In 2000, the Institute of Medicine released a report describing the health care safety net in the United States as "intact but endangered." In particular, the report emphasized the precarious financial situation of many institutions that provide care to Medicaid, uninsured, and other vulnerable patients; the changing financial, economic, and social environment in which these institutions operate; and the highly localized, "patchwork" structure of the safety net. One of the five key recommendations in the report concerns the need for data systems and measures:
"The committee recommends that concerted efforts be directed to improving this nation's capacity and ability to monitor the changing structure, capacity, and financial stability of the safety net to meet the health care needs of the uninsured and other vulnerable populations."1
1 Lewin ME, Altman S, editors. America's health care safety net: Intact but endangered. Institute of Medicine committee on the changing market, managed care, and the future viability of safety net providers. Washington, DC: National Academies Press; 2000. p. 213
In response to this recommendation, the Agency for Healthcare Research and Quality (AHRQ) and the Health Resources and Services Administration (HRSA) are leading a joint safety net monitoring initiative. An expert meeting in November 2000 provided an overview of the issues involved in establishing a monitoring system. Those attending the meeting recommended a monitoring system with four main goals:.
To accomplish these goals, there is a critical need to develop clearer knowledge of what needs to be measured, identify data and measures that are currently available, identify opportunities and strategies to develop data capacity, and assess the feasibility of monitoring these areas. Given the highly localized nature of the provision of safety net services, data and measures need to be examined at the smallest possible geographic level. This data book is the first product of the joint AHRQ/HRSA safety net monitoring initiative, bringing together 118 measures, largely from 1999, at the city, county, metropolitan, and State levels in two books. The first book, Monitoring the Health Care Safety Net—Book 1. Data for Metropolitan Areas, presents data from 30 States and the District of Columbia, including 354 counties and 171 cities in 90 metropolitan areas. Altogether, the data in this book describe the health care safety net in the places where 75 percent of the total American population lives and where 80 percent of Americans with family incomes below the Federal poverty line live. Figure 1-1 shows the places described in this book. The second volume, Monitoring the Health Care Safety Net—Book 2. Data for States and Counties, shows data from all 1,818 counties in these States, including both metropolitan and nonmetropolitan counties. A Web-based Safety Net Profile Tool (available at http://www.ahrq.gov/data/safetynet/) provides electronic access to the data and can be used to generate easy-to-use reports on geographic areas covered in these two books.
Figure 1-1: Areas Included in Data Book
The data and analyses in this book represent the first attempt to link the characteristics of the safety net to its outcomes on a widespread basis. Because the effort was limited to the use of existing data, geographic areas were selected to maximize the availability of data. Rates of preventable/avoidable hospitalizations are one key measure of safety net performance and outcomes; therefore, a prerequisite for inclusion of geographic areas in these books was the availability of hospital discharge data. We had access to such data for 30 States and the District of Columbia, all of which are included.
The goal of these data books is to help State and local health officials, planners, and analysts assess the capacity and viability of their existing safety net providers and understand the characteristics and health outcomes of the populations served. By presenting a set of common measures and indicators across many communities, these data books are intended to provide key information and benchmarks to inform policymaking and planning processes. The companion book, Monitoring the Health Care Safety Net—Book 3. Tools for Monitoring the Health Care Safety Net, offers strategies and concrete tools for assessing local health care safety nets. These books provide information that can help policymakers understand the status of their local and regional safety nets by providing measures and comparisons with other communities for a wide variety of characteristics of the safety net. Those who are working toward expanding the capacity of local safety nets and improving the health status of the populations they serve need such information to address factors that are within their control. In addition, understanding those factors over which they have little or no control may help inform their policy decisions. Similarly, valuable lessons can be learned from comparisons with other geographic areas and from analyses that lead to a better understanding of the determinants of safety net outcomes.
1 The Institute of Medicine report2 found that these types of data, which policymakers need for monitoring and improving the health care safety net, were either unavailable or incomplete. Building on that report, we have used resources from the U.S. Department of Health and Human Services to develop the first comprehensive array of safety net indicators and to assemble the data that support them, focusing on demand for services, financial support and structure of services, the health care and population context in which services are provided, and outcome/performance measures. In addition, we include analyses of the data to help policymakers begin to understand the determinants of health care outcomes and performance of the safety net. These data and analyses are presented in the context of a model that provides an overview of how the characteristics of the safety net and the factors that affect it relate to one another.
2 Ibid.
The Institute of Medicine defines the safety net as
"Those providers that organize and deliver a significant level of health care and other health-related services to uninsured, Medicaid, and other vulnerable patients."3
3 Ibid.[Page 21]
In particular, they define a group of "core safety net providers":
"These providers have two distinguishing characteristics: (1) by legal mandate or explicitly adopted mission they maintain an "open door," offering access to services to patients regardless of their ability to pay; and (2) a substantial share of their patient mix is uninsured, Medicaid, and other vulnerable patients."4
4 Ibid.[Page 21]
Core providers include a variety of health centers (e.g., Community Health Centers, Migrant Health Centers, the Health Care for the Homeless Program, School-Based Health Centers, and the Public Health Housing Program), community-based clinics, public hospitals, and many teaching hospitals as well. A substantial amount of safety net care is provided in hospital emergency departments, which, as a condition of participation in the Federal Medicare program, are required to provide medical screening exams and stabilizing treatment to all patients, regardless of their ability to pay. In addition, a considerable quantity of health care for safety net populations is provided in private physicians' offices. "Appendix B: Key Terms and Concepts" describes some of the ways health care for low-income populations is financed.
Assessing safety net viability and performance requires looking beyond institutions, buildings, and offices to include the study of the needs and health outcomes of the populations being served. While no single accepted definition of the population potentially served by the safety net exists, two common markers are used to help describe the population: income and health insurance coverage.
In 2001, 32.9 million Americans had family incomes below the Federal poverty line (a discussion of the Federal poverty line below). Many analysts expand this definition for the purposes of studying low-income populations and the health care safety net by studying the population below 200 percent of the Federal poverty line. In 2000, 200 percent of the Federal poverty line was an income of $35,920 for a family of two adults and two related children, and 85 million Americans had family incomes below this level.5
5 U.S. Census Bureau. Detailed poverty tables: 2001. Table 2. Available at: http://ferret.bls.census.gov/macro/032002/pov/new02_001.htm. Accessed April 24, 2003.
Populations served by the safety net typically lack health insurance coverage, are covered by Medicaid, or are low-income individuals with limited private insurance coverage (the underinsured). Throughout the entire first half of 2001, 46.0 million Americans were uninsured; the number who were uninsured at any point during that period would be larger. An additional 29.4 million Americans under age 65 had only public insurance coverage at some point during the first half of 2001.6 In total, depending on the definition used, between 75 and 85 million Americans constitute the potential population that may be eligible for safety net services. Some of these individuals will not need health care, use the safety net, or seek free or reduced-price care.
6 Medical Expenditure Panel Survey. MEPS household compendia of tables. Household health insurance tables: 2001. Table 1. Available at: http://www.meps.ahrq.gov/CompendiumTables/01Ch1/AllTables.pdf. Accessed April 24, 2003.
The composition and functioning of the health care safety net varies tremendously from one local area to another. In some areas, public hospitals are major providers of safety net care; in others, there may be none. Community Health Centers and other health center programs provide varying amounts of care in different areas. The types of transportation available, the number of providers willing to accept Medicaid patients, and the proportion of people who are uninsured all differ from one place to another. The informal phrases "all safety nets are local" and "when you've seen one safety net, you've seen one safety net" capture this concept well.
In these data books, we use data from the smallest local level available-including counties and cities-to provide this information. However, much can be learned from comparing different areas and understanding larger-scale patterns, so we also present information at more aggregated levels, including Metropolitan Statistical Areas (MSAs), States, and regions of the country, in an effort to provide a sense of context. For example, if a rate in a given county seems high, it may be helpful to understand the extent to which it is similar to the rates in surrounding areas or how it may differ from broader regional patterns. In addition, certain regional variations are known to exist in the U.S. health care system, including differences in levels of uninsurance, health maintenance organization (HMO) penetration, and public hospital presence. Chapter 2 describes the ways in which different levels of geographic information are useful in examining the safety net.
The need to understand barriers to accessing health care and disparities in health outcomes for vulnerable populations is receiving increasing attention among policymakers and health planners. This is particularly true as reducing the number of uninsured Americans remains a major challenge and all levels of government face budget constraints in the provision of health care for low-income populations. In this environment, the status and performance of the safety net is of growing concern. At the Federal level, efforts are under way to expand health center programs. At the State and local levels, policymakers are looking at alternatives to expanding insurance coverage to help ensure access to timely and effective care.
In developing these data books, there was a need to recognize the complex nature of access and the multiple contributors to disparities in health outcomes. No single factor can ensure "access to care," and no single measure of health outcomes can assure policymakers that "needed" care is being delivered and that the safety net is performing adequately. Accordingly, these data books are compiled on the basis of two major principles: (1) the underlying theoretical model for monitoring the safety net must recognize the multiple components of optimal health and the broad range of factors that can mediate how these factors affect health outcomes and safety net performance, and (2) multiple measures of health outcomes and performance must be incorporated to provide as complete a picture as possible for understanding potential access problems. These data books include three types of outcome and performance measures:
Although these measures reflect only a small range of the ideal outcomes for evaluating the safety net, these data books were limited by the need to use readily available data, as no new data were collected. In addition, the desire to show the same measures across all areas included in the books required that we limit the measures to those available for a broad range of areas. However, as described in Chapter 7, the range of outcomes and performance among communities as documented by these measures was large, and some safety nets performed well on some measures and poorly on others. Accordingly, while additional measures would be helpful, these measures effectively illustrate the complexity of the access problem and provide a useful baseline for understanding health outcomes and safety net performance in these communities.
The underlying theoretical model for monitoring the safety net and organizing these data books is reflected in Figures 1-2 and 1-3. First, it is critical to recognize that optimal health can be affected by several interrelated elements (Figure 1-2). Health is obviously affected by genetics, and the impact of genes on the incidence of chronic diseases, cancer, immune system response, and other conditions is becoming increasingly well understood. The impact of the environment on health outcomes is also well documented. In developing countries, improvements in sanitation and water quality often outweigh any other potential intervention in efforts to improve health outcomes. Even in the United States, the impact of air quality, water quality, housing (including lead paint exposure), and other environmental factors is significant, with low-income populations or particular neighborhoods often differentially exposed to detrimental environmental factors.
Figure 1-2: Assessing Optimal Health: Basic Model
While what you are born with and the environment in which you live are of obvious importance, how you live your life (lifestyle and behavior factors such as nutrition, exercise, substance abuse, and smoking) and how you manage your health (personal health maintenance efforts such as immunizations, checkups, and responding to symptoms of ill health) are also critical elements of optimal health. Differences among population subgroups on these factors are also well documented and can contribute to differences in health outcomes among populations and communities.
The availability of health care is a critical component in any model of optimal health. However, it is often difficult to measure, and the presence of resources alone is not adequate to capture the underlying concept. True availability of safety net care means that patients without insurance or with limited incomes can afford to see a provider (the door is "open"), can get an appointment, and have an acceptable care experience (including waiting times and being treated with dignity, respect, and cultural sensitivity).
Provider performance matters as well. An "available" provider may be a prerequisite, but there is no doubt that the clinical content of care, the ability to provide referrals and coordinate care, and the quality of information systems to support patient care can also have a direct and substantial impact on health outcomes. Provider performance is linked to another important component of health: disease/condition self-management. Once a patient is sick, optimal health outcomes are often directly related to the ability to adhere to complex treatment regimens, to identify the symptoms of acute flareups, and to self-manage the inevitable ups and downs of most chronic conditions. Hospitalizations for these chronic conditions, often resulting from lack of effective management, are among the most costly components of U.S. health care expenditures and differ dramatically among population subgroups.
These major elements of optimal health are mediated by a broad range of personal and contextual factors, which may differ dramatically among population subgroups and communities (Go to Figure 1-3). Among the personal factors that can have a mediating influence on these elements are knowledge about health habits, disease management, resource availability, and so on; perceptions of the health care delivery system (Is the door open to me? Will care be provided in a respectful and dignified manner? Will the doctor listen to me?); personal characteristics such as health beliefs, attitudes about self-efficacy, motivation and confidence to manage chronic conditions, attitudes toward benefits and risks; personal resources (health insurance coverage, income, informal support network of family and friends, and so on), and a whole range of personal circumstances, such as stress and competing life demands.
Figure 1-3: Assessing Optimal Health: More Complete Model
These personal factors do not exist in isolation, but are affected by and can collectively affect a series of contextual factors. Among them are the structure of the local health care delivery system, including the presence or absence of public resources, ownership mix of inpatient facilities, level of market competition, and the level of safety net support; the nature of the public health system; the community environment (from local crime rates to the quality of housing stock); the community infrastructure (the level of social capital as reflected by community-based organizations, the strength of public social service and educational resources, the adequacy of the local transportation system, and so on); neighborhood characteristics (location of transportation, neighborhood cohesion, and so on); the civic environment/civil culture (political will to meet the needs of disadvantaged populations, level of social cohesion, willingness and capacity for innovation and cooperative action among major institutions); and national, State, and local policies such as Medicaid, Community Health Center and other health center initiatives, and environmental programs.
Both personal and contextual factors influence the ability of a safety net to meet the needs of vulnerable local populations. These data books capture as many of these factors as possible, given the limitation of relying on existing data. Many of the personal characteristics are included in Chapter 3, which discusses the demand for safety net services (health insurance, poverty, disability, and prevalence of HIV/AIDS) and in Chapter 6, which provides an overview of the characteristics of communities (racial/ethnic composition, national origin, education, and so on). Information on community contextual factors related to the safety net is included in the Chapter 4 discussion of financial support for the safety net (the level of Medicaid coverage, disproportionate share payment levels to hospitals, presence of Community Health Centers, and uncompensated care pooling mechanisms); in Chapter 5, which describes the health system structure and context (hospital ownership mix, teaching hospital presence, level of competition, degree of uncompensated care cost shifting, bed and health workforce supply, and so on); and in Chapter 6, which provides data on issues such as the local economy, unemployment, housing stock, and crime rates.
This book also includes some preliminary analysis of how these personal characteristics and contextual factors affect health outcomes and the performance of the safety net. While further in-depth work is required to understand the relationships among these factors, Chapter 7 documents some interesting findings that have potentially important implications for policymakers, analysts, and researchers.
This volume, Monitoring the Health Care Safety Net—Book 1. Data for Metropolitan Areas, has three main parts.
This section includes Chapters 1 through 7 and provides an overview and synthesis of the information in the main tables that follow.
Provides an overview of the role of geography in examining the health care safety net.
Describes measures related to demand, including the size of the uninsured population, the percent of the population living below the Federal poverty line, the percent of the population with a disability, and the number of AIDS cases per 100,000 population.
Measures include the extent of Medicaid coverage, Disproportionate Share Hospital payments, percent of the population below 200 percent of poverty covered by Medicaid, and Community Health Center grants.
Measures include the types of hospitals in each local area by ownership and teaching status, the market concentration of uncompensated and Medicaid hospital discharges, number of physicians per capita, and number of emergency department visits per capita.
Measures include racial/ethnic composition, unemployment and crime rates, and levels of education.
Measures include inpatient hospitalizations that could have been prevented with better ambulatory care; prenatal care, low birth weight and preterm births; and the percent of the low-income population with no usual source of health care. This chapter includes analysis and synthesis of the relationship between the measures included in the previous chapters and safety net outcomes and performance.
Chapters 3 through 7 provide a sense of the range of variation of many of the measures as well as how the characteristics of the safety net relate to outcomes and performance.
This section includes 15 detailed tables with 118 measures, showing data for each of the 354 counties, 171 cities, and 94 county residuals in 90 metropolitan areas. More detail on each of the measures is included in "Appendix A: Technical Information." These tables show all data for all areas included in this book and can be used to look up data on specific areas, to compare across areas, or to compare one measure with another.
This section includes two appendixes. "Appendix A: Technical Information" provides information on the data and measures used, methodological information, and geographic details. "Appendix B: Key Terms and Concepts" addresses issues related to health insurance, other ways that health care for low-income populations is provided, medical debt, and poverty.
Monitoring the Health Care Safety Net—Book 2. Data for States and Counties provides 14 detailed tables with 114 measures for 1,818 counties in 30 States plus the District of Columbia. All measures in this book are the same as those in Book I, although data on self-reported measures of access to care are not available.
The Safety Net Profile Tool, available at http://www.ahrq.gov/data/safetynet/profile.htm, can be used to generate reports that compare multiple measures for one or more geographic areas. This Web site also provides electronic versions of all the data included in Books I and II.
Monitoring the Health Care Safety Net—Book 3: Tools for Monitoring the Health Care Safety Net is a collection of papers that describe methods for assessing the state of local health care safety nets.
Specific geographic areas were included or excluded based solely on the availability of data; additional information is included in Chapter 2 and in "Appendix A: Technical Information." The data themselves are drawn from more than a dozen sources. Detailed information on the data sources used for these books is also included in "Appendix A: Technical Information."
Every effort has been made to ensure that the existing data used for these books have been processed accurately. However, errors may still exist, either due to an oversight in the data processing or problems with the original data.