Describing Total Population Health: A Review and Critique of Existing Units

Abstract

Total population health is a key tenet of health care reform efforts, evident in initiatives such as the National Quality Strategy, shifts toward population-based payments, and community benefit requirements for tax-exempt hospitals. Representing total population health in a way that guides best practices and establishes shared accountability for geographic communities, however, remains a challenge in part because of differences in how stakeholders define populations. To better understand the landscape of potential denominators for population health, this study examined a selection of relevant geographic units. The approach included a comprehensive review of health services and public health research literature as well as recent pertinent health policy documents. Units were characterized based on whether they: exhibit “breadth” of coverage across the whole US population; are “accurate” or grounded in health care utilization patterns; are “actionable” with mechanisms for implementing funding and regulation; and promote “synergism” or effective coordination of public health and health care activities. Although other key components of a total population health unit may exist and no single identified unit possesses all of the aforementioned features, several promising candidates were identified. Specifically, healthcare coalitions link health care and public health domains to care for a geographic community, but their connection to utilization is not empiric and limited funding exists at the coalition level. Although Accountable Care Organizations do not uniformly incorporate public health or facilitate coordination across all payers or providers, they represent an effective mechanism to increase collaboration within health care systems and represent a potential building block to influence total population health.

Introduction

Substantial attention is being devoted to ensuring that the $3 trillion health care industry in the United States¹ delivers high-quality care at the lowest possible cost. Population health is a national priority throughout these efforts and is one of the 3 central aims of the National Quality Strategy, which is mandated under the Affordable Care Act (ACA). By 2018, 1 of every 2 Medicare dollars will be spent in novel payment models, including population-based reimbursement.² In addition, the ACA provision requiring nonprofit hospitals to assess and address needs of their communities provides a $7.4 billion opportunity for population health. Although the delivery of uncompensated care is a traditional mechanism for nonprofit hospitals to justify a benefit to their community, increasing insurance coverage and evolving payment models have created an opportunity for these hospitals to identify novel ways to improve the health of their community beyond the hospital walls.^3,4

Despite this widespread policy focus on population health, a common definition of population remains elusive. To insurers, for example, it is a group of beneficiaries, while to health care providers it is a patient panel. Hospitals focus on inpatients while both policy makers and public health officials consider all of the people living in their jurisdiction. As the landscape of our health care system shifts, health care measures are increasingly being tied to financial incentives and penalties in order to emphasize accountability and promote collaboration among a diverse set of stakeholders. At the same time, there has been increased recognition of the importance of social and environmental factors in shaping health outcomes. The vision of health care and public health sharing responsibility for the well-being of their overlapping and intertwined communities is at the core of the concept of total population health,⁵ but bridging the competing definitions of the population makes this difficult.

As a result, health improvement efforts have largely been focused on relatively narrowly defined subsets of the population rather than on a more comprehensive “total population health” perspective. The Institute of Medicine has recently defined total population health (TPH) as the “health status and health outcomes of persons within a geopolitical area.”⁶ In contrast to a focus on populations covered by a single insurer or those served by a single hospital or health care system, TPH includes the entire population and is consistent with the reality that state and local funding decisions, regulations, and surveys occur within a geopolitical space.⁷

Although relevant subpopulations can and do exist within a broader population, a focus limited to populations that are defined by payers or providers fails to address the inherent locality in how health and health care are experienced from the patient perspective. Figure 1 demonstrates the distinction between the subpopulations typically considered by insurers and health systems compared to how an individual living as a part of a community might identify the broader population. Geographic approaches are common in public health, public safety, and emergency management, and although geography has long been used to describe differences and disparities in health care outcomes throughout the United States,⁸ no clear and consistent construct for characterizing populations has emerged. The study team acknowledges that a geographic framework may be less useful in referral conditions, but most primary care and almost all emergent and acute care takes place locally. The team sees the applicability of a geographic frame to be strongest in this context; aligned incentives across stakeholders would allow for shared accountability and incentive to improve population-based health care (eg, time to see a primary care provider across a community, cardiac arrest survival rates within a metropolitan area).

FIG. 1.

Different perspectives on population health. A color version of this figure is available in the online article at www.liebertpub.pop

In order to align incentives across the public health sector and competing hospitals and health systems, communities need to better understand the stakeholders committed to improving their health. Conversely, public health and health care stakeholders require a clear sense of the community for which they are responsible for improving health outcomes. A common definition of populations could be used to bridge health care and public health approaches to emerging and evolving policy mechanisms, such as the hospital community benefit and community health needs assessment,⁹ toward the ultimate end of moving total population health from catchphrase to reality.

The study team sought to describe many of the existing approaches to population health through a geographic lens and to characterize their ability to comprehensively describe a community. To do this, the team performed a comprehensive review of health services and public health research literature as well as pertinent health policy documents and analyzed potential “denominator” units that might be used to provide a framework for community and population health.

Identifying and Examining Candidate Units

Candidate units (Table 1) were selected based on a review of those used in health services or public health research and identification of others created by recent health policy initiatives. Because of the heterogeneity of the health geography landscape and the lack of standardization in approaches and terms, this review is not intended to comprise all units that have ever been used for population health analysis. Rather, the goal was to find measures that are either commonly referenced for these purposes or that were recently highlighted in national health policies. Following identification of the units, queries of PubMed and Google Scholar databases were conducted, using Google searches of “.gov” domains to identify additional reports. The search methodology was targeted to collect articles that centered on units within the context of population health, public health, community health, or health care geographies. The study team used a snowball sampling method to gather additional sources.

Table 1.

Candidate Units for Describing Total Population Health ^*

Accountable Care Organization: Established under the Affordable Care Act, a group of providers that assumes responsibility for a patient population based on its health care utilization

Census Units: Constitutionally mandated decennial count and survey of the US population among other measures, conducted by the US Census Bureau for a variety of administrative and statistical purposes

• County: Primary legal division of most states often associated with governing authority or administrative purposes, including local health department organization

• State: Major political and administrative unit of the United States, which is also the primary geographic unit for the purposes of Census data

• Public Use Microdata Area: Geographic areas with a minimum population of 100,000 used for the purpose of disseminating small-area data while preserving confidentiality

• Zip Code Tabulation Area: Generalized geographic representations of the US Postal Service zip code service areas

Core-Based Statistical Area: Geographic entity delineated by the Office of Management and Budget to identify core urban counties or equivalents with a population of at least 10,000 persons and associated adjacent counties with a high degree of economic integration, as evidenced by commuting patterns

Dartmouth Atlas: Based at the Dartmouth Institute for Health Policy and Clinical Practice and supported by a group of funders led by the Robert Wood Johnson Foundation, it uses Medicare and other data to document and analyze health care markets

• Hospital Referral Region: To represent regional health care markets, hospital service areas are aggregated based on utilization patterns for major cardiovascular surgery and neurosurgery

• Hospital Service Area: To represent local health care markets for hospital care, zip codes of Medicare patient residence are assigned to the zip code in which the majority of their hospitalizations occur

• Pediatric Surgical Area: To represent regional markets for pediatric surgery in New England, hospital service areas were aggregated based on utilization patterns of common ear-nose-throat procedures and appendectomies

• Primary Care Service Area: To represent Medicare patient travel for primary care, areas were aggregated to reflect the market to which patients travel to receive the majority of primary care services

Emergency Medical Service Region: Organized at the state level, emergency medical services systems administer and coordinate emergency care within assigned boundaries to meet the needs of local communities

Healthcare Coalition: A group of health care organizations, public health, and safety partners that collaborate in order to improve the health and resilience of their community, especially in the context of public health crises

Rational Service Area: Discrete geographic unit that represents potential primary care utilization based on travel time to services

• Health Professional Shortage Area: A rational area or neighborhood where travel time to primary care services exceeds 30 minutes, thus making it eligible to receive a number of federal benefits

• Medically Underserved Area: A neighborhood designated as underserved based on a calculation involving primary care physician numbers, infant mortality, poverty rates, and population age making it eligible for a number of federal benefits

Trauma Service Area: Categorization used in Texas and Florida for the purpose of regionalizing trauma care to coordinate care for patients statewide and to guide trauma center development, respectively

Presented in alphabetical order.

Given the absence of broad agreement on what constitutes an ideal total population health unit, the study team identified major themes in the candidate units to establish an evaluation framework. For example, many of the units, including the census units, accounted for the whole of the population geographically (breadth). The Dartmouth Atlas units, as well as rational service areas, rely on utilization of health care services to determine “boundaries” (accuracy). Healthcare coalitions are tied to funding and encourage coordination between public health and health care (actionability and synergism). Experts in population health have recommended that units be tied to geopolitical boundaries.⁶ With these themes in mind, the key components were chosen because of their ability to describe a balanced concept of total population health across multiple stakeholders and at multiple levels.

In order to evaluate a heterogeneous group of geographic units, the study team uses 4 key components as common vocabulary: breadth, accuracy, actionability, and synergism (Table 2). In this context, breadth is the degree to which a unit, in sum, provides total geographic coverage of the United States, which is an essential component for supporting TPH. Accuracy is the extent to which the unit reflects health care utilization patterns. This is critical for demarcating populations in a way that incorporates the appropriate stakeholders and thus helps to ascertain whether or not the unit can promote accountability. Actionability refers to the degree to which funding or regulatory actions occur at the level of the candidate unit that could be used to influence outcomes. Finally, synergism represents whether the unit encompasses both public health and health care, which is meaningful because of the increasing impetus for stakeholders to effectively coordinate activities across the continuum of care and at multiple levels of intervention.

Table 2.

Key Components for Describing Total Population Health

Component	Description
Breadth	The degree to which the units achieve total coverage of the US population
Accuracy	The degree to which the unit reflects health care utilization patterns
Actionability	The degree to which the unit might be used when leveraging improvements in total population health, considering the presence of state or federal regulatory or funding actions related to health or health care
Synergism	The degree to which the unit encompasses both public health and health care activities

Candidate Units to Measure Total Population Health

Census and statistical units

Over the decades since its establishment in the early 1900s, the US Census Bureau's geographic framework has adapted to the needs of a variety of federal and state agencies and has increasingly incorporated new technology in order to improve specificity. Current geographic areas are tabulated in 2 major categories: administrative/legal and statistical.¹⁰

Administrative/legal geographic units generally originate from federal or local laws, statutes, or court decisions, and are associated with boundaries for such purposes as elections and local programs.¹⁰ States and counties are perhaps the most familiar of these units. To some degree, residency in a certain state or county confers a sense of membership and community. Voting, public education, and many health care funding and planning functions occur within these administrative boundaries. States administer the Medicaid program, license health care professionals, and receive federal funding for the health insurance marketplace and a number of public health programs. Counties, which are the primary legal division of most states, tend to serve as the level of organization for local health departments.¹¹ The comparison of certain health outcomes between geographic locations at both the state and county level has served as a driving force for health care quality improvement and public health initiatives.¹²

Statistical units, on the other hand, are delineated by the Census Bureau in order to establish standard entities for data analysis or programmatic needs, and they are frequently utilized in health services research for the purpose of describing geographic disparities.¹⁰ Multistate regions, for example, are frequently used to highlight poor healthcare outcomes in the southeastern United States. When patient-level sociodemographic data is unavailable, small census tracts, block groups, or zip code tabulation areas can be used to overlay income or education estimates based on a patient's area of residence. These units are utilized as geocodes for a number of national health surveys, which are typically available at the county level, although some can be geocoded at the census block.¹³ Public use microdata areas are unique statistical units delineated by states to represent “communities with relatively homogenous characteristics” at a population size of 100,000–200,000. These units are utilized to disseminate data from the American Community Survey.¹⁴

Among the various statistical census units, core-based statistical areas (CBSAs) deserve special acknowledgement because of their use by the Centers for Medicare & Medicaid Services (CMS) to determine regional reimbursement rates. Since the 1940s, the federal government has recognized the importance of defining metropolitan areas. Although there have been a variety of permutations of the specific parameters, these units have generally attempted to delineate core metropolitan areas and associated adjacent counties with a high degree of social and economic integration.¹⁵ In addition to being used frequently in health services research because of the ease of overlay between data sets,¹⁶ CBSAs are used to assign hospitals to a certain labor market for the purposes of adjusting Medicare reimbursements.¹⁷

With the exception of CBSAs, which fail to account for nonmetropolitan populations, census and related statistical units have the benefit of breadth and total coverage of the US population. Administrative census units are tied to health care funding and regulation and can span the domains of health and health care, creating a demonstrated mechanism for actionability. Medicaid funding allocation and practitioner accreditation as well as public health department administration, for example, occur at the state level. However, all of these units fail to accurately reflect health care utilization patterns, thus limiting their applicability as an accountable unit for healthcare quality improvement (Table 3). Importantly, census and administrative units have a variety of functions beyond the health sector. Accordingly, the most significant limitation of these units for capturing total population health is that they cannot be changed specifically for this purpose.

Table 3.

Characterization of Candidate Units

	Key Components
Unit	Breadth	Accuracy	Actionability	Synergism
Accountable Care Organization	-	●	●	-
Census Units
County	●	-	●	○
State	●	-	●	●
Public Use Microdata Area	●	-	-	-
Zip Code Tabulation Area	●	-	-	-
Core-Based Statistical Area	-	-	○	-
Dartmouth Atlas
Hospital Referral Region	●	○	-	-
Hospital Service Area	●	○	-	-
Pediatric Surgical Area	○	○	-	-
Primary Care Service Area	●	○	-	-
Emergency Medical Service Region	●	○	●	○
Healthcare Coalition	●	○	●	●
Rational Service Area	○	○	○	-
Health Professional Shortage Area	-	○	●	-
Medically Underserved Area	-	○	●	-
Trauma Service Area	○	○	●	-

● The unit meets the criteria for the indicated component to a high degree; ○ The unit meets the criteria for the indicated component to a limited degree; - The unit does not meet the criteria for the indicated component

Dartmouth Atlas

In recognition of the limitations of census and related units, the Dartmouth Atlas was developed 20 years ago to better delineate naturally occurring health care markets. Geographic units identified by the Dartmouth Atlas project include hospital service areas (HSAs), hospital referral regions (HRRs), primary care services areas (PCSAs), and pediatric surgical areas (PSAs). These units were developed with the goal of reflecting actual patient utilization patterns of health services, which were found to correspond poorly with the units of analysis, such as counties, that had been used traditionally.⁸

Developed in the 1990s, HSAs and HRRs reflect hospital utilization patterns of Medicare beneficiaries. HSAs were tabulated by assigning zip codes of patient residence to the zip code in which those patients received the majority of their hospital care, and HRRs were defined by further grouping HSAs to reflect utilization of services for major cardiovascular surgery and neurosurgery. This method resulted in the definition of 3436 HSAs and 306 HRRs. The original HSA and HRR boundaries have not been redrawn, thus maintaining consistency in the database.⁸

More recently, with funding from the Health Resources and Services Administration (HRSA), the Atlas came up with PCSAs using Medicare data in an attempt to augment the visibility of primary care availability. For these purposes, primary care providers include both primary care clinicians (physicians and nurse practitioners) and facilities, such as Rural Health Clinics and Federally Qualified Health Centers.¹⁸ The Dartmouth Atlas approach was considered to address limitations of existing units that lacked standardization, such as the rational service areas used to designate underserved populations for federal funding.¹⁹

PSAs, limited to the northern New England region of the United States, reflect utilization patterns for pediatric surgical procedures (ie, ear-nose-throat procedures, appendectomies). Unlike the other Dartmouth Atlas units, PSAs are derived from a combination of Medicaid and commercial claims data and are thought to capture 66% to 90% of the total pediatric population in those regions.²⁰

The Dartmouth Atlas units have been widely adopted in health services research for the purpose of examining geographic variations in healthcare. Both the Congressional Budget Office and the Institute of Medicine have used these units to analyze national geographic variation in health care spending.^21,22 However, they are not currently tied to any health care measurement, funding, or regulation.

With the exception of PSAs, which are confined to the northeast region, the Dartmouth Atlas units are promising both in their breadth of coverage and their accuracy in reflecting health care utilization. However, an important caveat to these units' accuracy is that most of them rely on Medicare claims data and consequently may not represent utilization patterns of uninsured, pediatric, or privately insured populations. Also, although the units' boundaries were drawn based on service usage patterns at the time of their development, the Atlas does not update boundaries. This allows for continuity of the database, but it limits the applicability of the units for present-day measurement. The Dartmouth units also are limited by their actionability. Because of hospital closures and other health care changes, a number of HSAs do not contain a hospital anymore, making them impractical for leveraging healthcare change. The Institute of Medicine has specifically recommended against utilizing a HRR-based index for value-based purchasing because HRRs do not represent a level at which health care decisions are made.²² All of the Dartmouth units are also limited in synergism, accounting only for hospitals or primary care clinics (Table 3).

From Dartmouth to accountable care organizations

Although Dartmouth Atlas units have not been directly applied for the purpose of health care funding, planning, or regulation, they heavily influenced a number of health policy initiatives including the structure of accountable care organizations (ACOs) under the ACA. Although not explicitly geographic in nature, ACOs functionally exist within a region and provide financial incentives to improve population health. Specifically, ACOs encourage collaboration between providers and hospitals in order to improve care for a defined patient population. Under CMS's Medicare Shared Savings Program, beneficiaries are assigned to an ACO if they receive a plurality of their primary care services within that ACO. The ACO is then eligible to participate in shared financial savings if it meets a minimum quality standard, which is benchmarked based on national data, and minimum cost savings requirements based on estimates from historical cost data.²³

Over the past several years, applicants have demonstrated strong interest in both government-supported and commercial ACOs, and these organizations have demonstrated promising early results in terms of improving health care quality. As of 2014, there were more than 600 ACOs throughout the country, covering 20.5 million Americans. Coverage across the nation tends to vary by state, with more than 15% of the population covered by ACOs in Maine and Vermont, while 0–2% of the Midwest population is covered by ACOs.²⁴ Although the full impact of the ACO program is still under debate, some evaluations have suggested that ACOs might be effective at improving health care quality and reducing costs.²⁵

Despite this early popularity and potential success, ACOs face significant limitations within the context of their breadth, accuracy, actionability, and synergism with respect to improving total population health. The current coverage of ACOs is limited, with fewer than 1 in 10 Americans being covered by an ACO. Also, although patients are assigned to ACOs based on utilization patterns, these organizations are not accountable for any defined geographic area and fail to adequately provide care for an entire community. Experts, including some of the pioneers of the ACO movement, have advocated for ACOs to assume responsibility for geographic populations (accountable care communities) rather than solely populations assigned based on utilization. As they are currently defined, however, ACOs lack both the financial incentives and the capabilities to broaden their scope or to coordinate with public health domains.²⁶

Underserved populations and the rational service area

Health Professional Service Areas (HPSA) and Medically Underserved Areas (MUA) originated in the 1980s to designate areas with demonstrated need for additional health care resources. Both of these units are delineated at the county level in nonmetropolitan areas and at the “community” level in metropolitan areas, where communities are defined as groups that display a strong “self-identity” related to homogenous socioeconomic or demographic characteristics. Health care need is demonstrated by the ratio of certain health professionals to population for HPSAs, and by specific criteria for MUAs, including primary care availability, infant mortality rates, and proportions of the population who are elderly or poor. Based on these designations, HPSAs and MUAs are eligible for certain federally funded programs, enhanced CMS reimbursement rates, and the use of National Health Service Corps members.²⁷

Although HPSAs and MUAs by definition do not cover the entire US population, several states have divided their total population into rational service areas (RSAs) for primary care in order to better assess health needs and promote health care planning state-wide. RSAs are mandated under health planning programs in a handful of states, and several other states have voluntarily developed these service areas. Although RSAs are generally defined based on travel time to primary care, exact specifications of these areas vary by state, and the concept has not been widely adopted throughout the nation. State-wide RSAs also are not standardized. Although many state-wide RSAs are only used in the delineation of MUA or HPSA shortage areas for the allocation of grants and resources, some states have utilized these units for statistical analyses and health care planning.²⁸

In sum, RSAs are significantly limited by their breadth, accuracy, actionability, and synergism. In response to some of these concerns, PCSAs were developed as part of the Dartmouth Atlas.¹⁹ However, these units are limited by additional concerns regarding accuracy and actionability, as already discussed (Table 3).

Coordination of health care during emergencies

The provision of health care during both individual emergency medical conditions and public health crises has long required consideration of the geographic context of health and health care. Emergency medical service (EMS) systems have been regionalized since the 1970s, and trauma service areas (TSAs) evolved in the early 1990s to better coordinate the planning of trauma care. In particular, planning for patients with trauma, burns, ST segment elevation myocardial infarction, and stroke is already done using a regional perspective, but discrete geographic units are infrequently described.^29,30 Under the Hospital Preparedness Program, healthcare coalitions represent another novel and encouraging method for regionalizing care and potentially capturing community and population health.

The modern EMS system began after a government report in the 1960s called attention to the lack of coordinated care for accidental injury. EMS systems were organized at the national level in 1973, but by the early 1980s, federal funding and oversight of these systems waned. Subsequently, these regions have been funded and organized at the state or local level at various degrees, leading to what is now a highly fragmented, decentralized, and variable EMS system.³¹ Accordingly, there is no consistent method for defining EMS region boundaries, and although areal coverage tends to be high, there are parts of the US population without adequate EMS access.³² Although differences exist, EMS agencies largely exhibit effective actionability, and some agencies have successfully developed synergism by engaging in community outreach and education. Nonetheless, these initiatives are neither widely adopted nor approached in a consistent way throughout the nation.^33,34

Similar to EMS regions, TSAs have been defined in Texas and Florida in order to assist in the planning and delivery of trauma care. In an effort to provide trauma care access to all citizens, Texas state legislation divided the state into 22 service areas with regional advisory councils, created a trauma facility designation protocol, and called for the development of a statewide trauma system registry.³⁵ Florida has delineated 19 TSAs for the original purpose of guiding trauma center development, with strict criteria for the number of allowed trauma centers per area.³⁶ Although certain TSAs in Texas have declared success in regionalized trauma care,³⁷ the system in Florida continues to explore ways to improve trauma care, including the establishment of additional regional trauma agencies for the purposes of better aligning with existing Regional Domestic Security Task Force boundaries.³⁶

Beyond the delivery of day-to-day emergency care, healthcare coalitions were created under the 2002 Hospital Preparedness Program with the goal of building on regional health care preparedness for and in response to major public health events. Coalitions are grant-funded entities defined as “a group of healthcare organizations in a specified geographic area that agree to work together to enhance their response to emergencies or disasters.”³⁸ These coalitions must determine their geographic boundaries and are encouraged to consider factors such as local government jurisdictions, topography, and patient referral patterns.³⁸ Although most healthcare coalitions began as hospital-centric organizations, they now incorporate multiple hospitals as well as EMS agencies, health departments, fire and police jurisdictions, and allied health services including providers of home health goods (oxygen tanks) and services (visiting nurses). Members have indicated that regional partnerships have been one of the most valuable aspects of coalition formation, and the support of day-to-day activities has been a strong motivator for participation in these coalitions.³⁹ Early analysis of coalitions indicates that they have a high penetration rate and can provide enough value to members to function independently of federal grant funding.⁴⁰ To date, coalitions have been focused on large-scale disaster preparedness response, and the ability of these units to capture the health and health care within a region is unknown. However, the collaborative multi-stakeholder structure of these entities offers promise should their scope and makeup continue to expand.

Overall, although the adoption of TSAs has been limited throughout the United States, health care coalitions and EMS regions have broad geographic coverage and have demonstrated successes in linking the domains of health care and public health. All of these emergency units demonstrate mechanisms for state or federal funding and regulation. However, these units are limited by substantial variation across the country, as well as their emphasis on emergencies and disasters, which detracts from their potential application to total population health measurement (Table 3).

Establishing Accountability for TPH

Accurate and appropriate measurement of the health of communities may promote transparency, align incentives, and ultimately establish accountability for total population health. Toward these ends, a TPH denominator must (1) in aggregate, account for the whole of the US population; (2) be grounded in health care utilization patterns; (3) have funding and regulatory actions that can be implemented; and (4) allow for effective coordination of public health and health care activities.

The present analysis of candidate units has revealed both significant gaps in current capabilities and opportunities for improvement moving forward. Many of the units have a high degree of coverage of the total US population and some are associated with health care funding or regulatory mechanisms. Most of them lack adequate accuracy, however, and very few are reflective of a wide scope of health and health care activities. State and county health rankings demonstrate the power of measurement as an impetus for policy change. However, without a direct tie to care delivery, these data are rarely actionable. Although promising population health management initiatives are under way, most take place within a contained hospital, health system, or payer. As a result, these programs lack the breadth required to be scalable. An ideal “denominator” for TPH would balance comprehensiveness of the community with the ability to be leveraged to catalyze change.

The study team recognizes that many of the units examined are associated with health care delivery rather than public health. This is intentional, as public health spending represents less than 3% of our country's nearly $3 trillion health expenditures while health care represents 85% of the expenditures.¹ The tremendous resources of the health care industry have yet to be broadly applied toward the end of improving total population health. The team submits that more fully incentivizing engagement of health care systems, in collaboration with public health departments and other key stakeholders, will be essential in order to achieve the goals of TPH.

Although significant limitations were evident throughout the reviewed candidate units, a few units demonstrate promising characteristics. Healthcare coalitions demonstrate early successes in linking health care and public health domains in order to care for a geographic community, but their connection to utilization is not empiric and limited funding exists at the coalition level. Although payment models exist for ACOs, their current structure focuses on a limited population health management strategy rather than a TPH perspective. Despite this, they represent an effective mechanism for collaboration across the health system to deliver high-value care. Despite wide variation, EMS regions demonstrate an actionable unit for delivery of regional health care and have had demonstrated success in public health outreach efforts.

Conclusions

Although the concept of population health has become a key component of the redesign of the US health care system, little clarity exists about how to define the population. The Community Health Needs Assessment, for example, allows hospitals to define their own denominator, and ACOs encompass limited subpopulations. In the midst of current health care reform efforts, systems are currently evolving that might be used to more appropriately measure and incentivize improved TPH. Further research is needed to gain a better understanding of how populations might best be defined in a way that is reflective of geopolitical communities as patients experience them, is inclusive of the total US population and any vulnerable subpopulations, is actionable among the different stakeholders in public health and health care, and has the flexibility to be leveraged at multiple levels. Operationalizing the unit, including local optimization and assessment of how the unit can be integrated in a complex, multilevel health system, will also require further conceptual and empirical consideration. The study team has described some key domains of population health that can be used to assess geographic constructs; future work may build on these and prioritize the importance of different domains.

Improving TPH requires alignment of incentives across the many stakeholders invested in the health of the population. To date, no geographic unit exists that achieves the goal of linking public health and health care in a manner that is comprehensive, accurate, and actionable. The ACA's focus on improving population health, renewed attention on understanding community benefit, and a national emphasis on health care delivery system redesign make this a key time to construct or identify a consensus definition of community and population that will allow for aligned incentives and benchmarking of health.

Footnotes

Author Disclosure Statement

Drs. Margolis, Wiebe, and Carr, and Ms. Baehr, Ms. Holland, and Ms. Biala declared no conflicts of interest with respect to the research, authorship, and/or publication of this article. This research was supported in part by an appointment to the Research Participation Program at the US Department of Health and Human Services administered by the Oak Ridge Institute for Science and Education. The findings and conclusions in this report are those of the author(s) and do not necessarily represent the views of the Department of Health and Human Services or its components.

Acknowledgments

We would like to acknowledge Aditi Gupta, Syeda Majid, MS, and Austin Kilaru, MD for their research assistance. We would like to thank Alicia Livinski, MA, MPH for her guidance and assistance.

References

Centers for Medicare & Medicaid Services.. National Health Expenditure Fact Sheet. 2015. https://www.cms.gov/Research-Statistics-Data-and-Systems/Statistics-Trends-and-Reports/NationalHealthExpendData/NHE-Fact-Sheet.html. Accessed October 1, 2015 .

Centers for Medicare and Medicaid.. Better care. Smarter spending. Healthier people: paying providers for value, not volume. 2015. http://www.cms.gov/Newsroom/MediaReleaseDatabase/Fact-sheets/2015-Fact-sheets-items/2015-01-26-3.html. Accessed May 29, 2015 .

Young

, Chou

, Alexander

, Lee

, Raver

. Provision of community benefits by tax-exempt U.S. hospitals. N Engl J Med. 2013; 368:1519–1527.

US Department of Health and Human Services. The Economic Impacts of Medicaid Expansion, Uncompensated Care Costs and the Affordable Care Act. Washington, DC: US Department of Health and Human Services; 2015.

Kindig

, Stoddart

. What is population health?. Am J Public Health. 2003; 93:380–383.

Institute of Medicine Committee on Quality Measures for the Healthy People Leading Health Indicators. Toward Quality Measures for Population Health and the Leading Health Indicators. Washington, D.C.: The National Academies Press; 2013.

Jacobson

, Teutsch

. An Environmental Scan of Integrated Approaches for Defining and Measuring Total Population Health by the Clinical Care System, the Government Public Health System, and Stakeholder Organizations. Washington, DC: National Quality Forum; 2012.

The Dartmouth Atlas of Health Care. http://www.dartmouthatlas.org/. Accessed June 18, 2015 .

Folkemer

, Somerville

, Mueller

, et al. Hospital Community Benefits after the ACA: Building on State Experience. Baltimore, MD: Baltimore County Hilltop Institute; 2011.

10.

US Census Bureau. Geographic Areas Reference Manual. Washington, DC: US Department of Commerce; 1994.

11.

National Association of County & City Health Officials. National Profile of Local Health Departments. Washington, DC: National Association of County & City Health Officials; 2014.

12.

County Health Rankings. County health rankings & roadmaps. http://www.countyhealthrankings.org/. Accessed June 18, 2015 .

13.

Centers for Disease Control and Prevention.. Geocodes. 2013. http://www.cdc.gov/rdc/b1datatype/dt123geocod.htm. Accessed June 18, 2015 .

14.

King

, Furukawa

, Buntin

. Geographic variation in ambulatory electronic health record adoption: implications for underserved communities. Health Serv Res. 2013; 48:2037–2059.

15.

US Office of Management and Budget. Standards for defining metropolitan and micropolitan statistical areas. Fed Regist. 2000; 65:82228–82238.

16.

Gaskin

, Dinwiddie

, Chan

, McCleary

. Residential segregation and the availability of primary care physicians. Health Serv Res. 2012; 47:2353–2376.

17.

Institute of Medicine, Committee on Geographic Adjustment Factors in Medicare Payment. Hospital Wage Index. Medicare Payment: Phase I: Improving Accuracy. Washington, DC: National Academies Press; 2011.

18.

Health Resources and Services Administration.. Primary Care Service Area Version 3.1 Methods. 2013. http://datawarehouse.hrsa.gov/DataDownload/PCSA/2010/PCSA_Version3.1_methods_9_17_2013.pdf. Accessed June 18, 2015 .

19.

Goodman

, Mick

, Bott

, et al. Primary care service areas: a new tool for the evaluation of primary care services. Health Serv Res. 2003; 38:287–309.

20.

Goodman

, Morden

, Ralston

, Chang

, Parker

, Weinstein

. Populations of HSAs and PSAs. In: Bronner

, ed. The Dartmouth Atlas of Children's Health Care in Northern New England. Lebanon, NH: The Dartmouth Institute for Health Policy and Clinical Practice; 2013:136.

21.

US Congressional Budget Office. Geographic Variation in Health Care Spending. Washington, DC: Congressional Budget Office; 2008.

22.

Institute of Medicine. Variation in Health Care Spending: Target Decision Making, Not Geography. Washington, DC: The National Academies Press; 2013.

23.

Centers for Medicare & Medicaid Services. Shared Savings Program. http://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/sharedsavingsprogram/index.html?redirect=/SharedSavingsProgram/. Accessed June 18, 2015 .

24.

Petersen

, Gardner

, Tu

, Muhlestein

Growth and Dispersion of Accountable Care Organizations: June 2014 Update. Salt Lake City, UT: Leavitt Partners; 2014.

25.

Centers for Medicare & Medicaid Services.. Fact Sheets: Medicare ACOs continue to succeed in improving care, lowering cost growth. 2014. http://www.cms.gov/Newsroom/MediaReleaseDatabase/Fact-sheets/2014-Fact-sheets-items/2014-09-16.html. Accessed June 18, 2015 .

26.

Noble

, Casalino

. Can accountable care organizations improve population health? Should they try?. JAMA. 2013; 309:1119–1120.

27.

Health Resources and Services Administration. Shortage Designation: Health Professional Shortage Areas & Medically Underserved Areas/Populations. http://www.hrsa.gov/shortage/. Accessed June 18, 2015 .

28.

Lopes

PM.

State-Wide Rational Service Areas for Primary Care Services: Lessons from Six States. Bethesda, MD: Health Resources and Services Administration; 2000.

29.

Adeoye

, Albright

, Carr

, et al. Geographic access to acute stroke care in the United States. Stroke. 2014; 45:3019–3024.

30.

Branas

, MacKenzie

, Williams

, et al. Access to trauma centers in the United States. JAMA. 2005; 293:2626–2633.

31.

MacKenzie

, Carlini

Characterizing Local EMS Systems. Washington, DC: National Highway Traffic Safety Administration; 2013.

32.

Federal Interagency Committee on Emergency Medical Services. 2011. National EMS Assessment. Washington, DC: National Highway Traffic Safety Administration; 2012.

33.

Morganti

, Alpert

, Margolis

, Wasserman

, Kellermann

. The state of innovative emergency medical service programs in the United States. Prehosp Emerg Care. 2014; 18:76–85.

34.

Brice

, Evenson

, Lellis

, et al. Emergency medical services education, community outreach, and protocols for stroke and chest pain in North Carolina. Prehosp Emerg Care. 2008; 12:366–371.

35.

Texas Department of State Health Services.. Trauma Systems History. 2011. https://www.dshs.state.tx.us/emstraumasystems/Etrahist.shtm. Accessed May 1, 2015 .

36.

Florida Department of Health. Amended Trauma Service Area Assessment. Callahan, FL: Florida Health; 2014.

37.

Mayo Clinic. Consensus-driven trauma criteria key to region's success. http://www.mayoclinic.org/medical-professionals/clinical-updates/trauma/consensus-driven-trauma-criteria-key-regions-success. Accessed May 1, 2015 .

38.

Barbera

, Macintyre

. Medical Surge Capacity and Capability. Washington, DC: US Department of Health and Human Services, Assistant Secretary for Preparedness and Response; 2010.

39.

Walsh

, Craddock

, Gulley

, Strauss-Riggs

, Schor

. Building health care system capacity to respond to disasters: successes and challenges of disaster preparedness health care coalitions. Prehosp Disaster Med. 2015; 30:112–122.

40.

Rambhia

, Waldhorn

, Selck

, Mehta

, Franco

, Toner

. A survey of hospitals to determine the prevalence and characteristics of healthcare coalitions for emergency preparedness and response. Biosecur Bioterror. 2012; 10:304–313.