Monitoring the Healthcare Safety Net

Book 1. Data for Metropolitan Areas

Chapter 3 - Demand for Safety Net Services

Introduction

Demand refers to the extent of need for safety net services. In any given local area, demand is affected by a wide variety of factors, particularly the size of the population potentially using safety net providers and the intensity of their need for services. The number of people who are uninsured or are covered by Medicaid, the size of the low-income population, and the number of individuals with major health problems all have an impact on the demand for safety net services. At the individual level, these factors-largely related to poverty and poor health status-affect personal health maintenance and disease/condition management. In addition, insurance status, poverty, and poor health influence the personal circumstances and resources available to individuals and families for accessing needed health care.

The data books include several measures that capture various aspects of the demand for safety net services, such as

• Percent of the population under age 65 that is uninsured
• Percent of the population under age 65 that is uninsured with family incomes below 200 percent of the Federal poverty line
• Percent of the population with incomes below 100 percent of the Federal poverty line:
  • Total
  • Ages 0-17
  • Ages 18-64
  • Age 65 and older
• Percent of the population with a disability:
  • Ages 5-20
  • Ages 21-64
  • Age 65 and older

AIDS prevalence per 100,000 population. Additional details on these demand measures are included in "Appendix A: Technical Information." Other measures of demand, such as disease prevalence/incidence, are not available at the county level.

Variation in Demand for Safety Net Services

Table 3-1 compares several demand measures across geographic areas. The information shown includes all counties, cities, and county residuals included in this book. (Go to Chapter 2 and "Appendix A: Technical Information" for more information on geographic areas.)

Although there is substantial variation in our demand measures across geographic areas, comparisons often obscure the considerable variation within these areas. For example, Table 3-2 documents the extent of variation within these geographic areas in the percent of the population living below poverty. While suburban counties average 8.3 percent of their population below poverty, this figure ranges from a low of 1.8 percent in the county residual of Arapahoe County, CO (the part of Arapahoe County outside the city of Aurora), to a high of 25.3 percent in Crittenden County, AR.

Table 3-3 shows variation in the percent uninsured within Census regions, which is considerably less than the variation in poverty rates. Even at this lower level of variation, however, there is a fourfold difference in uninsurance rates in the Northeast, ranging from 7.7 percent in the Portland, ME, MSA to 31.6 percent in Jersey City, NJ. The lowest level of variation occurs in the Midwest, where a greater than twofold variation exists between the Minneapolis-St. Paul, MN, MSA at 8.7 percent and Chicago, IL, at 19.2 percent.

Figure 3-1 shows the percent of the population below 200 percent of poverty for all 90 MSAs, with a subset of the MSAs labeled. In the Northeast, MSAs such as Springfield and Worcester, MA, have the lowest uninsurance rates for their low-income populations, while Hartford, CT, and Jersey City, NJ, have the highest. In the South, the Tennessee MSAs have the lowest uninsurance rates among the low-income population, likely due to the TennCare program, which covers low-income residents through a State-sponsored, managed-care- based program. These data, however, are from 1999, and the TennCare program has changed considerably since then. In contrast, nearly three of every five low-income individuals in the Augusta-Aiken, GA-SC, MSA are uninsured. In the Midwest, St. Louis and Chicago have the lowest and highest rates, respectively. However, Chicago's Midwestern high of 36.7 percent of the low-income population lacking health insurance is far lower than the highest rates in the West, with San Francisco at 49.2 percent and Los Angeles at 45.0 percent.

[D] Select for text description.

In Table 3-4, tremendous variation in AIDS rates among metropolitan areas is evident. MSAs vary the most in the Northeast and the West, and considerably less in the South and Midwest. Of the MSAs included in this book, the San Francisco, CA, MSA has the highest rate of AIDS cases (2,113 per 100,000 population), while the Albany-Schenectady-Troy, NY, MSA has the lowest (94 per 100,000).

How Is the Demand for Safety Net Services in One Locality Affected by Neighboring Localities?

The demand for safety net services in any given community is affected by many factors. Key among these is the size of the population seeking services. In many urban areas, multiple political jurisdictions are in close proximity to one another, and there may be little regional planning related to the safety net. While each jurisdiction draws from its residential and business tax base to provide local services, individuals seeking health care may cross jurisdictional boundaries. Suburban populations may seek care in their central cities, and one suburb may draw population for health care services from an adjacent suburb. Given State funding for safety net services, including the Medicaid and State Children's Health Insurance Programs, providing safety net services may be more complex in areas near State borders, as residents from one State seek care in another. The extent to which one locality's safety net is in demand to provide services to the residents of another depends on where people live, where providers are located, payment systems, the availability of different types of services, and the transportation system.

How Demand for Safety Net Services Is Related to Safety Net Performance and Population Outcomes

Table 3-5 shows the relationship among four demand measures, representing different aspects of the need for services facing States and localities, and each of the outcomes or performance measures. (Additional information on the outcome measures is available in Chapter 1 and in "Appendix A: Technical Information.")

Higher levels of poverty have moderate to very strong associations with higher rates of negative outcomes, including potentially preventable hospitalizations for all ages examined and all three birth outcomes, at the place/county level. These relationships are not as strong at the MSA level, where higher poverty rates are moderately associated with increased preventable hospitalizations for children and for adults ages 40 to 64, and with higher rates of late or no prenatal care. An increasing proportion of the population of an MSA living below 200 percent of the Federal poverty line is moderately associated with an increasing probability of low-income individuals having no usual source of care and having no doctor visit in the past year.

Overall, the demand measures presented in Table 3-5 show a clear and consistent pattern, with greater safety net demand associated with worse outcomes. This pattern may reflect situations in which the demand for safety net services exceeds the local capacity available to provide them. Issues related to the supply of safety net services and financial support for them are discussed in the next two chapters.

Figure 3-2 shows the relationship between child and adult poverty based on all the geographic areas included in this book. A very strong relationship exists between child and adult poverty, with an R2 of 0.936, representing nearly perfect agreement between the two measures. The notable exceptions (outliers) are towns such as East Lansing, MI, and Chapel Hill, NC, where adult poverty is considerably higher than the levels of child poverty would indicate. It is not surprising, however, to find comparatively extreme poverty in college towns, where many students have little direct income.

[D] Select for text description.

The need for multiple measures of demand for safety net services may not be immediately evident. However, while poor individuals may be highly likely to be uninsured, there is virtually no relationship between the level of poverty in a geographic area and the proportion of its population that is uninsured. As Figure 3-3 shows, neither of these measures can be predicted by the other with any accuracy much greater than chance alone. Knowing only the level of poverty or the level of uninsurance in a given area provides limited information on demand for safety net services.

[D] Select for text description.

Figure 3-4 displays the ratio of the population below poverty in central cities relative to their respective suburban counties. In general, as might be expected, central cities have considerably higher poverty rates than their surrounding suburban areas. In the metropolitan areas where poverty is most concentrated in the central cities-as in Milwaukee, WI, Hartford, CT, and Baltimore, MD-the city poverty rate may be 4 to 6 times that in the suburbs. Notable exceptions to this pattern include the Fort Myers-Cape Coral, FL, and Bakersfield, CA, MSAs, where the central cities have lower poverty rates than the suburbs. In a few areas, such as Ft. Lauderdale, FL, and Modesto, CA, poverty is evenly distributed across the cities and the suburbs. A nearly one- to-one central city-to-suburb ratio of poverty may occur because an entire area is fairly well-off, as in the Fort Myers area, or because poverty is widespread throughout the metropolitan region, as in Jersey City, NJ.

[D] Select for text description.

A Tale of Four Cities: The Uninsured Population

States and localities have a wide variety of health care responses to similar levels of poverty within their populations. One of the most common strategies is to use Medicaid and/or the State Children's Health Insurance Program to extend health insurance coverage to low-income uninsured residents. The extent to which individuals enroll in such coverage varies for several reasons, including program budgetary constraints, ease of enrollment and continuing eligibility verification, and residents' perceptions of both their need for care and the stigma associated with the program.

Table 3-6 shows four MSAs that have similar population sizes (ranging from 403,070 to 480,091 residents) and a fairly narrow range of poverty rates, from 11.6 percent in Saginaw-Bay City-Midland, MI, to 16.0 percent in Modesto, CA. However, the proportion of the population that is uninsured in each of these areas, particularly the proportion of the low-income population that is uninsured, varies dramatically. At the high end, in the Augusta-Aiken, GA-SC, MSA, 28.6 percent of individuals under age 65 are uninsured, and 56.8 percent of the population is living on incomes below 200 percent of the Federal poverty line. This MSA crosses State boundaries: approximately 35 percent of the population lives in South Carolina, with the remainder residing in Georgia. This situation poses an added challenge when considering the provision of health care services to the low-income population in this area, as residents may cross State boundaries to reach their preferred or most convenient health care provider.