Poor performance in achieving population health goals is all too familiar. So is the accompanying every-decade ritual in the United States: the announcement of a new round of planning to create health goals for the nation (eg, Healthy People 2020), followed by a wave of enthusiasm and then disenchantment (eg, "the problems with the data arise from . . ."), search for the guilty (eg, "but they were never at the table"), punishment of the innocent (eg, "with this reorganization, our agency looks forward to . . ."), and reward for the uninvolved (eg, "we should never forget that America offers the world's highest-quality health care").

Lost in this drama are the numbers: for the 281 measurable public health performance objectives tracked for Healthy People 2010, only 10% met their targets (1). Although progress was made toward meeting nearly 50% (n = 138) of the objectives, 20% (n = 57) grew worse. Disparities in health outcomes for ethnic minorities also remain a failure. One of the most glaring disparities is in the African American community, in which 48% of adults suffer from a chronic disease compared with 39% of the general population. Why do we keep falling short of the bars we have set for ourselves in population health and health equity?

Several factors contribute to these poor results. First, multiple and unconnected sectors lack shared responsibility for outcomes. Consumers, providers, insurance companies, employers, and government agencies all vie for individual advantage in our fragmented health care system, avoiding responsibility for unimpressive outcomes. Second, the health care system lacks cooperation and collaboration in achieving population-level goals. Emmanuel and Fuchs (2) characterize this as "the myth of shared responsibility." Third, no public or private entity has overall responsibility for improving population health. This situation contributes to a willingness to proclaim victory for hard work, rather than meaningful improvement (3). Finally, moving toward improved population health and health equity requires understanding what works and what does not, and a willingness to agree on the price we pay for each. Sustained cooperation and shared responsibility among stakeholders in different sectors of a comprehensive public health system are necessary (4).

The public health response promotes community partnerships and cooperation as represented in the essential services. Public health agencies have come to recognize that community partnerships are a necessity in health improvement and that major health initiatives require community coalitions (5). Results are mixed, but the empirical evidence base for the effectiveness of partnerships to improve population health is growing (6-9).

In response to these problems, we offer a framework to guide collaborative action to improve population health. We also outline key processes for promoting community/system change and population health improvement. We conclude with 7 recommendations for strengthening collaborative partnerships for population health and health equity.

Collaboration is difficult to establish and maintain. First, stakeholders often have differing goals or understanding of the problem, which leads to disagreements and a devaluing of others' preferred strategies and approaches. Partners who share responsibility for naming and framing the problem may find it easier to bridge those differences. Having common goals makes it easier for stakeholders to see their potential contribution to healthier communities.

Second, stakeholders often focus narrowly on only a few of the many factors that contribute to the problem. They typically use interventions to address these through familiar channels of influence; yet improving population health requires comprehensive and coordinated approaches that address 1) multiple personal and environmental factors (eg, knowledge and skills, access to services and support, policies and living conditions), 2) multiple sectors (eg, health, education, government), 3) multiple ecologic levels (eg, individuals, organizations, communities, broader systems). Stakeholders are more likely to see the work they do as particularly needed; thus, shared responsibility among organizations working in multiple sectors is rare.

Third, working at multiple ecologic levels is challenging. Different determinants of health have different areas of policy action and related actors (eg, Medicare, federal officials; air quality, regional actors; school nutrition, local people). Few partnerships coordinate collaborative action across multiple levels. Fourth, working together requires flexibility on the part of stakeholders' organizations and those who fund them. Yet many nonprofit organizations and governmental agencies have policies that limit their capacity to share resources and responsibilities.

Fifth, measurement of accomplishments is also a challenge. Many initiatives do not have accurate or sensitive measures of success at the level of the whole community. Changes in the community or system — the unfolding of new programs and policies — need to be measured to see what was actually implemented and its contribution to more distant population-level outcomes. The merit of longer-term efforts is difficult to assess and adjust to without such measures of environmental change.

Sixth, incentives for population-level improvement, such as outcome dividends, are rare. Without effective incentives for improving population health, the time and effort of collaborating with partners may go unrewarded. Working together across organizations is challenging because of competition for limited funding. The prevailing contingencies of reinforcement help secure discrete resources for individual organizations, not groups of organizations to improve population-level outcomes for which responsibility is shared.

Seventh, our knowledge of how to effect change in communities and systems to produce substantial improvements in population health is limited. We need a better understanding of how key collaborative processes, such as action planning or community mobilization, can yield environmental changes that will improve population health. Stakeholders may lack the experience or training required to make the community or system changes needed to affect public health.

Finally, public health has promoted best practices or programs that work as a way to ensure that the most effective approaches are implemented. The problem is that evidence-based programs are typically tested with small numbers of individuals and evidence of comprehensive and context-appropriate strategies that actually improve population health is rare. Researchers and practitioners have begun to reorient their efforts to population health using frameworks and related processes (9-12).

We have adapted the Institute of Medicine framework for collaborative public health action in communities (Figure 1) (4,8). This framework, like other related frameworks (11), is iterative and interactive, with interdependencies between the phases and related processes. For instance, the first phase (assessment and collaborative planning) is oriented to indicators of success, such as reduced rates of childhood obesity or diabetes, that define the endpoints noted in the last phase (achieving improvement in population health and health equity). Emerging evidence suggests that 12 collaborative processes, such as action planning and making outcomes matter, may facilitate change and improve related outcomes in population health (Figure 1) (9,13,14).

We conclude with 7 key recommendations for strengthening collaborative partnerships to assure health for all:

These recommendations aim to ensure conditions — including monitoring and feedback systems, training and technical support, and group incentives for progress — that can foster the success of broad collaborative partnerships (25). Such conditions should make it easier and more likely for multisectoral partnerships to achieve progress in improving population health and health equity.