6
Toward a Research Agenda for Population Health1
A portion of the workshop was set aside for facilitated small group discussions on a research agenda for population health. Participants were asked to identify research priorities based on their own experiences and what they heard over the course of the workshop discussions. Attendees divided into five groups to consider research questions and issues in one of three broad research areas:
- Research on understanding population health problems or issues, their root causes, trends over time, differences within subpopulations, and related issues
___________________
1 This section is based on the reports by Christine Bachrach, research professor, University of Maryland; Alina Baciu, senior program officer, National Academies of Sciences, Engineering, and Medicine; Michelle Frisco, associate director, Population Research Institute, The Pennsylvania State University; Amy Geller, senior program officer, National Academies of Sciences, Engineering, and Medicine; Mary Lou Goeke, executive director, United Way of Santa Cruz County, California; Marthe Gold, visiting scholar, The New York Academy of Medicine; Paula Lantz, professor and associate dean for academic affairs, Gerald R. Ford School of Public Policy, University of Michigan; Michelle Larkin, assistant vice president, Robert Wood Johnson Foundation; Sarah Linde, chief public health officer, Health Resources and Services Administration; Phyllis Meadows, associate dean for practice, School of Public Health, University of Michigan, senior fellow, Health Program, The Kresge Foundation; Lisa Simpson, president and CEO, AcademyHealth; and Steven Smith, clinical assistant professor, Pharmacotherapy and Translational Research, University of Florida. These reports were not meant to infer a consensus from the discussions, and the statements have not been endorsed or verified by the National Academies of Sciences, Engineering, and Medicine.
- Research on designing, implementing and evaluating the effect of different types of policies, services or other interventions that aim to improve population or community health and/or reduce socioeconomic and health inequities
- Research to improve the dissemination of data, research results, and evidence to wide audiences, and to improve the translation of evidence into policy and practice in ways that have a positive effect on communities and populations
Each group was tasked with listing three top research priorities to help focus their discussion, and they were asked to answer the following questions for each topic:
- Who are the primary audiences for the results or answers from this research?
- Who is most likely to conduct the desired research (e.g., academics, health systems, government, industry)?
- What resources and incentives are needed for this research to get done?
- Are there any significant challenges to producing this research, and how can these challenges be overcome?
Attendees reconvened in plenary session and the group facilitators reported on their groups’ conversations, followed by an open discussion moderated by Lisa Simpson of AcademyHealth. Highlights shared by individual facilitators are presented in Box 6-1.
UNDERSTANDING POPULATION HEALTH PROBLEMS AND ISSUES
Facilitator Marthe Gold of The New York Academy of Medicine and the City College of New York reported that her group considered the methods and infrastructure needed for research on understanding population health problems or issues. Various group participants mentioned several principles to inform researchers, including
- the need for equity subgroup analyses in all studies;
- the importance of capturing cost information;
- the need for integrated data systems that connect electronic health records, public health data, and data on social determinants of health; and
- the challenges of funding and studying multifactorial interventions.
Participants also focused their conversation on the following areas for research.
Cross-Sector Interventions
Participants discussed the need for research on interventions that span across sectors, Gold said. This would be done by transdisciplinary teams of investigators, and the audience would be decision makers in government, and funders. The research would likely be conducted with support from health foundations but also, more broadly, from foundations that are interested in quality of life, or the social circumstances of people, for example. It was suggested that highlighting the benefits of the research for other sectors would help to garner their support.
Prioritizing Resources
Some participants noted that research is also needed to determine how best to prioritize resources and efforts when funding is scarce. One model discussed by participants was how funding might be allocated if all governmental monies were considered together and a global budget developed across sectors. Gold added that this is a complicated modeling scenario. Another suggestion was a global budget that considers health and health care together, in both the public and private sectors, and how to invest in what creates health. The audience for this research was identified as policy makers at the local, state, and federal levels, and Gold added that the Office of Management and Budget would be an important audience in itself. Economists and experts in modeling would be needed for this research. Some participants also noted that decision makers from different sectors can help inform development of the research questions by reflecting on the challenges and decisions they face. Funding for research prioritized in this manner could come from foundations focused on health, or broader sources for cross-sector issues (including government funding).
Framing “Health”
The third priority for research identified by some participants was the need to evaluate the effectiveness of using the term health as a motivator for action in other sectors (non-health sectors, as well as health care partners outside of population health). Participants discussed, for example, whether framing issues in terms of health-in-all-policies was effective, or whether different kinds of language would be more effective. The audience for this research would be population health researchers and advocates, Gold reported, and the research would be conducted by communication scientists and public policy analysts. Foundations would be key funders of this work, Gold said, as they have been pivotal in thinking about broadening the notion of health and social determinants of health.
UNDERSTANDING THE ROOT CAUSES OF POPULATION HEALTH ISSUES
Two breakout groups considered the priorities for research on understanding the root causes of population health problems. Facilitator Michelle Larkin of the Robert Wood Johnson Foundation summarized the discussions by her group on research to identify the root causes of population health problems. Participants were interested in focusing the discussion on three main categories of research: employment, education, and housing.
Employment
Larkin reported that the group discussed many potential research questions and issues in the area of employment. Topics included the effect of wage levels and benefits on health outcomes; the role that labor plays in health broadly; the continuity of employment and its role in health; disparities in the labor force; participation in the labor force (unemployed and looking for work, as well as those who are not part of the labor force by choice); and job insecurity. A point was made about the need to understand the time dimension of research in this area, and the payoff for an intervention over a specific time period. This would be particularly important for funders, as well as for businesses that would look to implement job creation investments. Participants in this group also discussed the need to look at the history of federal job creation efforts, and the implications of those efforts for population health; state and local social costs of joblessness; and the evidence for developing toolkits for employers to help them better facilitate health through employment.
The primary audiences for these research questions were identified as state and local policy makers, mayors, the National Conference of State Legislatures, local chambers of commerce and business leaders, owners and investors in business, workforce investment boards, the public, and anchor institutions in the community in their role as employers. Participants discussed that this research would best be conducted by multidisciplinary teams, and the need to engage business schools and public health schools was noted.
With regard to resources, Larkin reported that several group participants discussed the need for the support of business, including individuals, companies, and associations that are starting to understand the value of a healthy and engaged workforce. It was also noted that the body of global research on employment and its effect on health is an available resource, and that there were models to learn from (e.g., trade apprenticeship models).
A variety of barriers were discussed, including existing federal and state policies around benefits requirements, and efforts by some employers to keep employees under 25 work hours per week to avoid those requirements. There are also trust issues associated with engaging business as a true partner in this type of research and translation. Other concerns raised were the overgeneralizing of findings, the funding needed, and how to identify true translators and champions from the employment sector to help move this forward.
Education
A priority identified by some participants for research in the area of education is to consider the mechanisms that connect health and education, Larkin reported. The audiences are similar to those for employment, with a focus on local school systems and anchor institutions. Key challenges noted were the entrenchment in the way education is paid for, school system equity challenges, and tax policy.
Housing
Larkin summarized that research questions under the topic of housing focused on the effect of housing conditions on health, and connecting remediation programs with health outcomes (e.g., reduction in asthma). Participants also discussed how investments by children’s health care institutions could take the form of tapping into community benefit funds to improve the housing conditions of the community. Research could also be done on the role of neighborhood quality in economic vibrancy; social connectedness and civic engagement; the importance of having affordable housing; and the implications of gentrification (i.e., when housing is no longer affordable and residents are pushed out of a community) on the health of a community. The audiences identified for these research issues were developers, zoning boards, community development, financial institutions, bankers, and anchor institutions.
Facilitator Phyllis Meadows of The Kresge Foundation and the University of Michigan reported that her group’s discussion of priority research needs also focused on the area of root causes of population health problems. Meadows noted that there were many diverse views in the group, making for a lively discussion. Some in the group felt that the root causes are already understood, while others, particularly participants who work in the social and behavioral sciences, felt there was more to be done to have a full understanding those issues. For example, much is known about the effect of poverty on health, and it was noted that the effect of capitalism cannot be discounted. An issue raised was the need to recognize that “root causes” has different meaning in different fields or to different audiences (e.g., the medical profession versus communities). The primary audience for research on root causes was identified as the practice community, which Meadows said includes policy makers and other decision makers. It was noted that the audience for research would ideally be engaged throughout the process, not just at the end when results are delivered. Specifically, the engagement of the practice community was discussed as a necessary element to really be able to move forward on addressing root causes. This includes community leaders and people who can implement the solutions identified. A few participants observed
that the workshop discussions thus far had not raised the notion of social movements, and they felt it was an important element for consideration.
The group discussed a range of challenges to conducting research on root causes, Meadows reported. Funding was raised in this group as a research need (e.g., funding levels, what is needed for dissemination, for adequately answering the research question[s]), and it was pointed out that funding which is allocated often will cover only one piece of the work. As such, Meadows explained, the research is lacking in its ability to capture the complexities of the issues and the root causes. Some in the group observed that some of the systems in place are perpetuating many of the problems. One specific area for research suggested was the need to consider how these often fragmented systems are affecting efforts in population health. Other challenges mentioned during the group discussions were the disconnection between researchers both within and across disciplines, a lack of sharing of information, the research capacity of partners, and the difficulties of determining causality. A question was raised about whether the focus should be on causality or association. A group participant suggested that very few people have a vested interest in changing existing structures and addressing root causes, and it is unclear whether there is actually a commitment to changing these structures.
Who should conduct this research on root causes depends on the situation, Meadows said. It was noted that, even though academia has the capability and capacity, academics might not always be the best ones to conduct this research. Group participants discussed the need to take into account the vested interest of those doing the research, and the potential for bias. In addition to funding, a key resource needed to move an agenda on the root causes of health is time, and Meadows relayed that the time needed for this kind of research is significant, especially when engaging community. Participants also discussed the infrastructure needed to collect, use, and share data.
In summary, Meadows said, the root causes of many population health problems are already known. What is needed to move forward is a clear definition of the problems, sustainable partnerships, and the political will. We have evidence, Meadows said, now we must agree that this is important research to do. Other participants noted the need for more public accountability, suggesting that public funds are not being dedicated toward work in population health. Updated methods and strategies might also be needed to develop a deeper understanding of these problems, and to move forward.
THE EFFECT OF POLICIES, PRACTICES, AND INTERVENTIONS THAT AIM TO IMPROVE POPULATION HEALTH
Facilitator Mary Lou Goeke of the United Way of Santa Cruz County summarized her group’s discussions on the research needed to understand the effect of policies, practices, and interventions that aim to improve population health. The group had a wide-ranging conversation across many research topics, she said, and she highlighted three main ideas that cut across the different areas of discussion.
- Measuring value—Group members discussed the need for better approaches to measuring the value of interventions that improve the health of a population, and better efforts to make the business case demonstrating the return on investment, or social return on investment, of these interventions. Goeke noted that some of the group warned against trying to monetize some of these interventions when discussing their positive effects. Participants also discussed the need for better ways to communicate the positive effects of population health interventions in general, in a simple, understandable way to those who need to know.
- Measuring impact—The difficulty in measuring multisector collective impact initiatives was also discussed, Goeke reported. Some initiatives incorporate many interventions, and it is very difficult to evaluate which interventions led to what effects.
- Raising awareness of existing research and tools—Goeke relayed a sense of frustration among some participants about the general lack of compilations of existing research, the lack of familiarity with the compilations that are available, and the lack of awareness among communities of the tools for evidence-based policies and practices they can draw upon.
IMPROVING THE DISSEMINATION AND TRANSLATION OF POPULATION HEALTH RESEARCH
Facilitator Steven Smith of the University of Florida shared the three research priorities identified by his group in the area of improving dissemination of research results and translation into policy and practice.
Framing the Message
What is the best way to frame research communication? What is the role of social media in research communication? How do we choose the “right” messenger to best tailor the message for the intended audience? The primary audiences for the results of this research, Smith conveyed,
would be researchers, academic communication offices, government communication offices, the public health sector, advocacy organizations, professional societies, and the various knowledge brokers who pass along results of research. This research would be conducted by communication scientists (Smith mentioned the FrameWorks Institute as an example). The primary resources and incentives needed would be funding, and Smith relayed that the group suggested public–private partnerships could be a good resource for conducting this type of research. A key challenge identified by some participants was the ability to maintain integrity while providing results. In other words, Smith said, researchers are often not comfortable with having to report results before they are ready, or with having to simplify the message to the point where necessary information and caveats are not discussed. He also noted that traditional models of research may be a challenge in this particular area.
Incentives for Translation of Knowledge to Practice
What is needed to incentivize researchers, and more generally, the research enterprise (including academia), to increase their involvement in dissemination and implementation research? The primary audiences for this research are academia and research funders, as well as journals, Smith said. Agencies or research funders would likely be the drivers of this research, and possibly academic researchers as well. Again, the key resource needed is funding. Smith also relayed the suggestion that funders be encouraged to require demonstration of impact. He noted that grantees usually have to describe their dissemination and implementation plan, but this is not well enforced, and many get by with simply saying they will publish a paper and present at national conferences. Forcing accountability would be a challenge, Smith noted. Another challenge identified was defining exactly how much is needed to incentivize academic researchers. It was observed that most people support the concept of dissemination of information, but far fewer actually act on it.
Connecting Researchers and End Users of Research
How can we meaningfully increase the exposure of researches to the end users of their research, and vice versa? The primary audiences for this research would be researchers and end users, but also intermediary organizations (e.g., professional societies, advocacy groups, funders). Researchers and users would be the primary people involved in conducting this type of research, Smith reported. Participants in this group highlighted the importance of funding to support the time commitment necessary to build collaborative relationships. In this regard, there could
be incentives for showing a strong level of collaboration (e.g., within academia there could be credit given for time devoted to building those relationships, beyond just an occasional meeting or letter of support from the end user or collaborators). Collaboration could also be incentivized by the accrediting bodies of academic institutions with which researchers are affiliated. A key challenge identified by the group was the time commitment for relationship building and interaction between researchers and end users at more than the surface level. It was suggested that greater value could be placed on pragmatic research and practical trials, Smith reported, and he added that some participants called for greater appreciation of the ability to translate rigorous research into practice in the community setting. One strategy suggested was to give researchers and end users a primer on the benefits to community members of the collaborative research relationship.
DISCUSSION
In the plenary discussion following the reports from the small group facilitators, participants discussed the issue of complexity, the challenges to generating political will and influencing change, and the need to better link costs and outcomes.
Complexity
George Isham, senior advisor at HealthPartners and co-chair of the roundtable, raised the concept of population health as a complex adaptive system that is situation dependent (e.g., affected by where people live and other factors). He asked whether enough attention is being given to the complex aspects of population health when considering root causes and identifying policy approaches to improve health across the country. Christine Bachrach of the University of Maryland responded that the idea that population health is a complex adaptive system does not necessarily change the questions about root causes. Rather, it emphasizes the need to pay much more attention to the interactions among root causes. In addition, it draws attention to the need to revolutionize the methods being used and the types of evidence being collected to provide answers for population health questions. As an example, she observed that the vast majority of researchers are still doing regression analysis. Simpson agreed and added that, with the explosion of data and technology, there is both a tremendous opportunity and the potential for making false associations that do not actually exist.
Simpson asked panelists for examples of complex interventions and evaluating multifactorial interventions. Where could the population
health field turn for lessons and strategies? Gold mentioned the successes in tobacco control and said that researchers are often asked to identify the one element that had an impact on lessening the use of tobacco (e.g., taxes, physician counseling, public campaigns). The answer is that a critical threshold of many different things must be achieved at the same time, she said, and funders need to understand the need to assess multifactorial interventions. Simpson questioned whether funding agencies have been receptive to this notion. A participant said that complexity science is a different way to frame how research is conducted, and that this is an area of expertise the Roundtable on Population Health Improvement might consider bringing on board. She agreed that traditional methods, such as the regression analysis approach that was mentioned, will not move the field to where it needs to go. Larkin added that generating standards of evidence for population health was also discussed in her group, including how those standards would work in a transdisciplinary research framework.
Political Will and Influencing Change
Bob Griss of the Institute of Social Medicine and Community Health expressed his opinion that the complexity issue is exaggerated, and what is really needed is a focus on the politics of the social determinants of health. He suggested that complexity is often used as an excuse by vested interests to stave off public policy and regulatory action. As an example, he cited the influence of the tobacco industry, which has stalled actions by the U.S. Food and Drug Administration to regulate menthol in cigarettes and e-cigarettes. He raised a concern that public health practitioners are “tinkering with the system” instead of addressing the root causes, and that root causes include a lack of commitment to the principles of public health and the principles of equality. He noted the importance of community engagement in driving the process of addressing root causes of population health problems. He added that, although individual health status can be a measure of inequality, real differences become apparent from comparative research between systems. International comparisons are highly relevant, he said, in the way they integrate medical care and public health into a seamless financial system. He observed that there was much discussion at the workshop about the erosion of funding for public health but limited discussion of the political movement needed to address these issues.
Simpson noted that the political will to act on the science and address the root causes of health problems was identified as a need by the group facilitated by Meadows (summarized on pages 46–47). She also referred participants to the National Research Council and the Institute of Medi-
cine report U.S. Health in International Perspective: Shorter Lives, Poorer Health, which includes international comparisons. She shared one example that ranks the life expectancy at birth of U.S. women as 16th in a set of 17 peer Organisation for Economic Co-operation and Development countries (NRC and IOM, 2013). She expressed her disappointment at how little the findings of this report have affected the conversation on health in this country. The question persists, Gold added, as to where to put our energies and our will, and what falls into the research space versus the advocacy space.
There is an argument to be made, Simpson said, for continuing to produce evidence that is current in its context. It is not that one more study suggesting the root cause of poor health is poor living circumstances will finally change minds. It is that when the window of opportunity opens for translating evidence into policy, the evidence to support that change has to be recent, Simpson said. She has heard from policy makers that the findings provided to support policy change are often too old, especially when discussing insurance and costs.
Meadows suggested that the field sometimes fails to be strategic and seize opportunities to use population health research in a way that can effect change. The root causes of population health problems have become very entrenched, and it is difficult to make any real gains on them, she said. She suggested that, in addition to new methods of how to study root causes, new methods and approaches are needed for how to influence change on these measures.
Isham raised the issue of shared value and said that a research agenda for population health needs to include implementation research on how to achieve shared value, including effective communication. He referred participants to the Roundtable on Population Health Improvement workshop summary Supporting a Movement for Health and Health Equity (IOM, 2014).
Simpson suggested that individuals and associations, such as the newly formed Interdisciplinary Association for Population Health Science, could use the outcome of this workshop to discuss population health research needs directly with funders in both the public and the private sector. The issues raised at the workshop (e.g., the need to focus on equity, development of new methods, understanding costs) will not be addressed unless those paying for research start asking for it, Simpson said.
Costs and Outcomes
Jean McGuire of Northeastern University suggested that some of the resources and obligations for achieving population health are within the health care industry, and that research is needed on the role of the
health care industry. The health care industry needs to be able to drive and invest in population health, she said, but it faces a mixed incentive and even a disincentive to do so. She referred to the comments by Ron Pollack of Families USA about the potential negative effects of increasing copays and other consumer-level costs on an individual’s ability to access care, and the question raised by Jenelle Krishnamoorthy of Merck & Co., Inc., about the cost effect of successful population health initiatives on entitlement programs (e.g., individuals living longer has increased costs for Medicare). Simpson agreed and repeated the earlier comment by Gold that cost dimensions need to be included in studies. A challenge is that return on investment can occur in a different sector; for example, some of the benefits of better asthma care for children are returned to the schools in the form of decreased absenteeism.
Bachrach noted that accountable care organizations are taking action to improve population health, and research is needed to determine the extent to which these actions are affecting upstream determinants of health. There is a need to understand the outcomes of organizations that are still focusing at the individual level relative to those that are addressing upstream determinants of health. Simpson added the need to understand what initiatives are working, or not working, in different sectors so that time and resources are not wasted repeating unsuccessful strategies.
A participant highlighted the gap between health management and population health and suggested the need for research on bringing these functions together. Provider systems need to know the impact of their work. Another participant emphasized the need to consider the effect of behaviors on health (e.g., tobacco use, physical activity, nutrition), and the interactions between behaviors and social determinants of health.
CLOSING REMARKS
The Roundtable on Population Health Improvement has been using six categories of drivers of population health improvement to organize its work: metrics, resources, relationships, policy, communication, and research/evidence, said David Kindig, professor emeritus at the University of Wisconsin School of Medicine and Public Health. Not all of these areas require new research, and there is much that can and should be done based on what is already known, he said. However, as noted during the workshop, he asserted that there are areas where new evidence is needed to move population health forward, particularly relative to metrics, resources, relationships, and communication. Research does not make policy, Kindig continued, but it is extremely important in the policy process. There is a need for increased funding and activity on policy-relevant research that can move the population health field forward, such
as the research priorities identified by participants during the workshop discussions. In closing, he referred participants again to the framework presented by Paula Lantz of the University of Michigan (see Table 2-1) and to the results of the survey on priorities conducted by the planning committee (see Chapter 4 and Appendix D) as resources for researchers and organizations to identify priorities for population health research and to inform the path forward.
