focused on the question of whether it is possible to succeed at that mission in the context of a policy framework that is still under construction, if not actively contested or sabotaged. “Our interest is not focused on whether we can meet the legal requirements of community benefit, as all of our participating systems are in compliance,” Gunderson said, “but we are not satisfied at all with what those funds, which are spent almost entirely on emergency room care, are achieving.” Gunderson said that, much as in the high-profile case of Jeffrey Brenner’s “hotspotting” effort in Camden, New Jersey, which provides intensive social supports to help keep vulnerable patients from being readmitted to the hospital, Stakeholder Health focuses on keeping highly vulnerable patients out of the hospital by taking
_________________
3For additional information on the Dignity Health grants, see http://www.dignityhealth.org/Who_We_Are/Community_Health/STGSS044512 (accessed July 11, 2014). For more on the Catholic Health East Trinity Health Investment Program, see http://www.trinity-health.org/documents/2010AnnualReport.pdf (accessed July 11, 2014), p. 24.
steps to address nonclinical needs that are linked with health outcomes. The Health Systems Learning Group, in close collaboration with the White House and the U.S. Department of Health and Human Services, has produced a monograph that includes a number of examples of successful models and case studies (Health Systems Learning Group, 2013).
One discovery to come from the learning group’s discussions is that health care systems have unique and timely data on the most vulnerable patients. Two members of Stakeholder Health, Dignity Health, and Loma Linda University Health, have shown other members the value of using social determinants data located on maps as a tool to guide their investment decisions, so they more carefully direct their health care assets in a manner that considers broader needs as articulated by their community partners, as they work together to transform health in their communities.
Focusing on the funding implications of the learning group’s findings, Gunderson said that the case for proactive engagement with the neighborhoods that are most vulnerable to health challenges is simple. “The largest single line item in almost any not-for-profit health system is charity care,” he said, with the second-largest item being bad debt resulting from high, uncollectable co-pays common in the new bronze insurance plan. Those two items are unmanageable once someone “crosses the sidewalk and becomes a patient,” Gunderson said. “The only place for us to engage these financial challenges is on the other side of the sidewalk.” The deliberations of the learning group suggest that it is possible to better engage people before they enter the health care system but that doing so will demand an entirely new set of competencies and practices that are unusual for health care systems.
The key insight to come from the learning group and from the experience of Stakeholder Health’s members, Gunderson said, is that it is possible for private health care systems to be significant partners in transforming the health of their communities by embracing an ensemble of practices. Although this ensemble is quite straightforward and logical, it represents a major change from the way things are done now. First, he explained, it requires being focused on place and considering the most socially complex people who at different times in their lives may be patients in their socially complex neighborhoods.4 Second, the ensemble of practices works best in large-scale partnerships that are focused on place-based tactics and that prioritize the most vulnerable neighborhoods.
_________________
4Social complexity is a way of articulating the complicated range of factors such as social and physical environment that produce health and reduce the risk of premature death. For more information on socially complex persons in socially complex neighborhoods see http://stakeholderhealth.org/wp-content/uploads/2013/09/HSLG-V11.pdf (accessed July 11, 2014), Chapter 5.
The third—and perhaps biggest—challenge will be to spend funds proactively rather than reactively. “Proactive mercy would be more decent and in fact far cheaper than reactive charity,” Gunderson said, “but the reality is that the vast majority of community-benefit funds are currently frozen quite solidly in reactive paradigms.”
For the purposes of reporting to the IRS, charity care and bad debt are separate line items that show up in different places in the financial accounts, but the reality is that they are indistinguishable in practice, Gunderson said. He also commented, as someone whose hospital is directly involved in charity care, on the perversity of a situation in which the biggest component of charity care involves emergency room costs associated with the poor and uninsured, but most hospitals—including his—focus a good amount of their marketing efforts on attracting patients to their emergency rooms, which are profitable for services provided to the insured and which represent a key transition to the hospitals becoming even more profitable through providing in-patient services. Stakeholder Health suggests, though, that health care systems can be important partners in community health despite this perversity. The key factor is that emergency room-based charity care is extraordinarily costly compared to proactive, place-based tactics. “Anything on the streets costs pennies compared to the dollars spent inside the walls,” Gunderson said, “but communities are very large places and consume a great many pennies if they are not highly focused in their efforts.”
One way to focus these efforts is to note that the recipients of charity care predictably and consistently can be found in certain locations and the data possessed by hospitals may be used to clearly identify the neighborhoods and even the streets where these recipients of charity care live. Hospital data can be mapped in real time to guide proactive strategies with great precision in terms of the range of services needed by people to address the social, biological, psychological, and spiritual components of their health, in part by building networks of partnerships and community webs of trust, Gunderson said. More importantly, he said, these data can be used to reveal to nonhospital partners the dynamics driving the behavior of patients, and these partners can then contribute additional intelligence that can, in turn, create powerful feedback loops that build trust and align networks.
As an example of such a process, Gunderson discussed some of the lessons learned from a population health–based, place-based strategy called Wellness Without Walls being carried out in Memphis. There, some 500 congregation-based partnerships are having a positive effect on the health of their communities—not by just simply informing congregations of the availability of clinical services or even public health services, but by combining the knowledge of the congregations about the journey of life
of their own members and neighbors with the knowledge of the medical professionals about the disease and injury conditions experienced along that journey. “At a very basic level, the congregational networks seem to be making it more likely that patients from their networks show up at the right door at the right time ready to be treated,” Gunderson said.
Perhaps the most counterintuitive lesson from the Memphis experience, he said, is that when the most vulnerable are invited—and Gunderson emphasized the word “invited”—through networks of trusted relationships, they are more likely to accept that invitation into the health care system earlier, which results in a lower disease burden and lower cost of care. Gunderson noted that though the number of patient encounters may increase, the average cost of those individual encounters decreases by an even greater degree. “That is where the cost is saved, not by restricting access, but actually by inviting access at earlier stages when it is likely to be useful,” he explained.
In discussing some of the details of the Memphis program, Teresa Cutts said that in 2010, the baseline year for the program, most of the charity care in Memphis was concentrated in a few zip codes and, in particular, in a specific neighborhood. It also turned out that the majority of the partnering congregations were in the areas where most of the charity care was concentrated. In 2011, charity costs had jumped, but when the program launched in 2012, the cost of charity care dropped by 7 percent from baseline and almost 9 percent in the targeted hotspot neighborhood.
A key to the success of this program was having “health navigators” on the ground in specific neighborhoods rather than being based out of the hospitals. The navigators were responsible for getting people into the health care system and making sure that they received the proper care from the proper providers. “We found that this has been very useful,” Cutts said, “because with many of our folks there were trust issues that kept them from using existing safety net clinics and other resources.”
Gunderson said that the Memphis program illustrates a key finding from Stakeholder Health—that an ensemble of practices sustained by enduring partnerships built on trust and focused on mercy and justice can improve access and thereby lower the overall cost of care in vulnerable neighborhoods. Most important, he said, this is not something that can be done to the community, but it can be done with the community, and it demands that health care systems and their public health agency partners change decades-old practices and learn to practice the art of humility that is required for the best working partnerships.5
_________________
5Details about the Congregational Health Network’s Memphis program can be found at http://www.innovations.ahrq.gov/content.aspx?id=3354 (accessed July 11, 2014).
One of the lessons that Gunderson and his colleagues learned when they implemented this model in their North Carolina communities is that many of their own hospital employees lived in the most vulnerable communities. “We came to understand that we must cross-train some of our own employees as community health workers,” he said, and, in fact, those workers have named themselves “supporters of health.” Gunderson said that he expects that in the most critical census tracts there will be an increase in low-acuity care and a decrease in high-acuity care, which will reduce overall costs.
THE ROLE OF AFFORDABLE HOUSING IN POPULATION HEALTH
Over the past 2.5 years, Valerie Agostino said, Mercy Housing has been thinking about affordable housing as health care and about organizations that provide affordable housing as viable health care partners. This journey into the world of health care has been frustrating because the health care world is complex and undergoing major changes. “I think the health care world is not quite ready to embrace affordable housing as a true partner yet,” Agostino said, “but I think we are on the brink of something happening in this area.” In her opinion, there is an opportunity for affordable housing providers to engage in well-defined program models that are tied to community wellness in their properties and that have them partner with health care providers to ensure that their tenants have good access to care and good adherence to care plans.
Currently, Mercy Housing is working on developing service models that would provide reimbursement for some of those services, but over the long-term Agostino would like to be able to partner with health care organizations to provide more units of affordable housing as a component of population health. Health care organizations, however, have not found the connection between affordable housing and population to be compelling enough to be interested in forming partnerships, she said. As a result, she and her colleagues are working to reframe their arguments around the fact that the value of some of the services provided in an affordable housing setting is high.
Agostino also discussed the Mission Creek project, an affordable housing community in the Mission Bay neighborhood of San Francisco. Located in an area that was previously a rail yard and is now a vibrant community, the 140-unit Mission Creek building is adjacent to public transportation, a multi-use recreational path, and a supermarket, and it houses a library and numerous indoor and outdoor activity spaces with natural light. Mission Creek has an adult day health care center on the premises, and it has a strong partnership with the San Francisco Depart-
ment of Public Health. What makes this facility particularly noteworthy is that there is access to a physician and a licensed case manager present at all times and that the residents have created a caring and engaged community. The residents are all low-income, frail seniors who were referred by the Department of Public Health and who came from other facilities or were previously homeless. Agostino said that her physician partner in the department of public health claims that there is a $30,000 savings for each participant living in the building, which is due to a reduction in hospitalizations and skilled nursing stays, but Agostino does not believe that the data are robust enough at this point to make that claim.
She acknowledged that this approach is constrained by the fact that funds to build public housing are in short supply and that Medicaid does not currently pay to support housing as an alternative to providing skilled nursing. However, Agostino said she is hopeful that the shifts occurring today in health care will provide further opportunities for funding affordable housing.
Debbie Chang asked how to better engage payers as partners in linking clinical care to population health, given that reducing hospital admissions can be a disincentive to engage in population health, because the payment system does not reward health systems for such reductions. There were several suggestions in response. Gunderson pointed out that charity care creates a context for innovation, because when a hospital is spending its own dollars, it is highly motivated to make its investments more effective. Cutts added that, in Memphis, Cigna became interested in being a partner in community engagement when it realized that the specific zip code being targeted had the highest inappropriate use of emergency room services in the nation and that Cigna therefore “had a dog in the fight.” Agostino offered that traditional health care payers may be more open to the concept if a pilot project—funded by the public health department, for example—results in promising data. Barnett said that while payers have to become partners in these efforts, today many of them will not even consider getting involved. “We have to find a way in the broader community to bring the payers to the table and find some way to light a little fire to get them engaged in this dialogue,” he said. He added, though, that in the California community health work study that he discussed, only two entities were able to demonstrate a significant return on investment, and both were health insurance plans who had the necessary claims data.
Chang then asked each panelist to name one breakthrough action that would increase the spread and scale of population health initiatives.
Gunderson replied that he would treat health care marketing like tobacco marketing. “There are social phenomena that we are allowing to be driven by marketing dollars,” he said, “and it is not just pharmaceutical companies that are doing it.” Cutts said that she would like to find ways to better share risks and potential savings among all of the stakeholders, particularly those working at the frontline of community engagement such as the pastors of the Memphis congregations. Agostino and Barnett seconded that idea, and Agostino added that she would like to see silos broken down so that all of the different players in a given community would operate with the interest of the community in mind rather than their own self-interests. Barnett said that he would like to see ways of getting more data to the local level as a way of creating a broader coalition of partners.
Paula Lantz of George Washington University shared a concern that health care systems and payers use the term “population health” to refer to medical care, instead of viewing it from the perspective of the social determinants of health (i.e., considering underlying, high-level factors such as income and education). Gunderson echoed that concern and said that changing the mindset will be a great challenge. Mary Pittman of the Public Health Institute asked what characteristics are most likely to make a health systems leader amenable to the population health mindset. Gunderson responded that humility is a key characteristic because the leader has to acknowledge the complexities and be willing to learn through experience, and he added that this quality cannot be taught. Furthermore, the leader must already believe that money is not being spent to its greatest potential. Cutts added a leader must be willing to take risks, and Barnett said that people with these qualities and mindsets do exist in hospital systems, but they are not necessarily the top decision makers. Andrew Webber of Maine Health Management Coalition told the workshop that incremental changes are occurring in Maine, where chief executive officers of health care systems are investing in primary and integrated care and have an understanding of the business incentives of population health management.
