Skip Navigation U.S. Department of Health and Human Services www.hhs.gov
Agency for Healthcare Research Quality www.ahrq.gov
Archive print banner

This information is for reference purposes only. It was current when produced and may now be outdated. Archive material is no longer maintained, and some links may not work. Persons with disabilities having difficulty accessing this information should contact us at: https://info.ahrq.gov. Let us know the nature of the problem, the Web address of what you want, and your contact information.

Please go to www.ahrq.gov for current information.

Chapter 5 — Safety Net Structure and Health System Context

Introduction

Understanding the structure of the local safety net and the local health care delivery system is critical for assessing the status and performance of a safety net. Having resources available to provide services for uninsured, low-income, and other vulnerable populations is important in meeting the needs of these populations. However, the ability of vulnerable populations to obtain timely and effective care and the performance of providers offering care to Medicaid and uninsured patients can also be affected by a broad range of other factors related to the local health care delivery system. These aspects of health system context include hospital ownership mix, level of competition among hospitals, the extent of managed care penetration, the degree of concentration of uncompensated care, the presence of facilities with an explicit mission to serve vulnerable populations (such as public hospitals, some not-for-profit hospitals, and Community Health Centers), and the supply of physicians.

Defining which providers constitute the local safety net can be difficult, with the discussions often laden with strongly held positions about who is a "true" safety net provider. The recent Institute of Medicine report defined the safety net as "those providers that organize and deliver a significant level of health care and other related services to uninsured, Medicaid, and other vulnerable patients," recognizing that most communities have a "core safety net" of providers. These providers have two distinguishing characteristics: "(1) by legal mandate or explicitly adopted mission they maintain an 'open door,' offering access to services for patients regardless of their ability to pay; and (2) a substantial share of their patient mix is uninsured, Medicaid, and other vulnerable patients."7

In this data book, we have not attempted to identify explicitly which providers are safety net providers and which are not, but rather provide as complete a description as possible of the overall provider mix. Included among our measures are

  • Hospital admissions by ownership type (public, not-for-profit, investor-owned).
  • Hospital admissions by teaching status (no, low, moderate, and major teaching).
  • Hospital outpatient capacity.
  • Presence of a Health Resources and Services Administration (HRSA)-sponsored Community Access Program initiative.
  • Presence of an HRSA-sponsored Community Health Center (refer to Chapter 4: Financial Support for Safety Net Services).

We also provide information on the "health system context" for the local safety net, including measures such as

  • Concentration and distribution of inpatient uncompensated care and Medicaid discharges.
  • Managed care penetration and extent of competition.
  • Physician supply per 100,000 population by type (adult primary care, pediatrics, obstetrics/gynecology, medical and surgical specialties).
  • Hospital bed supply and admission rates.
  • Emergency department visit rates.

Return to Contents

Variation in Safety Net Structure and Health System Context

Important differences exist across communities in the composition of the inpatient delivery system. For example, 256 counties, or 67 percent of those included in our data book, do not have a public hospital. Almost 37 percent of the total population in the areas contained in the data book live in counties with no public hospital. As documented in Table 5-1, public hospital presence is more common in cities than in suburban areas, with higher levels of public hospital market share in the South and West. In 22 counties-typically counties with smaller populations on the fringe of urban areas-the public hospital is the only inpatient facility.

Investor-owned hospitals are also unevenly distributed, with much higher levels of market penetration in the South (19.6 percent of hospital admissions) and the West (19.0 percent), and very low shares in the Northeast (1.8 percent) and Midwest (2.2 percent). In 19 counties in the data book, an investor-owned hospital is the only inpatient provider in the county-these counties are all suburban, and all but one are located in the South.

Table 5-1: Hospital Ownership Status by Area Type and Region, 1999

Area Percent of Admissions by Ownership Status
PublicInvestor-OwnedNot-For-Profit
MSA12.4%11.6%76.0%
Suburban County8.3%10.4%81.3%
Metropolitan County*14.6%12.2%73.2%
Northeast7.9%1.8%90.3%
South16.5%19.6%63.9%
Midwest6.0%2.2%91.8%
West16.8%19.0%64.2%
All Areas12.4%11.6%76.0%

* Metropolitan counties are those counties in metropolitan areas containing a city with 100,000 or more people.

The distribution of teaching hospitals also varies significantly, as shown in Table 5-2. Major teaching hospitals have a larger market share in metropolitan counties (30.7 percent) than in suburban counties (11.2 percent), with a very high market share in the Northeast (42.3 percent of all admissions). Overall, major teaching hospitals have a 24.1 percent market share in the areas included in the data book, although 24 of the 90 Metropolitan Statistical Areas (MSAs) have no major teaching hospital presence (these 24 MSAs represent 12 percent of the total population in the areas contained in this book).

Table 5-2: Hospital Teaching Status by Area Type and Region, 1999

Area Percent of Admissions by Teaching Statusa
NoneLowModerateMajor
MSA46.1%11.4%18.5%24.1%
Suburban County64.2%7.2%17.5%11.2%
Metropolitan Countyb36.7%13.6%19.0%30.7%
Northeast29.8%10.3%17.5%42.3%
South57.8%11.9%17.5%12.8%
Midwest52.5%7.1%13.7%26.6%
West47.3%14.3%22.7%15.7%
All Areas46.1%11.4%18.5%24.1%

a No teaching = no residents; low teaching = 1 to 4 medical residents per 100 staffed beds; moderate teaching = 5 to 14 medical residents per 100 staffed beds; major teaching = 15 or more medical residents per 100 staffed beds.
b Metropolitan counties are those counties in metropolitan areas containing a city with 100,000 or more people.

The competitive context of local health care markets and the ability of individual inpatient providers to respond to market pressures also differ substantially across metropolitan areas. One indirect measure of competition is the level of managed care penetration in an area, which in many circumstances can exert pressure to keep hospital charges lower. Managed care plans are often more aggressive in areas where they have large market shares. As illustrated in Figure 5-1 HMO market penetration is quite variable. In some MSAs in California and the Northeast, managed care penetration is greater than 60 percent, while in other areas-such as the Charleston-North Charleston, SC, and Kalamazoo-Battle Creek, MI, MSAs-the share is less than 20 percent.

Figure 5-1: HMO Market Penetration Metropolitan Areas, 1999

Figure 5-1: HMO Market Penetration Metropolitan Areas, 1999
[D] Select for Text Description.

The ability of a safety net hospital to respond to market pressures is a function not only of the overall price competitiveness of the local market but also of the hospital's own payer mix. If a hospital has a high level of uncompensated care and a small commercial payer base, it may face difficulties in shifting the costs of nonpaying patients by raising charges. One indicator in the data book, the cost-shifting index, provides a measure of this phenomenon, showing the average by which area hospitals must raise charges to commercial patients to make up for the revenue lost through the provision of uncompensated care. As illustrated in Figure 5-2 these levels are quite high in some areas, most notably Jersey City, NJ, and many metropolitan areas in Florida. Within these communities with high cost-shifting indices, individual hospitals with large uncompensated care loads and narrow commercial payer bases may be at substantial financial risk, as they find it difficult to "shift" these burdens in an increasingly competitive market.

Figure 5-2: Hospital Cost-Shifting Index Metropolitan Areas, 1999

Figure 5-2: Hospital Cost-Shifting Index Metropolitan Areas, 1999
[D] Select for Text Description.

Physician supply also differs across communities. As shown in Table 5-3 physician supply per 100,000 residents is consistently higher in the Northeast, with levels of adult primary care providers about 20 percent above average, pediatricians per 100,000 about 42 percent higher, and medical specialists 33 percent higher. The West has lower levels of physician supply across all types, up to 25 percent lower than national averages. Thirty-five counties have no pediatricians in the county-these counties are largely lower population areas on the fringe of urban centers, mostly in the South and Midwest.

Table 5-3: Physician Supply by Area Type and Region, 1999

Area Physician Supply per 100,000 Population
Pediatricians Adult Primary Care Providers Medical Specialists
MSA82.388.736.4
Suburban County72.273.429.3
Metropolitan Countyc87.796.940.2
Northeast116.6105.348.6
South81.080.136.0
Midwest70.091.932.3
West63.079.628.5
All Areas82.388.736.4

c Metropolitan counties are those counties in metropolitan areas containing a city with 100,000 or more people.

Return to Contents

How Safety Net Structure and Health System Context Are Related to Safety Net Performance and Population Outcomes

Little work has been done to document the impact of safety net structure and context on outcomes for vulnerable populations. Such an analysis is, of course, complicated by the myriad factors that can influence such outcomes and the difficulties of linking specific aspects of structure or context to outcomes.

In simple bivariate analyses examining individual measures and their association with outcomes, a relatively small level of association is observed. For example, as shown in Table 5-4, the mix of hospital ownership status in a community has little or no association with most measures of patient outcomes, although higher levels of investor-owned hospitals were associated with higher levels of "no usual source of care" or "no physician visit in the past year." Although these data cannot indicate a causal link between access problems and high levels of investor-owned hospitals, they do indicate that in communities with high levels of for-profit hospitals, levels of access problems are often also higher. Teaching hospitals are generally associated with higher levels of potentially preventable hospitalizations and worse birth outcomes. Again, no causal link can be made, and these data may simply reflect the fact that teaching hospitals are typically located in highly urbanized, central city areas that typically have more significant access problems. Go to the discussion in Chapter 7 for the results of analyzing multiple indicators together.

Table 5-4: Association Between Ownership and Teaching Status and Outcomes (Place/County and MSA Levels)

Outcome Measure Association With Outcome Measures (R2)d
Percent of Admissions in Public Hospitals 1999 Percent of Admissions in Investor-Owned Hospitals 1999 Percent of Admissions in Major Teaching Hospitals 1999
Place/County Level Preventable Hospitalizations, Ages 0-170.0030.0000.040+
Preventable Hospitalizations, Ages 18-390.011-0.0000.060+
Preventable Hospitalizations, Ages 40-640.0080.0000.069+
Late or No Prenatal Care0.010+0.0000.122+
Low Birth Weight (Full-Term Births)0.0010.0000.127+
Preterm Births0.0010.0040.043+
MSA Level Preventable Hospitalizations, Ages 0-170.0000.0080.185+
Preventable Hospitalizations, Ages 18-390.0170.0120.091+
Preventable Hospitalizations, Ages 40-640.0250.0050.115+
Late or No Prenatal Care0.0020.0000.116+
Low Birth Weight (Full-Term Births)0.0010.0010.132+
Preterm Births0.0150.0040.005
No Usual Source of Care (Low Income)0.0700.386+0.243-
No Physician Visit in Last Year (Low Income)0.0050.317+0.219-

d The higher the R2, the stronger the association. The "+" and "-" indicate the direction of the association. A "+" indicates that the outcome/performance measure increases as the factor increases, and a "-" indicates that the outcome/performance measure decreases as the factor increases.

Table 5-5 shows the relationships between physician supply and safety net outcomes. Similar to the findings for hospitals, physician supply has little association with rates of preventable/avoidable admissions or birth outcomes. However, there is a moderate to high association between lacking a usual source of care and having no physician visit in the past year, with higher levels of supply associated with lower levels of barriers to care.

Table 5-5: Association Between Physician Supply and Outcomes (Place/County and MSA Levels)

Outcome Measure Association With Outcome Measures (R2)e
Pediatricians 1999 Adult Primary Care Providers 1999 Obstetrician/Gynecologists 1999 Medical Specialists 1999
Place/County Level Preventable Hospitalizations, Ages 0-170.021+--0.038+
Preventable Hospitalizations, Ages 18-39-0.058+-0.072+
Preventable Hospitalizations, Ages 40-64-0.037+-0.035+
Late or No Prenatal Care--0.009+-
Low Birth Weight (Full-Term Births)--0.057+-
Preterm Births--0.039+-
MSA Level Preventable Hospitalizations, Ages 0-170.161+--0.159+
Preventable Hospitalizations, Ages 18-39-0.042+-0.085+
Preventable Hospitalizations, Ages 40-64-0.033+-0.050+
Late or No Prenatal Care--0.008-
Low Birth Weight (Full-Term Births)--0.110+-
Preterm Births--0.012-
No Usual Source of Care (Low Income)0.258-0.201--0.308-
No Physician Visit in Last Year (Low Income)0.409-0.336--0.398-

e The higher the R2, the stronger the association. The "+" and "-" indicate the direction of the association. A "+" indicates that the outcome/performance measure increases as the factor increases, and a "-" indicates that the outcome/performance measure decreases as the factor increases.

Not surprisingly, higher levels of emergency department visits are moderately associated with higher levels of preventable hospital use and worse birth outcomes (Go to Table 5-6). Emergency departments are often considered the "safety net for the safety net" and can be an important window on the performance of the health care delivery system for low-income and other vulnerable populations. High levels of emergency department use may indicate an inability to obtain care in other settings, or it may reflect dissatisfaction with or poor performance of the ambulatory care delivery system in an area. Interestingly, higher emergency room use is also associated with lower levels of lacking a usual source of care and having no physician visit in the past year, illustrating the difficulty of interpreting these data and the limits of some access to care measures. Higher levels of managed care penetration have low to moderate associations with lower preventable hospitalization rates and improved birth outcomes. Greater levels of uncompensated care (as reflected by the cost-shifting index) and an increasing concentration of discharges in high-burden hospitals (those with a cost-shifting index greater than or equal to 0.25) are slightly to moderately associated with higher levels of preventable hospitalizations and worse birth outcomes.

Table 5-6: Association Between Health System Context and Outcomes (Place/County and MSA Levels)

Outcome Measure Association With Outcome Measures (R2)f
Cost-Shifting Index, 1999 Percent of Discharges in High Burden Hospitals 1999 HMO Penetration Rate 1999 Emergency Department Visits per 1,000 Population 1999
Place/County Level Preventable Hospitalizations, Ages 0-170.111+0.057+-0.110+
Preventable Hospitalizations, Ages 18-390.081+0.039+-0.211+
Preventable Hospitalizations, Ages 40-640.080+0.059+-0.181+
Late or No Prenatal Care0.236+0.127+-0.072+
Low Birth Weight (Full-Term Births)0.112+0.046+-0.250+
Preterm Births0.051+0.023+-0.154+
MSA Level Preventable Hospitalizations, Ages 0-170.364+0.231+0.128-0.175+
Preventable Hospitalizations, Ages 18-390.214+0.110+0.190-0.165+
Preventable Hospitalizations, Ages 40-640.212+0.178+0.053-0.130+
Late or No Prenatal Care0.198+0.120+0.058-0.030
Low Birth Weight (Full-Term Births)0.130+0.036+0.178-0.237+
Preterm Births0.050+0.0140.209-0.123+
No Usual Source of Care (Low Income)0.0050.0460.0000.219-
No Physician Visit in Last Year (Low Income)0.0720.0010.0510.230-

f The higher the R2, the stronger the association. The "+" and "-" indicate the direction of the association. A "+" indicates that the outcome/performance measure increases as the factor increases, and a "-" indicates that the outcome/performance measure decreases as the factor increases.

These data can be useful in beginning discussions on the impact of the structure and context of the local safety net. More research is required to isolate the effects and interactions of these factors with one another and to provide information to policymakers for understanding the performance and potential vulnerability of a local safety net.

Return to Contents

Tales of Two Cities

The extent of these differences in safety net structure and health system context is perhaps best illustrated by comparing a few metropolitan areas. The populations of the Newark, NJ, and Portland-Vancouver, OR-WA, MSAs are roughly the same size, with approximately 2 million people. Yet the context in which these local safety nets must function differs enormously, as shown in Table 5-7. Portland represents a typical West Coast market, with relatively high levels of health maintenance organization (HMO) penetration (47.5 percent), while Newark is more representative of the Northeast, with lower levels of managed care (24.9 percent). The supply of hospital beds and physicians also differs dramatically, with levels in Portland (70 physicians per 100,000 persons and 1.7 beds per 1,000 persons) significantly lower than those in Newark (110 physicians per 100,000 persons and 3.6 beds per 1,000 persons), reflecting historic differences in medical practice and possibly the impact of managed care. Emergency department use in Portland is also significantly lower (253 visits per 1,000 persons) than that in Newark (390 per 1,000 persons), indicating different utilization patterns among these populations. The higher penetration of managed care in Portland, OR, may indicate a greater level of competition in the health care system than in Newark. However, the Portland safety net is less threatened by the potential disequilibrium that can result from large numbers of hospital providers having disproportionate levels of uncompensated care. The cost-shifting index in the Portland MSA is 5.7, compared with 21.0 in Newark, and the percent of hospital discharges in high-burden hospitals in Portland is only 11.7, compared with 31.5 in Newark.

Table 5-7: A Tale of Two Cities: Portland, OR, and Newark, NJ, 1999

  Portland, OR MSA Newark, NJ MSA
HMO Market Penetration47.5%24.9%
Physicians per 100,000 Persons70.4110.2
Hospital Beds per 1,000 Persons1.73.6
Emergency Department Visits per 1,000 Persons253.1390.2
Cost-Shifting Index5.721.0
Percent of Discharges in High-Burden Hospitals11.7%31.5%

Even larger differences in safety nets can exist within a single State. For example, in Orange County, CA, there is no public hospital, and 38.5 percent of admissions are in investor-owned facilities. In San Francisco, 21.2 percent of admissions are in public hospitals, and there are no investor-owned facilities in the county. San Francisco also has a substantial teaching hospital presence, with more than half of all admissions in major teaching hospitals, while Orange County has no major teaching hospital (Go to Table 5-8).

Table 5-8: A Tale of Two Counties: Orange County and San Francisco, CA, 1999

Percent of Admissions in Orange County San Francisco City and County
Public Hospitals0.0%21.2%
Not-for-Profit Hospitals61.5%78.8%
Investor-Owned Hospitals38.5%0.0%
Major Teaching Hospitals0.0%51.4%
Other Teaching Hospitals25.2%16.7%
Non-Teaching Hospitals74.8%31.9%

Return to Contents

Shifting Hospital Costs for Uninsured Patients

In most communities, the hospital costs of uninsured patients must be "shifted" to other paying patients. A few States have pooling systems to help even out these burdens, since some hospitals in a community have much higher levels of uncompensated care than do others. More typically, however, a hospital must raise charges to commercial patients to cover the expenses of patients who cannot afford to pay.

The ability of a hospital to "shift" these costs to other payers usually depends on three factors: (1) the size of the burden, (2) the hospital's payer mix, and (3) the competitive environment of the hospital. It is harder to pass along to commercial payers a large amount of uncompensated care than a small amount. For small amounts, a minor increase in charges to other payers can compensate for these costs; larger burdens, however, require a significant increase in charges. If the commercial payer base is small, the amount by which charges must be increased can be substantial. In a competitive health care market, where managed care plans are bargaining for the best price possible, hospitals with large uncompensated care burdens and/or a narrow commercial payer base are put at a substantial financial disadvantage, further threatening their payer base.

The data book contains four measures to help policymakers gauge the extent and potential impact of cost shifting in a community. The first of these is the "cost-shifting index," which is the average percent by which area hospitals must raise charges to commercial patients to make up for the revenue lost through the provision of uncompensated care. A high number indicates that a large amount of cost shifting is required by area hospitals. Second is the index of market concentration of Medicaid and uncompensated care discharges, which indicates the extent to which the market share of uncompensated care and Medicaid patients is concentrated in a small number of hospitals. Here a high number means that the burden is concentrated in a few hospitals and therefore not spread evenly throughout the community. The third measure is the percent of hospital discharges in the community that are in "high-burden hospitals," where a high number means that a relatively large portion of the market is facing potential cost-shifting problems. Fourth, a "Gini coefficient" indicates the proportion of area patients who would have to change hospitals to equalize uncompensated and Medicaid discharges across all area hospitals. In addition, the level of HMO penetration in the community is provided to give a proxy measure of the potential level of price competition in the community, with higher levels of managed care indicating potentially greater levels of competition, which may make cost shifting more difficult for high-burden hospitals.


7 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.


Return to Data Book Contents
Continue to Chapter 6

 

The information on this page is archived and provided for reference purposes only.

 

AHRQ Advancing Excellence in Health Care