What’s Foundational to Population Health? Using Research to Inform Post-Election “Keep or Kill” Decisions

Post-election policy debates concerning the future of the Affordable Care Act focus heavily on the 20 million people who gained health insurance coverage under the law. But the future of less-visible components of the law – especially its population health provisions – may be equally important to the nation’s long-run health and economic wellbeing. Which provisions should be scaled up or down as U.S. policymakers revisit health reform over the next four years?

The ACA stimulated a surge of new interest in the science and practice of improving population health. The aim of this work is to improve health status for large groups of people and entire communities, rather than doing so one patient at a time through an expensive and fragmented medical care system. This work includes upstream strategies for keeping people healthy in order to reduce the flow of patients into the medical system with preventable health conditions. The ACA established population health improvement as a central component of the Triple Aim strategy on which many of the law’s payment and delivery system reforms are based. Since then, hospitals, health systems, insurers, pharmaceutical manufacturers, information technology companies and large employers have created new organizational units devoted to population health. And public health professionals and preventive medicine specialists also claim large swaths of territory in the population health frontier.

Despite this enthusiasm, population health strategies remain controversial due to their heterogeneity and lingering uncertainties about their effectiveness. Over the past 6 years, a wide array of population health initiatives have been developed and tested, including those supported by the ACA’s Prevention and Public Health Fund and the Center for Medicare and Medicaid Innovation. What is lacking in the emerging field of population health is a coherent description of the essential ingredients, backed up by sound research. What capabilities does every community need in order to succeed in improving health status for the population at large? A 2012 National Academy of Medicine study called for additional research to define this set of foundational capabilities and to quantify the resources required to establish these capabilities in every community.

This month the journal Health Affairs published results from one of our new studies that sheds light on this question by combining information from a long-running national survey of communities, “hard” data on health outcomes, and strong econometric methods. Specifically, our study examines whether certain combinations of population health activities lead to improved health outcomes when implemented over time by community organizations.

Measuring Population Health Activities

We followed a national cohort of more than 300 communities over a 16 year time period to examine the extent to which community organizations work together in implementing a set of activities designed to improve health status in the community at large. The population health activities measured in this study are based on practices long recommended by the National Academy of Medicine and other scientific and professional advisory groups. These activities include conducting regular assessments of health status and needs in the local area, developing shared priorities and plans for health improvement, educating community residents and leaders about health priorities, investing resources in shared health priorities, and evaluating the results of these investments. These activities are reflected in many current national guidelines and model practices for health improvement, including CDC’s Community Health Improvement Navigator program, the Robert Wood Johnson Foundation’s Foundational Public Health Services framework, and the U.S. Department of Health and Human Services’ Public Health 3.0 framework.

Prevalence of Population Health Activities in U.S. Communities

We used cluster analysis methods to classify communities into comparison groups based on the spectrum of population health activities implemented in each community and the constellation of organizations engaged in implementing these activities. Communities that implement a broad spectrum of population health activities through dense networks of collaborating organizations are classified as comprehensive delivery systems for population health activities, while the remaining communities were classified either as conventional or limited delivery systems based on their spectrum of activities and collaborative networks.

Estimating Outcomes Attributable to Population Health Activities

We linked survey data on population health activities in more than 300 communities with county-level health outcome measures indicating deaths from potentially preventable conditions during the period 1998 to 2014. We also linked these data with a rich set of measures reflecting community demographic, socioeconomic, and health resource characteristics. Collectively, these data allow us to track how the spectrum of population health activities in each community changes over time, and how these changes relate to health outcomes. Using the method of instrumental variables analysis, we generate causal estimates of how changes in population health activities impact community mortality rates after adjusting for other determinants of health in the communities.

By carefully analyzing data spanning 16 years, the results showed that deaths from preventable causes such as cardiovascular disease, diabetes, influenza and infant mortality declined significantly among communities that implemented a broad spectrum of population health activities through dense networks of collaborating organizations. Preventable deaths were more than 20 percent lower in the communities with the strongest networks supporting population health activities, compared to communities with less comprehensive networks. These differences in mortality persisted after controlling for a wide range of demographic, socioeconomic, and health resource characteristics in the communities, including using causal estimation methods that control for unmeasured community differences.

Drawing Conclusions about What Is Foundational

These results give us the clearest picture yet of the health benefits that may accrue to communities when they build broad, multi-sector networks to improve population health. Achieving better outcomes in this study was not simply a matter of implementing widely-recommended activities involving assessment, planning, priority-setting, resource deployment, and evaluation. Communities needed to implement a broad spectrum of activities and engage a full range of partners in these activities in order to reduce mortality. The results suggest that strong collaborative networks help communities arrive at the best decisions about how to invest limited resources in high-impact health solutions.

The population health activities examined in the study include those now incentivized through the federal Affordable Care Act and related health reform initiatives. Tax-exempt hospitals are required to conduct community health needs assessments in their local service areas, develop community health improvement plans, and report annually on their expenditures related to community benefit activities. And state and local public health agencies are required to undertake similar activities in order to meet voluntary national accreditation standards. The communities that achieved significant reductions in mortality in this study, however, progressed beyond health assessment and planning activities to include shared investment of resources along with monitoring and evaluation activities.

Perhaps most importantly, the communities that achieved sizable reductions in mortality appeared to do so by engaging broad networks of organizations in implementing population health activities rather than relying on independent and uncoordinated efforts. The network effects appear to be major drivers of these results, which are consistent with a growing body of research indicating that community networks can function as force multipliers.

Taken together, these results suggest that foundational ingredients of population health activities include mechanisms that:

· Engage dense networks of stakeholders across the medical, health and social sectors

· Support recurring cycles of community health assessment

· Develop shared priorities and plans for health improvement

· Educate community residents and leaders about health priorities

· Invest resources in shared health priorities

· Monitor and evaluate the results of investments

We recommend that these key ingredients receive careful consideration in the post-election “keep or kill” policy debates about the future of ACA’s population health components.

What Can We Say about Cost and Value?

A complete understanding of the value of population health activities to society requires estimates of both benefits and costs. Estimating the resources required to implement population health activities is beyond the scope of our Health Affairs study, which focused only on estimates of health impact. Nevertheless, post-election “keep or kill” debates about ACA and population health are likely to focus heavily on costs as well as benefits.

We recently conducted a related study to generate first-generation estimates of the costs required to implement Foundational Public Health Services as recommended by the National Academy of Medicine and as defined by a Robert Wood Johnson Foundation supported expert panel process. The services defined and measured in this new costing study do not perfectly match the population health activities measured in our Health Affairs study, but the two sets of measures are closely aligned conceptually. As a result, it is instructive to compare the health impact estimates from our Health Affairs study with the cost estimates from the costing study.

Results from the costing study indicate that full implementation of foundational population health activities in every U.S. community would require $26.3 billion per year in 2016 dollars, a total that is about 70% larger than the estimated $15.4 billion currently spent on these activities in 2015-16. Combining these cost estimates with the mortality reduction estimates produced in the Health Affairs study (and making some conservative assumptions about how mortality reductions translate to life expectancy gains), we estimate that the cost per life-year gained through investments in population health activities is on the order of $7,200 – a number that implies foundational population health activities are potentially a very wise investment.

Of course these estimates of value are very crude back-of-the-envelope calculations, requiring further analysis and refinement. Our research team along with many others are continuing to study these issues with the hope of informing ongoing decisions in health policy and health systems. Our research is part of the new Systems for Action research program created by the Robert Wood Johnson Foundation as part of its national action framework for building a Culture of Health. Based at the University of Kentucky, Systems for Action supports research that evaluates mechanisms for aligning medical care, public health, and social services in ways that improve health and wellbeing.

I invite you to stay up to date on the research progress by following this blog, the Systems for Action website, my research archive, and my twitter feed.

Equal Protection? What Zika and Big Data Reveal About Emergency Preparedness Across the U.S.

Equal protection is a principle worth striving for in emergency preparedness as in other spheres of human endeavor. My new blog post for the National Health Security Preparedness Index uses the big data included in that Index to explore inequities in health protection. Not every American community will experience Zika transmission. But every community needs the ability to prevent, detect, and manage emergencies in ways that minimize morbidity and mortality. Read more here.

Preparedness levels since 2013 have varied widely across domains and states.

New Research Opportunity Focuses on System Interactions, Intersections, and Spillovers

Science tells us with increasing precision that health and well-being in American communities is shaped by a complex web of determinants and contributing factors extending far beyond the reaches of the medical care sector. The physical environment, human and social capital, and economic opportunities and constraints exert a profound influence on health at molecular, person and population levels. One consequence of this fact is that actions taken outside of the health sector can have profound and even profoundly unintended consequences on health and well-being, and these effects are often distributed unevenly across society. Efforts to solve a transportation or housing or education problem, for example, may shift exposures to health risks and protective factors in ways that persistently alter the health trajectories of affected population groups. Economists and other social scientists refer to these phenomena variably as interactions, synergies, externalities, or spillovers across sectors, institutions and systems.

Unfortunately, our knowledge about these system spillovers is incomplete and still evolving, and so our ability to manage these spillovers in ways that optimize population health outcomes and minimize health inequities remains quite limited. The result is our current constellation of fragmented approaches for delivering and financing medical care, public health, and social and community services in the U.S., and our many missed opportunities for collective actions in improving health.

Earlier this month the Robert Wood Johnson Foundation (RWJF) announced a new funding opportunity to help build scientific knowledge about how to solve these systems problems as part of its $25 million investment in research to build a Culture of Health. The Systems for Action research program will develop a rigorous body of applied scientific work that reveals how best to align, coordinate, and integrate the many delivery and financing systems that shape health and well-being in American communities, including but not limited to the medical and public health sectors. This new work is not starting from scratch, but rather it builds from the strong foundation of health services research (HSR) and public health services and systems research (PHSSR) that has helped us understand and improve delivery and financing systems found within the health sector.

We invite you to review the research agenda developed for this exciting new program, along with the new funding opportunity announcement. Join us this Friday December 18 from 1-2pm for a webinar on the new research opportunity (registration required).

Midlife Public Health Crisis: Economists Discover What the Epidemiologists Missed

Amid the hustle of the APHA Annual Meeting this week some may have missed the striking new paper on the declining health status of middle-age white Americans by Princeton’s Anne Case and her colleague/spouse this year’s Nobel prize winning economist Angus Deaton. Making elegant use of some of our key national health datasets, the researchers show that a confluence of mental health problems, chronic pain, inability to work, and substance abuse track closely with an escalation in suicides, alcohol and drug deaths since the turn of the 21st century. Most shockingly, these trends have been sufficiently persistent and large to drive up all-cause mortality at midlife among white non-Hispanic Americans during the period 1999-2013. The dramatic rise in mortality and midlife distress among those age 45-54 was concentrated among those with fewer years of education.

The rise in mortality and decline in life expectancy among such a sizable demographic group over a relatively brief time frame is nearly unprecedented in the 21st Century American population. Case and Deaton note that only the HIV/AIDS epidemic has done this much damage to public health in modern times. The demographic timing of this health shock during prime working and earning years also means that the economic consequences are considerable and likely to contribute to a sluggish U.S. economy.

These new findings on the declining fate of white mid-life Americans, combined with our existing knowledge about the persistent health disparities faced by racial and ethnic minority groups in the U.S., should be a call to action for those of us who study public health delivery and financing systems. It is not clear who if anyone is being well served by our status quo approaches to protecting and improving population health. Surely we can find ways of retooling the U.S. public health system so that it becomes more responsive and more effective in detecting and addressing key population health dynamics.

Such retooling will require strengthening the public health system’s engagement with and influence on larger social, economic, and environmental systems that shape health and wellbeing, which are likely at the heart of America’s midlife crisis. Check out our new Systems for Action Research Agenda for ideas on how systems and services research can help us get there.

I invite you to comment on this blog, tweet at me, nudge me on linkedin, and follow my research archive.

A Decade After Katrina: Are We Better Prepared, and How Can We Know?

National Preparedness Month is upon us, and as we pass the 10 year anniversary of Hurricane Katrina, policy and public attention turns to the question of whether we are better prepared today than in years past. Does the nation have the necessary capabilities to blunt the health and economic consequences of disasters, pandemics and other large-scale public health emergencies? Looking just at the money, we know that federal outlays for public health preparedness and response programs have been in steady decline since Katrina (see figure). And our best data sources suggest that state and local government expenditures dedicated to public health preparedness are negligible. So what does this mean for preparedness levels across the US?

A new tool for measuring the nation’s preparedness and health security capabilities has been in development over the past several years: the National Health Security Preparedness Index. An initial version of the Index was released in December 2013, and the second, current version was released a year later in December 2014. The Index is definitely a work in progress as the science and practice of preparedness continues to evolve, and particularly as we learn how to measure important preparedness constructs and capabilities such as those outlined in the National Health Security Strategy.

We recently conducted a series of validation and simulation studies with the current version of the Index, and based on these results we have released a series of recommended improvements to the Index methodology and measures. My video overview of these recommended changes can be viewed here:

The Index is currently inviting public comments regarding these proposed updates to the Index methodology and measures. You can access the report on Index recommended updates and the public comment form here. The public comment period extends throughout National Preparedness Month and ends on September 30, so let us hear from you.

From Minneapolis to Milan to Manila: What We Liked and Learned at the Summer Meetings

The students are back at my door, punctuating the end of an active summer conference season for those of us who study public health delivery systems, policy and economics. This post highlights a few of the studies presented during the summer meetings just in case you missed some. I find these meetings invaluable not only for scientific networking and idea-generation, but also for the early access they give us to newly emerging findings and novel applications of research methodologies. The journal articles will come, but for now they are still months away or more. I have no time to be comprehensive here given the unfinished manuscripts on my own desktop, so this post dashes unabashedly through just a few of the meeting sessions in which I had a hand. But in doing so, I hope to pique reader interest in browsing other sessions and studies featured at these meetings.

My summer science always starts with the AcademyHealth Annual Research Meeting (ARM) in June, which met for the first time in Minneapolis this year. I had the good fortune to chair a session at this meeting featuring new research on the interactions between health care and public health delivery systems. You can view my Heard in the Hallways preview of this session here:

Particularly notable in this session, Case Western’s Scott Frank and his colleagues present a comparative analysis of community health assessment and improvement activities led by hospitals vs. public health agencies in Ohio, which won AcademyHealth’s award for Best Research in the Public Health and Population Health category. This study, conducted through our RWJF-supported Public Health PBRN Program, found substantive differences in the types of health conditions and intervention strategies targeted by the two types of organizations, suggesting that there are some benefits (complementarities) when each type of organization leads their own assessment and planning process, along with some missed opportunities for coordination.

A related study featured in this session by Emily Johnson and colleagues at the University of Colorado finds that state laws requiring hospitals to publicly report their expenditures on community benefit activities results in significantly larger expenditure levels for both charity care and nonclinical activities. One of our RWJF-funded Postdoctoral Scholars in PHSSR, Sharla Smith from the University of Kansas Medical Center, rounded out the session with her study of how public health inter-organizational networks respond to economic shocks and governance structures. Using data from our National Longitudinal Survey of Public Health Systems, she shows that governance exerts a stronger influence on inter-organizational connectedness than do resource constraints.

For the second year in a row we organized a methods workshop on social network analysis in health services research, which featured some innovative research on physician networks using claims data by Harvard’s Bruce Landon, along with studies of interdisciplinary care teams by the University of Minnesota’s Doug Wholey, and our own work using network analysis to evaluate multi-sectoral public health delivery systems (see figures). We also joined colleagues at RWJF in presenting a session on the three new research programs being implemented as part of RWJF’s Culture of Health action framework. These sessions represent only a small fraction of the new research on public health services and systems presented at the ARM, including an entire adjunct Interest Group Meeting on public health systems research.

A month later many of us took our science abroad for the World Congress of the International Association for Health Economics in Milan, where the conference theme of nutrition and economics revisited the famous Latin maxim “De Gustibus Disputandum Non Est” – should preferences be treated as fixed or endogenous in health economic research? The conference featured a dizzying array of new studies conducted around the globe on public health issues that include taxes on sugar-sweetened beverages, enhanced food labeling requirements, incentives for fruit and vegetable intake, the economics of disaster preparedness, novel family planning interventions, and substance abuse prevention policies. As part of our panel on the economics of public health delivery, Johns Hopkins economist David Bishai presented new research demonstrating how county-level public health expenditures reduce disparities in mortality rates between counties with high and low African American populations. We followed this with an updated analysis of how Medicald spending and public health spending interact at both state and local levels to influence population health, resulting in some unintended consequences involving crowd-out. Maryam Nejad at the Institute for the Study of Labor rounded out our session with an analysis of how suburbanization fuels obesity rates over time, using people who move from city centers to the suburbs to identify these effects. The interplay of studies conducted in the U.S., in other high-income countries, and in low and middle-income countries at this meeting allowed for particularly insightful scientific discussion and debate.

August brought us to the 7000+ island nation of the Philippines and its capital city of Manila for the 2015 edition of the Global Forum on Research and Innovation for Health. Sponsored by the Council on Health Research for Development, this meeting pulls together a large and diverse collection of several thousand researchers, representatives from health industries, and policy officials from national ministries of health, finance, science and technology. The Global Forum provides a platform for sharing research on strategies for improving health and economic development in low-resource settings, with a focus on low and middle income countries and their relationships with more advanced economies. As part of a session on public-private partnerships, we shared updated research on the health and economic value of multi-sector alliances for public health delivery in the U.S., using data from the National Longitudinal Survey of Public Health Systems. Along the way we shared lessons learned about how to support productive research collaborations using the mechanism of practice-based research networks (PBRNs), and also reflected some of our work related to disaster risk reduction as part of the National Health Security Preparedness Index. Many other fascinating sessions at this meeting featured research on disaster preparedness, non-communicable disease control, harnessing big data in health, and building flexible research infrastructure.

As did Milan, Manila left me surprised and intrigued by the range of research studies on public health delivery, financing, and economics that are now underway in African, Asian, Latin American, and Caribbean countries, among other settings. Clearly there is much to be learned by comparing the structures, processes, and outcomes that play out across these diverse systems and population groups.

Of course these meetings over the last three months are far from the only venues where good science on public health delivery, policy and economics were discussed this summer. Not included in this post, but not to be overlooked, is the venerable NACCHO Annual Meeting that convened in July with an active research track featuring studies from many colleagues in PHSSR.

Don’t make us wait for the journal articles. If you saw some research not-to-be-missed this summer, you can post a comment about it below, nudge me on LinkedIn, tweet at me, or just show up during my office hours. My own research presentations and products always can be found in the research archive.

Miller’s Rule: Improving Public Health Requires Understanding Institutions and Delivery Systems

This week we lost one of the great public health services researchers of the 20th Century in Dr. C. Arden Miller, who died on Sunday after more than a half-century of scientific inquiry to improve population health. His research spanned a wide continuum: vaccine development, reproductive health services, infant mortality, and—for much of the last three decades of his career—public health delivery systems. Miller believed that a key to improving America’s sub-optimal and inequitable health status lies in discovering ways of improving the fragmented and diffuse delivery systems for prevention and public health initiatives. And so early in the 1970s he began using the theories and methods of health services research to study public health institutions and delivery systems, while most of the HSR field focused narrowly on medical care delivery and financing.

Miller’s work provided the foundation for much of the contemporary research on public health delivery and financing, including our national surveys of public health delivery systems and our methods for measuring public health system performance. Indeed, his enthusiasm for this area of inquiry is what attracted me to the PHSSR field as a graduate student from more “mainstream” research interests in health care financing and care management. Methodologically, Miller taught us that one can learn only so much about delivery systems from afar through surveys and administrative data. At some point, one must go down to the factory floor, observe the operations and talk to the people who do the work of producing population health. His comparative case studies of U.S. local health departments (books pictured) and cross-national comparisons of maternal and child health systems (featured in a 1994 New York Times article) are powerful examples of the merits of a mixed-methods approach.

Miller’s long legacy of action-oriented public health research lives on in the work of the many, many researchers and health professionals whom he taught and mentored. Farewell and God Speed Dr. Miller.