Health Affairs, 23, no. 4 (2004): 234-244 doi: 10.1377/hlthaff.23.4.234 © 2004 by Project HOPE	New Online   How Would Obama, McCain Cover The Uninsured?   Debating Cost Of Uninsured   Try Medicare-For-All   HA Blog Top 10

This Article Abstract Reprint (PDF) Data Supplement Submit a response to this article Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services E-mail this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to Citation Manager Reprints & Permissions Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (16) Citing Articles via Google Scholar Google Scholar Articles by Cunningham, P. Articles by Hadley, J. Search for Related Content PubMed PubMed Citation Articles by Cunningham, P. Articles by Hadley, J. Related Collections Access To Care Business Of Health Consumer Issues Health Reform Insurance Coverage Managed Care Managed Care - Consumers Safety-Net Systems Health Spending

DataWatch

Expanding Care Versus Expanding Coverage: How To Improve Access To Care

Abstract

 
The Bush administration has proposed expanding insurance coverage as well as community health centers (CHCs) to increase access to care for uninsured people. This paper examines the relative effects of insurance coverage and CHC capacity on access to care. Communities that have both high insurance coverage and extensive CHC capacity tend to have the best access, although the former appears more important. Funding of insurance coverage expansions is likely to produce greater gains in access than if an equivalent level of funding were invested in CHCs. Policymakers should consider CHC expansions as complementary to insurance coverage expansions rather than as a substitute.

These two strategies represent two distinct ways of increasing access to care. Insurance coverage expansions are targeted directly at uninsured people and attempt to increase access by removing the financial barriers that most uninsured people report as their main reason for not getting needed medical care.¹ By contrast, expansions of CHCs and other safety-net providers are usually not targeted specifically at uninsured people, but rather at specific geographic areas that are considered to be medically underserved. Such an approach implicitly assumes that insurance coverage is not a sufficient guarantee of access for low-income people or that expanding insurance coverage is an inefficient way to improve access. Instead, CHCs attempt to increase access not only for uninsured people in the service area but also for lower-income insured people who may not have convenient access to other mainstream medical care providers or who face high out-of-pocket costs in using these providers.

Research from the past thirty years shows conclusively that health insurance coverage has a strong effect on access to care.² However, much less is known about the relative effectiveness of CHCs in improving access. The existing studies generally are limited to a specific geographic area or are based on users of CHCs.³ These studies tend to show that CHCs increase access to primary care, although access to specialty care remains a problem.⁴

This paper has two basic objectives: (1) to describe variations in insurance coverage and CHC availability across different U.S. population subgroups and communities; and (2) to examine the relative effects of increasing insurance coverage versus CHC expansions on access to care for low-income people.

Study Data And Methods

Top Study Data And Methods Study Results Discussion Editor's Notes NOTES

 
Data sources. The data for this study are based primarily on the 1998–99 Community Tracking Study (CTS) household survey, supplemented by data on CHCs. The CTS household survey is designed to produce representative estimates for the U.S. population and sixty randomly selected communities of individuals’ health insurance coverage, access to care, use of services, and perceived quality.⁵ The entire sample for the CTS sites includes about 55,000 people, although the primary sample of interest in this study includes nonelderly people (under age sixty-five) with incomes below 200 percent of the federal poverty level (n =12,700).

To examine the availability of CHC resources, data from the household survey were linked to data on CHCs from the Uniform Data System (UDS) maintained by the Bureau of Primary Health Care at the Health Resources and Services Administration (HRSA).⁶ The UDS includes detailed information on revenues, services, and users for all Federally Qualified Health Centers (FQHCs) that receive federal grants. These include community and migrant health centers, health care for the homeless programs, and public housing programs. "Look-alike" facilities that meet the statutory requirements of the FQHC program but do not receive direct federal grants are also included.

Measures of access to care. We examined several widely used measures of access to care and medical care use, including the likelihood of having a usual source of care; having an unmet medical need or postponing getting care in the past year; the likelihood of an ambulatory medical care visit; a visit for a routine health problem or general exam; and a visit to a hospital emergency department (ED). While higher levels of use usually denote better access, higher use of the ED is often associated with lack of access to more appropriate primary care sources.

Measuring CHC capacity. Our objective was to measure both the availability and capacity of CHC resources for the survey sample. To do this, individual CHC sites were linked to each person in the CTS household survey by ZIP code. We computed the distances (in miles) to every CHC within the CTS site for every person. We then constructed a measure that aggregates the total grant revenue across all CHC facilities that are within five miles of people’s residences.⁷

Analysis. After describing variations across populations and communities in safety-net capacity, insurance coverage, and access to care, our analysis then used multivariate regression methods to assess the relative effects of insurance coverage and CHC capacity on access to care for low-income people.⁸ The regression models estimated the quantitative impacts of individual insurance coverage and CHC capacity on the measures of access to care, controlling for the effects of other personal and area characteristics.

The statistical method used to estimate the regression coefficients adjusted for the potential bias due to reverse causality. That is, better access to care may not only be the result of having insurance coverage or higher CHC resources, but may also reflect (at least in part) decisions by those with greater health care needs or greater demand for care to enroll in insurance coverage or live close to CHC facilities. In the case of CHC resources, it may also reflect decisions to locate CHCs in medically underserved areas where access to care is known to be a problem, as well as the fact that the size of the federal grant for individual CHCs is based in part on the amount of uncompensated care that CHCs provide.

Study Results

Top Study Data And Methods Study Results Discussion Editor's Notes NOTES

 
Rates of coverage and CHC availability. Despite the fact that all but four of the sixty CTS sites have CHCs, only 38 percent of all people under age sixty-five lived within five miles of a CHC (Exhibit 1). Uninsured people were only somewhat more likely than privately insured people were to live close to CHCs (44 percent versus 36 percent), while about half of enrollees in Medicaid or other state coverage programs lived close to CHCs.

The greater availability of CHCs to lower-income people and racial/ethnic minorities largely reflects the criteria used to define medically underserved areas. Location in such an area is a necessary precondition for the location of federally qualified CHCs, and many of these include neighborhoods in large urban areas with high concentrations of poor people and ethnic minorities.⁹

Variation across communities. Much variation in insurance coverage and CHC availability also exists across the twelve communities that served as CTS case study sites.¹⁰ Rates of insurance coverage for nonelderly people ranged from 92 percent in Boston to 78 percent in Miami (Exhibit 2). The extent of variation in CHC capacity is even greater (even when Lansing, which has no CHCs, is excluded), ranging from 12 percent living within five miles of a CHC in Little Rock to 90 percent in Miami.¹¹

These findings reflect the fact that CHC resources are not allocated nationally based on the level of need (such as uninsurance rates) across metropolitan areas; rather, their distribution largely reflects local decisions to seek funding to create or expand CHCs. Also, applicants for CHC funding must demonstrate that they are part of a medically underserved area, a determination that takes into account poverty rates, infant mortality rates, concentrations of racial/ethnic minorities, and availability of primary care physicians but not explicitly the number of uninsured people. Since these areas are often defined for specific neighborhoods, rather than entire counties or metropolitan areas, need for CHCs often varies within communities as much as it does across communities.

Effects on access to care. To examine how differences in insurance coverage and CHC capacity may affect access to care among low-income people, we classified a subset of the sixty CTS communities into four groups, based on being in the top or bottom third of communities in terms of insurance rates and of having a strong or relatively weak CHC presence. Thus, the four groups include communities that were strong on both insurance coverage and CHC capacity, that were relatively weak on both measures, and that were strong on one measure and weak on the other. About half of the sixty sites were classified into these four groups, comprising about 58 percent of the low-income population. To examine differences in access to care among low-income people in these four groups, we computed adjusted means based on the regression model described above.¹²

The results in Exhibit 3 suggest several inferences. First, the highest levels of access to care were found in communities with both strong insurance and strong CHC presence. Low-income people in these communities had the highest observed levels of health care use among all four groups of communities, and the lowest levels of perceived difficulty getting care among all low-income people we studied.

Third, among communities with comparable levels of insurance coverage, access to care was generally better in communities with a strong CHC presence than in those with a weak CHC presence. For communities with low rates of insurance coverage, low-income people in communities with a strong CHC presence were more likely to have a usual source of care and more likely to have an ambulatory care visit, although increased ambulatory care use in communities with high CHC capacity appears to reflect greater use of hospital EDs than in communities with low CHC capacity.

The finding that greater CHC capacity appears to increase use of the hospital ED deserves further discussion. One might expect ED use to be lower in areas with strong CHC capacity, since CHCs provide a more appropriate source of primary care. In fact, earlier research with the same data showed that close proximity to a CHC did appear to reduce use of EDs among uninsured people, who rely on EDs as a source of primary care to a much greater extent than insured people do.¹³

However, higher ED use among all low-income people in areas with strong CHC capacity may reflect lack of access to specialty care rather than primary care. As noted in recent studies, CHCs often have difficulty providing or referring patients for specialized tests or treatments, and there are typically fewer specialists available for referrals in the medically underserved areas that CHCs serve.

Potential for further increases in access to care. Although the findings above indicate that both increasing insurance coverage and expanding CHCs will improve access to care, it is less clear whether equivalent levels of public investment in either approach will yield equivalent outcomes. To examine this more directly, we used the estimated regression models to simulate the effects on access of doubling CHC grant revenue nationally, from the current budget of approximately $2 billion to $4 billion. Such an increase would be much larger than the Bush administration’s CHC expansions, which increase direct funding by about one-third.

We then simulated the effects on access that would occur if the same amount of funding were used to expand insurance coverage. Based on previous research, our analysis assumed that expanding coverage by $2 billion would increase the coverage rate of low-income people from the baseline mean of 69.4 percent to 71.8 percent.¹⁴ For comparison purposes, we also simulated universal coverage’s effects on access.

The results of these simulations (Exhibit 4) show that on balance, insurance coverage expansions would lead to larger increases in access to care among low-income people than if an equivalent level of funding were used to expand CHC capacity. CHC expansions would produce an almost equivalent increase in the percentage with a usual source of care, as would insurance coverage expansions (from 78 percent to roughly 80 percent for both types of expansions), while CHC expansions would produce a somewhat higher increase in the percentage with any ambulatory care visit. However, insurance coverage expansions would result in larger increases in visits for general examinations, as well as a decrease in ED use (whereas CHC expansions would increase ED use, as discussed under the results for Exhibit 3). Insurance coverage expansions would also lead to somewhat larger decreases in the percentage who put off care or were unable to get needed care.

The simulations are intended to demonstrate the different magnitudes of the effects of insurance coverage and CHC expansions on access and to provide a rough approximation of what alternative policies to increase access could achieve. However, the simulations are limited by the fact that current funding levels for CHCs and insurance coverage programs are on vastly different scales. While a $2 billion expansion would double the current federal support of CHCs, it would represent a relatively small expansion in insurance coverage.

Discussion

Top Study Data And Methods Study Results Discussion Editor's Notes NOTES

 
It should come as no surprise that insurance coverage expansions are a more effective and efficient way to increase access to care among low-income populations, since this is consistent with the fact that lack of money—rather than lack of doctors—is the major reason why low-income people, and uninsured people in particular, have poor access to medical care.¹⁶ Interestingly, while greater investment in CHCs increases overall service use and may improve access to primary medical care (such as through more low-income people having a usual source of care), continued lack of access to specialty services and providers may explain why high CHC capacity does not reduce use of the ED or unmet medical needs.

Also, insurance coverage expansions have a natural advantage over CHC expansions because they directly target people with the greatest financial barriers to care. By contrast, CHC expansions target uninsured people indirectly and focus primarily on increasing the availability of providers in the urban neighborhoods and rural areas where they are in short supply. While such areas often have high concentrations of very poor uninsured people, many, if not most, uninsured people live outside of these areas. Instead, because most uninsured people are in employed families and not among the poorest Americans, they are likely to be dispersed more widely in middle-class urban and suburban areas that are not medically underserved but also have no easy access to CHCs and other safety-net providers.

Insurance coverage expansions also have their limits. Both public coverage and private coverage are voluntary, and many who are eligible for expansions choose not to enroll for a variety of reasons. In the case of providing subsidies for the purchase of private insurance coverage, out-of-pocket costs for premiums may still be a barrier for many low-income uninsured people, and deductibles and copayments may still impede service use. Administrative barriers in public programs (such as long application forms, face-to-face interviews, asset tests, and six-month redetermination periods) inhibit many eligible people from enrolling or reenrolling, and noncitizens (that is, recent immigrants) are often ineligible or afraid to enroll, even if they meet all other eligibility requirements.

Moreover, insurance coverage expansions must offer adequate incentives (that is, reimbursement rates) to physicians to take on newly insured patients. For example, low Medicaid rates relative to those of private payers often discourage many physicians from accepting Medicaid patients, even though the patients themselves incur little or no out-of-pocket expenses for their use of services.¹⁷ Thus, while coverage expansions generally are the most effective way to reduce financial barriers to care, new enrollees could still face other barriers to care and need to rely heavily on safety-net providers.

CHCs also provide services that are often unavailable or more difficult to obtain through private health care providers; this may not be fully accounted for in our results. For example, CHCs often provide important "enabling" services to low-income people who encounter language and transportation barriers. Also, CHCs may still be the preferred or only source of care for some services that are especially sensitive or frequently not covered by insurance or available through private health care providers, such as reproductive health, mental health, and substance abuse services. Finally, while provider capacity for low-income people in an area can also be increased to some extent by raising reimbursement rates for public programs, such incentives would likely be insufficient to cause many more physicians to locate in medically underserved areas that are otherwise served by CHCs and other safety-net providers.

In sum, insurance coverage expansions and safety-net expansions should not be viewed as alternative solutions to the same problem but as complementary approaches. Insurance coverage expansions should still be viewed as the main tool for removing the financial barriers to the primary and specialty care services that many uninsured people need but cannot afford. Even with these expansions, important gaps in services for vulnerable populations and underserved areas will remain. Bolstered by additional revenue from insurance coverage expansions, CHCs and other safety-net providers will still play an important role in filling these gaps.

Editor's Notes

Top Study Data And Methods Study Results Discussion Editor's Notes NOTES

 
Peter Cunningham (pcunningham{at}hschange.org) is a senior health researcher at the Center for Studying Health System Change in Washington, D.C. Jack Hadley is a principal research associate at the Urban Institute, also in Washington.

The authors thank Paul Ginsburg, Len Nichols, Richard Sorian, and Steve Zuckerman for providing helpful comments on an earlier draft. Excellent programming assistance was provided by Lucy Hsu of Social and Scientific Systems Inc.

NOTES

Top Study Data And Methods Study Results Discussion Editor's Notes NOTES

 

1. B.C. Strunk and P. Cunningham, "Treading Water: Americans’ Access to Needed Medical Care, 1997–2001," Tracking Report no. 1 (Washington: Center for Studying Health System Change, March 2002).
2. See J. Hadley, "Sicker and Poorer: The Consequences of Being Uninsured," Medical Care Research and Review 60, no. 2 (Supplement 2003): 3S–75S[Abstract/Free Full Text]; and Institute of Medicine, Care without Coverage: Too Little, Too Late (Washington: National Academies Press, 2002).
3. See summaries in R.M. Politzer et al., "Inequality in America: The Contribution of Health Centers in Reducing and Eliminating Disparities in Access to Care," Medical Care Research and Review 58, no. 2 (2001): 234–248[Abstract/Free Full Text]; and A. Dievler and T. Giovannini, "Community Health Centers: Promise and Performance," Medical Care Research and Review 55, no. 4 (1998): 405–431.[Abstract/Free Full Text]
4. M.K. Gusmano, G. Fairbrother, and H. Park, "Exploring the Limits of the Safety Net: Community Health Centers and Care for the Uninsured," Health Affairs 21, no. 6 (2002): 188–194.[Abstract/Free Full Text]
5. R. Strouse et al., "Report on Survey Methods for the Community Tracking Study’s 1996–97 Household Survey," Technical Publication no. 15 (Washington: HSC, November 1998).
6. Bureau of Primary Health Care, BPHC Uniform Data System Manual: 2000 Revision (Bethesda, Md.: BPHC, 2000).
7. In preliminary analysis we investigated alternative radii of 2.5, 5, and 10 miles for aggregating the measure of CHC capacity. The results were qualitatively similar, which suggests that the measure is not very sensitive to the selection of a radius within a ten-mile distance. We selected the five-mile radius as a reasonable approximation of what in reality are variable market areas that depend on population density and means of transportation.
8. Details of the multivariate statistical analysis can be found in J. Hadley and P. Cunningham, "Care vs. Coverage: A Statistical Analysis of the Effects of Safety-Net Availability and Insurance Coverage on Low-Income People’s Access To Care," available as an online data supplement, content.healthaffairs.org/cgi/content/full/23/4/234/DC1.
9. It should be noted that what is considered to be an "accessible distance" may differ for metropolitan and nonmetropolitan areas. Ideally, one would want to construct different thresholds for rural and urban areas for defining CHC capacity, although the differences in size and population density of nonmetropolitan areas make it difficult to derive a single standardized definition that would apply to all nonmetropolitan areas. Thus, we controlled for metro/nonmetro status and the size of the metro area in the analysis of the effects of CHC capacity on access.
10. These twelve communities were randomly selected from the sixty sites to serve as "case study" sites. As such, they are broadly representative of metropolitan areas with 200,000 or more people. In addition to in-depth visits to these twelve sites, survey samples for them were much larger than those of the other forty-eight communities, and therefore the precision of the site-specific estimates for these twelve communities is much greater.
11. One limitation of the measure of CHC capacity is that it is not standardized based on the potential number of users (that is, the number of people in the service area). Thus, communities such as Boston, Seattle, and Miami may have very high capacity in part because these CHCs serve very densely populated neighborhoods and thus have a larger potential pool of users within a five-mile radius than is the case in some smaller communities. In the analysis of the effects of CHC capacity on access, we at least partially controlled for this by including measures of metropolitan statistical area (MSA) type and size, as well as a county-level measure of population density.
12. The regression model was used to compute predicted values for individuals, assuming that they were included in one of the four groups in Exhibit 3. Group-specific means were computed for both insurance coverage and CHC capacity for each of the four groups in Exhibit 3. Regression-adjusted means were then computed for each of the four groups by using the group-specific means for insurance coverage and CHC capacity, and the actual values for all other variables. The regression-adjusted mean reflects the average of the predicted values for individuals.
13. J. Hadley and P. Cunningham, "Availability of Safety Net Providers and Access to Care of Uninsured Persons," Health Services Research (forthcoming).
14. The estimated increase in coverage rates resulting from a $2 billion coverage expansion is based on recent estimates of the incremental cost (in 2001 prices) of additional medical care that would be used by a person who obtains either private or public coverage after being uninsured for a full year, which was estimated at $1,015 per person on average. Extrapolating from these estimates, we estimated that $2 billion could provide coverage for an additional 1.97 million low-income people, or an increase of about 2.4 percentage points in the proportion of low-income people with insurance coverage. For estimates of the incremental costs of insurance coverage, see J. Hadley and J. Holahan, "Covering the Uninsured: How Much Would It Cost?" Health Affairs, 4 June 2003, content.healthaffairs.org/cgi/content/abstract/hlthaff.w3.250 (14 April 2004). The simulations were computed as the average of the predicted values for individuals based on the regression model. Baseline values of the access measures (Exhibit 4) were defined as the predicted values of each access measure computed using the baseline average level of insurance coverage (69.4 percent) and individuals’ actual values of CHC resource capacity. For insurance coverage expansions (Exhibit 4), the value of the probability of having coverage was increased by 2.4 percentage points to 71.8 percent for all people. To simulate CHC capacity expansion, individuals’ actual values of CHC capacity were increased to a minimum value of $3.25 million for people facing little or no CHC capacity; for people living in areas with larger CHC capacity, we assumed a 25 percent increase over the existing level of CHC capacity. These assumptions doubled the mean value of CHC capacity available from $3.1 million to $6.2 million, consistent with the simulation of a national doubling of CHC grant revenues from $2 billion to $4 billion.
15. The estimate of $35–$70 billion is based on Hadley and Holahan, "Covering the Uninsured."
16. Strunk and Cunningham, "Treading Water."
17. See, for example, S. Berman et al., "Factors That Influence the Willingness of Private Primary Care Pediatricians to Accept More Medicaid Patients," Pediatrics 110, no. 2 (2002): 239–248[Abstract/Free Full Text]; A. Coburn et al., "Effects of Changing Medicaid Fees on Physician Participation and Enrollee Access," Inquiry 36, no. 3 (1999): 265–279[ISI][Medline]; and J. Perloff et al. "Medicaid Participation among Urban Primary Care Physicians," Medical Care 35, no. 2 (1997): 142–157.[CrossRef][ISI][Medline]

                      What's this?

This article has been cited by other articles:

				N. L. Cook, L. S. Hicks, A. J. O'Malley, T. Keegan, E. Guadagnoli, and B. E. Landon Access To Specialty Care And Medical Services In Community Health Centers Health Aff., September 1, 2007; 26(5): 1459 - 1468. [Abstract] [Full Text] [PDF]

				J. Hadley, P. Cunningham, and J. L. Hargraves Would Safety-Net Expansions Offset Reduced Access Resulting From Lost Insurance Coverage? Race/Ethnicity Differences Health Aff., November 1, 2006; 25(6): 1679 - 1687. [Abstract] [Full Text] [PDF]

				H. Bailit, T. Beazoglou, N. Demby, J. McFarland, P. Robinson, and R. Weaver Dental safety net: Current capacity and potential for expansion. J Am Dent Assoc, June 1, 2006; 137(6): 807 - 815. [Abstract] [Full Text] [PDF]

				C. S. Lesser and P. B. Ginsburg Strategies to Enhance Price and Quality Competition in Health Care: Lessons Learned from Tracking Local Markets Journal of Health Politics Policy and Law, June 1, 2006; 31(3): 557 - 567. [Abstract] [PDF]

				M. K. Gusmano, V. G. Rodwin, and D. Weisz A new way to compare health systems: avoidable hospital conditions in Manhattan and paris. Health Aff., March 1, 2006; 25(2): 510 - 520. [Abstract] [Full Text] [PDF]

				C. R. Gresenz, J. Rogowski, and J. J. Escarce Dimensions of the Local Health Care Environment and Use of Care by Uninsured Children in Rural and Urban Areas Pediatrics, March 1, 2006; 117(3): e509 - e517. [Abstract] [Full Text] [PDF]

				P. J. Cunningham Medicaid/SCHIP Cuts And Hospital Emergency Department Use Health Aff., January 1, 2006; 25(1): 237 - 247. [Abstract] [Full Text] [PDF]

				D. Hargunani Healthcare Coverage for Children: Are We Making Progress? AAP Grand Rounds, April 1, 2005; 13(4): 46 - 46. [Full Text] [PDF]