7

PRIORITIES FROM USER PERSPECTIVES

Virtually every sector in the nation is, directly or indirectly, a significant stakeholder in the future of health services research. Meeting participants and discussants, including, in particular, those from the policy, consumer, clinician, and payer communities, discussed priorities for the field of health services research over the next decade. These discussions highlighted important research priorities for improving health and health care outcomes as well as priorities for improving the infrastructure needed to support and transform health services research. A consistent theme was the importance of focusing on implementation science as core among the challenges and opportunities.

PRIORITIES FROM A POLICY MAKER’S PERSPECTIVE

Charles (Chip) Kahn, president and chief executive officer of the Federation of American Hospitals, identified four areas of health services research relevant to both policy and practice: coverage, quality, accountability, and transparency. Using that lens, he articulated three research priorities for the field.

First, Kahn commented that health services research has not been very successful in helping policy makers move away from fee-for-service and that, therefore, it is critically important to understand the impact of alternative payment models on health outcomes. The greatest successes remain the Social Security Amendments of 1983 and the diagnosis-related group codes. The Medicare Access and Children’s Health Insurance Program Reauthorization Act of 2015 “shows that there is a lot of work to do,” according to Kahn. “Rather than value-based purchasing—because I don’t know what value really is—I’d say we need results-based purchasing, and the field needs to tell us and guide policy makers as to what that ought to be.” A number of natural experiments are taking place, such as those presented by differences in coverage between California and Texas. “We need to understand that better,” he said, particularly “the policy implications.” Whatever payment models are developed, they need to incorporate risk

adjustment and socioeconomic disparities, he added. As previously mentioned, health services research also needs to incorporate broader social issues such as loneliness, he pointed out.

Second, while the field of health services research has done a good job of producing measures both of quality and performance, Kahn commented that there is a need for additional research that demonstrates which quality measures are critical to improving outcomes. “But we have a Tower of Babel here,” he said. Many measures have been produced and used, and many measures are required by payers, “but I would argue that we don’t really understand which measures are successful in moving the dial. We don’t have a good way of assessing gaps and then figuring out how to fill them.”

Finally, he noted that big data and artificial intelligence are going to enhance care, with both the public sector and academia moving forward in these areas. But even if some analytics are great, not all will lead to improvements, and, in fact, some may result in harm. Also, data analytics are going to influence policies regarding payment, accountability, and transparency in currently unknown ways. “It’s essential that the understanding of the implications of these new analytics be well understood,” he said. “Frankly, those who are producing them and doing great work are not the people necessarily to answer that question.” While heavy regulation is not the answer, the field needs to look at this issue from a broad perspective and make recommendations so that policy makers can understand and respond appropriately, he said. “We have to figure out how to make analytics work for patients [and] caregivers and at the same time figure out how that is also going to work in a policy context.”

PRIORITIES FROM A CLINICIAN’S PERSPECTIVE

William Bornstein, system chief medical officer and chief quality and patient safety officer for Emory Healthcare and professor of medicine in the Emory University School of Medicine, took a more granular approach by discussing what he needs to take care of patients. Type 2 diabetes encapsulates many of the challenges he and other providers face. Patients, family members, and health care providers have a belief that diabetes is brought on by inactivity and obesity, although genetic and biological factors are obviously also involved. This belief can lead providers and family members to conclude that people with diabetes either do not understand that lifestyle changes can ameliorate the disease or are not motivated to change. In reality, these patients do understand and are extraordinarily motivated to change, said Bornstein. “We fail them in this regard by not being able to help them.”

He also pointed out that the social determinants of health contribute to diabetes. Patients who live in food deserts or in neighborhoods where it is unsafe to walk face challenges in changing their behavior, “but honestly I don’t find that my more affluent patients are having tremendous success in this regard.” The barriers appear to be ubiquitous, which calls for a much better understanding than exists today of how providers can help patients change their behavior.

These issues are not unique to diabetes, leading Bornstein to suggest three research priorities for the field of health services research moving forward. The first is the need to understand the levers for behavior change. “Intuition in this regard is usually dead wrong,” he said, “and the usual incentives that we think motivate human behavior really are not fully explanatory.” Behavioral economics has provided some insight into how people change, including the idea that their behavior can be “nudged” through subtle messages, but this research is still in its early stages.

The second priority is to deal more effectively with overburden and burnout among providers. “If providers are suffering, we can’t effectively or optimally relieve the suffering of our patients,” Bornstein said. New care models may reduce the burden on providers. Additionally, electronic health records had great promise, but that promise has not been realized, Bornstein said. Providers still need better and more effective decision support. Furthermore, new kinds of errors are occurring that would not be occurring in the absence of these systems. “We need more research in terms of how we optimize clinical decision support and how we reduce overburden that’s coming directly from the confinements that EHRs have been designed with.”

Finally, building on earlier points raised by Katie Martin, Bornstein mentioned that the financial risk of care has been changing. “It’s shifting to providers, it’s shifting to patients. There are undesirable consequences of that, and there are desirable consequences. Behavioral economics can help us understand how more effectively to mitigate those undesirable consequences while retaining the benefits.”

Related to these research priorities raised by Bornstein, earlier comments from Gerard Anderson highlight the importance of future research focused on determining the most effective models of managing the challenges of high-need, high-cost patients. Robert Phillips, vice president of research and policy at the American Board of Family Medicine and adjunct faculty at Georgetown University and Virginia Commonwealth University, also suggested that there was a need for more research focused on understanding how to identify protective community factors, health risks, and resources for patients and communities before patients become high-need. One way to do this is to learn more about managing the interfaces of the ecology of medical care boxes, as mentioned by

Jack Westfall and Andrew Bazemore earlier in the workshop. How can home health care change to do things that hospitals used to do? How can patients transition into primary care to keep them from coming back to the hospital? Work at the whole-person and whole-community levels can yield more progress than carving up problems by organ system or disease, he said.

PRIORITIES FROM A CONSUMER ADVOCATE’S PERSPECTIVE

The American Association of Retired Persons (AARP) uses health services research to understand the impact of policies or interventions on consumers, observed Lina Walker, vice president of health security in AARP’s Public Policy Institute. It also uses health services research to engage with consumers and help them make better decisions about their health and well-being. With this background, she highlighted several research priorities for the field as well as several approaches for improving the infrastructure available to support health services research.

One area in which health services research falls short is in its treatment of the social determinants of health. Researchers typically rely on ZIP codes or dual status (Medicare-Medicaid beneficiaries) to estimate social determinants, but these are not sufficient as proxies, Walker said. “Social determinants include transportation, food and security, housing, income, social connectedness, and I would argue . . . that ZIP codes and dual status barely capture any of those factors.” A cross-cutting conversation is needed, she said, about how to capture, collect, and integrate data on social determinants with claims and clinical data.

The second point she made is that health services research needs to understand how to capture and reflect the value that consumers place on their choices. For example, convenience is a powerful factor among the people AARP represents. When people are moving into a nursing home, they want to find a facility that is convenient to family and friends. “Those considerations are very much a part of health care decision making. They are not outside the realm of health, and for those reasons, they have to be reflected in health services research.”

She pointed out that health services research has two obligations: to identify and focus on the issues that matter most, and to provide the information that helps people make informed decisions. A more expansive scope of health services research would advance human health and well-being by integrating and assessing across systems.

Related to this, David Balch, of the Patient Advocate Foundation, noted the importance of understanding the “financial toxicity” associated with shifting

health care costs to patients. When patients have excessive medical bills, they begin skipping meals, rent payments, utility payments, and car payments, but, eventually, they hit a “breaking point” that affects their medical care. “One way that health services research could intervene on behalf of those patients is to figure out how to have a conversation early and often about those areas and figure out what interventions can help patients manage those aspects of financial toxicity.”

Finally, Walker noted that readmissions and death are outcomes often used to reflect health care quality, but these are not necessarily the outcomes that are most important to patients. She told of a conversation with a car service driver who was in constant pain from neck surgery. A quality measure that considers only readmissions would miss his problems. The quandary is that not enough evidence is available to develop such a measure, but the lack of a measure makes it hard to develop the evidence. “How do we break that cycle?” How do we assess care match with patient goals? Some Medicaid programs and health systems are developing their own measures and evaluations, though they may not be published. Some systems have employed more patient-reported outcomes or have tried to incorporate values in decision making. Walker noted that in order to move the science and improve care, it is important to continue to support the development of and share information about patient-centered outcomes that measure health care quality from the patient and family perspective.

PRIORITIES FROM A PAYER’S PERSPECTIVE

Mary Applegate, medical director for the Ohio Department of Medicaid, noted that the field of health services research is ripe for disruption, and this observation led to her recommendations for research and infrastructure-related priorities over the next decade.

First, building on earlier themes, Applegate commented on the misalignment of incentives in academia that were previously mentioned by Blumenthal and Ferris (see Applying Health Services Research in Practice and Policy, chapter 3) - incentives which focus on publication and promotion, and do not lead to generation of the type of evidence that is needed by payers. “What I need in the Medicaid program is an inch deep and a mile wide, because, in Ohio, I have three million people I need to take care of.” On a related note, Applegate highlighted the need for incentives that encourage researchers to focus on implementation science. She pointed out that there are effective models by which payers can partner with academic researchers. For instance, academic medical centers can work with an honest broker who could be contacted by payers when they have an outstanding question that could be addressed through health services research.

The honest broker could then identify an academic medical center with the expertise and the data needed to address the question. In Ohio, Applegate has established an honest broker system that enables her to work closely with academic researchers on implementation issues.

Applegate also argued that there is a need for research that can be completed faster in order to inform critical health policy decisions. As an example, she stated that more than 4,000 people are dying in Ohio every year for reasons related to opioid use. “I need to know with predictive modeling who is at risk for misuse, dependency, overdose, and death—those are four different endpoints—and we need to know at the point in service what the deal is so that it can shape my conversation at the time that I’m actually seeing the patient.” This is already possible with infant mortality, she noted, where the EHR can provide information about who is at risk for preterm birth and whose babies are at a greater risk for infant mortality. When patients hear that their babies are at risk, they do things differently.

Applegate also highlighted the importance of designing payment policies and health care systems that engender trust. In order to do this, she pointed out that the outcome measures used by health services researchers should be coproduced by researchers, patients, and clinicians to ensure that they are relevant to end-users. Indeed, the need is so great that her system has been producing such measures on its own. The problem is that different states have created measures that are similar but not the same, “which means we haven’t capitalized on what health services research can actually do for us. This is a plea to please work directly with the people who are going to be using the results of the research.”

In general, greater cooperation and standardization among the states could lead to faster improvement in each of the states, she said. For example, some standardization of evaluation across states would be “a whole lot better than what we have now, which is an evaluation state by state.” Cooperation among states could help build links between clinical practice and public health. Standardization also could increase collaboration among private sector efforts to harness EHR data for health services research.

Other participants highlighted the importance of understanding the impact of provider consolidation on health disparities and other health outcomes and of establishing better measures for value in health care since prevailing measures were developed based on the current fee-for-service payment structure.

Applegate concluded by pointing out that innovation is fun. For example, she raised the question of whether it would be possible to use block chain technology to link datasets without having to worry about privacy concerns, and the impact of such a system on value-based purchasing. These types of innovations will require partnerships, including partnerships with information technologists,

educators, and patients, as well as a focus on implementation science. Such partnerships represent a new way of operating, she concluded, but the information they generate can improve health and save lives.

CROSS-CUTTING NATIONAL PRIORITY: LEVERAGING DATA

Drawing on conversations from earlier panels as well as the discussion that ensued following the presentations of stakeholder specific priorities for the field of health services research, additional themes emerged. The first relates to the need for organizations to cooperate in sharing and assessing the data essential to health services research, and the second relates to the need for a national conversation regarding the ultimate goals of the health care system: the focus on health.

Regarding data sharing, participants considered the prospects for pooling data from different health systems to answer questions in health services research and about health systems performance. Bornstein expressed the view that “the patients we serve and the citizens of this country are supportive of more access to this kind of data. The barriers seem to be more at the level of legislation and regulation.” Also, when issues are raised about privacy, managers tend to shy away from data sharing out of concerns about potential HIPAA violations. However, HIPAA is relatively permissive, Bornstein said, reiterating a point Brennan made earlier. “We need to take stronger positions on these things. We can do more under today’s regulatory environment, but I would [also] like to see the regulations and legislation change.”

Kahn referred to the possibility of technological solutions to the problem. Again, reiterating earlier points made by Brennan, Kahn noted that technologic advances allow for data deidentification in ways that both prevent reidentification and preserve the utility of the data. However, these advances may not solve the policy problems, because “privacy is the kind of area where the few, not the many, make policy.”

Walker suggested the need for funding streams that require collaboration among diverse stakeholders, which would provide incentives for sharing data. On this issue, Rainu Kaushal, of Weill Cornell Medicine, noted that several of the academic medical centers in New York City have been sharing data through PCORnet. “The tools are out there,” she said. “It’s the political and cultural issues” that are preventing more widespread data sharing. Joe Selby of PCORI noted that he sees clinical data being combined with health plan enrollment data or Medicaid enrollment data in individual projects, but not more widely. “If all the parties become interested, if there is something in it

for everybody, it will happen,” he said. “But I haven’t seen anything yet to make me optimistic.”

Darshak Sanghavi, of OptumLabs, commented on the large number of data-pooling activities taking place in the private sector. The resulting data are typically proprietary but are often available for researchers. Applegate agreed and noted that private industry usually has health researchers embedded within their own organizations to understand these data. “Is there a way to have a virtual community so that we have methods and some standardization so that we can better understand what’s going on?” she asked. “That’s a challenge for the field.” Added Kahn: “We have to figure out the best way to make that world work, because we’re a commercial country, and this is going to be commercial at the end of the day.”

CROSS-CUTTING NATIONAL PRIORITY: FOCUS ON HEALTH

With regard to the second theme, Sandro Galea stated that “we are long overdue for a national conversation about what we’re trying to do” in health care. If the goal of US health care is to have people die healthy, so that morbidity at the end of life is compressed, different decisions would be made about public- and private-sector research and development. “If it’s true that we’re trying to each live the longest possible life as healthy as possible, then it changes a lot of what we do. It changes a lot of the questions we ask. It changes what we invest in. [But] I don’t think we’ve had a national conversation.” On a related note, Gerard Anderson observed that the underlying factors that influence health raise very broad issues for health services research.

Michael Chernew pointed to the inevitable complications raised by the extensive cross-subsidization that occurs in the health care system. This feature of the system raises questions about authority, about who is going to do what, and about how much one group should pay for benefits, such as convenience, that flow to other groups. Some of these questions could be answered by policy makers trying to foresee the future of health care needs in the United States. For example, as more care is delivered in outpatient facilities and as technology progresses, health services research will need to change its areas of emphasis. However, that can be difficult, given the unpredictability of the changes going on in the health care system. In general, Chernew argued that policy makers should try to have a “soft touch” on delivery systems to allow maximum flexibility. “If you could give a broad set of parameters about quality and spending and allow the delivery systems and organizations to innovate underneath that, I think you’d be better

off than if you tried to prescribe very specific ways in which organizations have to produce care with very specific types of resources.”

This observation led to a discussion of the influence of health services research on the health care system. One way of looking at the field’s history is to conclude that influential papers have not always affected policy in a meaningful way. However, participants objected to that interpretation, pointing to, for example, the ways in which health services research affected the adoption of accountable care organizations or section 2713 of the Patient Protection and Affordable Care Act, covering preventive health services. In some cases, health services research has simply moved the conversation in particular directions, but in other cases, such as through the work on checklists and patient activation (Hales and Pronovost, 2006), health services research has had a direct impact on health care. Furthermore, as Anderson pointed out, it takes time for research to make its way into practice in health care.

As an extension of these themes, the panelists discussed the possibility of organizations that normally compete with each other collaborating to set priorities on issues that are important to them all in improving the health of their clients and populations. Focusing specifically on funders of health services research, Timothy Ferris, of the Massachusetts General Physicians Organization, noted that funders often differentiate themselves by looking for the next “great idea, and maybe that’s a luxury that we can’t afford.” Kahn noted that legislation sets priorities for many federal agencies, which limits their ability to establish their own priorities. Instead, the private sector, including foundations, may need to engage in priority-setting processes while the government “can do what it can do.” However, Walker pointed out that legislation is not necessarily an obstacle. “There’s always room to maneuver,” she said. “These are all overlapping sets of priorities.”

Finally, participants noted the importance of ensuring that the field of health services research communicates and demonstrates its value in the design of successful policies and in the improvement of health and health care outcomes. On a related note, several workshop participants commented on the need for health services research to make its work and findings more accessible to nonspecialists, people “sitting around their kitchen tables.” For example, Enekwechi cited a recent project she conducted on improving birth outcomes through payment policy levers in which the language of health services research and payment policies was less well understood by some of the stakeholders. “Health services researchers talk to each other, but we don’t translate very well.”

This page intentionally left blank.