This Article Abstract Submit a response to this article Services E-mail this article to a friend Alert me to new issues of the journal

T H E  U N I N S U R E D : 
P A Y I N G  F O R  C A R E
W E B  E X C L U S I V E   

12 February 2003  
  

How Much Medical Care Do The Uninsured Use, 
And Who Pays For It?

A large amount of money from existing government sources is potentially
available to finance expanded insurance coverage.


Jack Hadley and John Holahan

ABSTRACT:

With the number of uninsured people exceeding forty-one million in 2001, insuring the uninsured is again a major policy issue. This analysis establishes benchmarks for the inevitable debate over the cost of expanding coverage: How much is being spent on care for the uninsured, and where does the money come from? This information is essential for assessing how much new money will be required for expanded coverage, how much can be reallocated from existing sources, and how a new financing system would redistribute the burden of subsidizing care for the uninsured from private to public sources.

How much medical care do the uninsured use, and who pays for that care? How much do the uninsured pay for their own care? How much is “uncompensated” care provided by hospitals, clinics, and physicians?¹ What sources of funds do medical care providers use to cover the costs of uncompensated care, and, in particular, how much comes from governments?

These are critical questions in the debate over extending health insurance coverage to the uninsured, because it is important to distinguish real increases in costs resulting from expanded insurance coverage from transfers of existing costs from one financing source to another. Real increases occur because having insurance increases the amount of care used by the uninsured. Cost transfers, on the other hand, represent shifts from those who currently pay for the care received by the uninsured to those who would pay if coverage were expanded.

By estimating how much medical care the uninsured use and who pays for it, this analysis seeks to determine the resources that are already in the medical care system and potentially available to help pay for expanded insurance coverage. Knowing the existing sources of payment for care will help policymakers identify where some of the money for new coverage could come from.

Estimating the total new costs that would be induced by expanded insurance and the sources of financing is also critically important to the debate, but such estimates exceed the scope of this analysis. Also, we do not address “crowding out,” the substitution of public insurance for private insurance by people who now have private coverage. If large numbers of people switched from private coverage to a subsidized public insurance program, then the costs showing up on the government’s side of the ledger would be much higher, although there would be an offsetting reduction in private payments for medical care.

Study Data And Methods

We use two independent approaches to estimate the costs and sources of payment for care used by the uninsured, because no single data source provides complete, unambiguous, and precise information.² One approach uses household survey data collected by the Medical Expenditure Panel Survey (MEPS).³ MEPS is the best available household survey for our purposes because it obtains information on services used from household respondents and then contacts providers to identify amounts and sources of payment for the respondents’ care. We pooled data from 1996, 1997, and 1998 and updated each year’s costs to 2001 prices.⁴

The second approach uses data from various surveys of providers’ revenues and expenses and from government budgets and agency reports, to determine how much care they delivered to the uninsured and to identify the sources of payment. Hospital data come primarily from the annual survey of the American Hospital Association (AHA). Estimates of uncompensated care provided by clinics are derived from information collected by government agencies that contribute to clinics’ funding. Physicians’ uncompensated care is inferred from recent physician surveys.

Both approaches are subject to potential biases. The MEPS data may be prone to underreporting by the uninsured and subject to inherent problems of following up with providers about specific episodes of care. The MEPS estimates must also be adjusted for systematic differences in definitions and populations between MEPS and the National Health Accounts (NHA, kept by the Centers for Medicare and Medicaid Services, or CMS). The provider-based approach may be biased because existing accounting methods do not clearly identify uninsured patients or the cost of their care, and payments received by providers and allocations made by governments are seldom explicitly earmarked as paying for the uninsured.

Using two approaches provides a basic cross-check of each set of estimates. If they are relatively close to each other, than one can be more confident that the true amount of uncompensated care falls within a certain range. If the estimates are very different, then the reliability and potential sources of bias inherent in each approach must be evaluated to determine whether one approach can be judged to be closer to the true amount of uncompensated care.

Estimates From MEPS

Estimating the cost of care for the uninsured. The MEPS estimates are based on the civilian, noninstitutionalized population under age sixty-five, excluding people covered by Medicare.⁵ Newborns, people who died during the year, and people who were institutionalized for part of the year are included for the portion of the year that they satisfied the basic MEPS criteria for inclusion.

MEPS defines expenditures as “payments made for health care services,” which excludes the cost of services for which there was no explicit and identifiable payment linked to a specific patient (except for services provided by public hospitals and clinics). For example, MEPS does not count provider revenues from general government appropriations to hospitals and Medicare and Medicaid disproportionate-share hospital (DSH) payments, since they are not payments for specific patients.

Data adjustments. MEPS’ unique definitions and methods for measuring expenditures lead to much lower national health spending estimates than those reported by the CMS. To correct for MEPS’s systematic underreporting, we used information from a detailed comparison of the MEPS and NHA estimates to develop an adjustment factor to align the MEPS estimates with the NHA.⁶

We first subtracted the estimates of Medicare spending from both sources, since they are not direct payments for care received by the uninsured. We also subtracted from the NHA expenditures attributed to the Department of Defense for military personnel, revenues from non–patient care activities reported by health care providers, and spending for long-term nursing home and long-term hospital care, which are not likely to be included as covered services by a program to extend coverage to the uninsured. These adjustments reduced the NHA spending total from $912 billion to $556.1 billion. The MEPS/NHA adjustment factor we used to inflate the MEPS estimates is 1.25, the ratio of the revised NHA expenditure level to the comparable MEPS expenditure level.

We collapsed the sources of payment identified by MEPS into five groups: self (out-of-pocket), private insurance (including Tricare, CHAMPVA, and workers’ compensation), public insurance (Medicare and Medicaid), other public sources (Veterans Affairs, or VA, other federal programs, and other state and local programs), and other sources (other private and other, unknown source).⁷ Information on insurance coverage is reported on a monthly basis, which enabled us to distinguish full- and part-year uninsured and to identify insurance payments for people with part-year coverage.

We also adjusted the MEPS data to include an estimate of uncompensated care from private providers. (MEPS imputes the cost of uncompensated care provided by public hospitals and clinics.) The estimate was based on the question, “How much would providers have been paid if the uninsured had been covered by private insurance?” The difference between this estimate and the amount providers actually received in payment from explicitly identified sources other than private or public insurance is an estimate of the value of care delivered by private providers with no explicit payment linked to a specific patient.⁸

We generated the amount of expected payment using MEPS data on total charges for both privately insured and uninsured people. We calculated the ratio of payments to charges for those with full-year private insurance coverage and then applied this ratio to the total charges for care received by people who were uninsured for at least part of the year. (Charge information is provided even if there is no payment.) Overall, payments (from all sources, including out-of-pocket) for care received by the full-year privately insured covered 81.5 percent of providers’ charges for that care.

We applied this ratio to the total charges for care received by the full- and part-year uninsured, excluding care paid for by private insurance, public insurance, or other public sources, to estimate the total payments that providers would have received if the uninsured had been covered by private insurance. This calculation produced an estimated “expected” payment of $54.6 billion. Actual total payments from these sources (self, other private, and unknown) made for the full- and part-year uninsured were $38.8 billion, which implies that $15.8 billion of uncompensated care was delivered by private medical care providers.

Amount of “uncompensated care” provided. Exhibit 1 presents the estimates of medical care spending by insurance status and source of payment. People who were uninsured during any part of the year received $98.9 billion in care, of which $34.5 billion was “uncompensated care” (that is, not paid for either out of pocket or by a private or public insurance source). This represents 35 percent of the care received by the uninsured but only 2.8 percent of total personal health care spending of $1,235 billion in 2001.⁹ The other 65 percent was paid for out of pocket and, mainly for the part-year uninsured, by private and public insurance sources.

Uncompensated care accounted for 60 percent of the care received by the full-year uninsured, with almost all of the rest ($14.1 billion, or 35 percent) paid for out of pocket. The part-year uninsured also received a substantial amount of uncompensated care, $9.9 billion, which accounted for 17 percent of their overall care.

Even taking uncompensated care into account, the full-year uninsured received about half as much care ($1,253 per person) as the privately insured received ($2,484). While some of this difference is attributable to differences in age and health status between the two groups, research that takes these factors into account still finds about a 50 percent differential.¹⁰ Thus, even though uncompensated care is the primary source of care for the full-year uninsured, it does not make up for or offset the effects of being uninsured on access to and use of care.¹¹

Sources of financing. Including uncompensated care delivered by private providers, private sources appear to account for just over 81 percent ($28.1 billion) of all uncompensated care (Exhibit 2). However, this greatly understates the role of governments because, as shown below, private providers receive sizable government funds to cover the costs of the uncompensated care they provide. Among explicitly identified government sources, the VA is the largest single source of funding, although these public sources exclude tax appropriations, government grants, and Medicare and Medicaid payments to private hospitals and clinics for care of the uninsured.

Estimates From The Provider-Based Data

Hospitals’ uncompensated care costs. In 1999 hospitals incurred $20.8 billion, 6.2 percent of their total expenditures, in costs for patients who did not pay their bills.¹² This proportion is slightly higher than the 6.0 percent reported in 1998, but it has been relatively constant over the past decade. Applying the CMS projection of a 13.3 percent increase in Medicare hospital payments between 1999 and 2001 produces an estimate of $23.6 billion in uncompensated care costs in 2001.¹³

Clinics and direct care programs. Freestanding clinics and health centers that are part of the health care safety net are a major component of the medical care system used by the uninsured.¹⁴ These providers are both privately and publicly owned, and they receive funds from all levels of government as well as from private sources, including payments from patients. While many of these clinics serve substantial numbers of low-income people, not all of their users are uninsured. For example, although 88 percent of the patients served by federally qualified health centers (FQHCs) have incomes below 200 percent of the federal poverty level, only 39 percent are uninsured.¹⁵

We obtained data from the major federal programs that provide grants to clinics and health centers: the Bureau of Primary Health Care, the Maternal and Child Health Bureau, the HIV/AIDS Bureau, and the National Health Service Corps. We also obtained budget information for the VA, the Indian Health Service, and local health departments, all of which provide care to the uninsured.¹⁶ To the extent possible, we obtained information on budget percentages allocated to medical services and the proportions of charges or users identified as uninsured or self-paying. The latter was used to estimate the share of spending devoted to the uninsured.

We estimate that community health centers and other direct care providers delivered $7.1 billion in care to the uninsured in 2001 (Exhibit 3). The VA accounts for more than half of this total. Centers that receive support from the Bureau of Primary Health Care are the next-largest source of care for the uninsured.

Uncompensated care from physicians. We used two different studies to estimate the value of uncompensated care provided by physicians. An analysis of data from the American Medical Association’s (AMA’s) 1994 Socioeconomic Monitoring System found that 67.7 percent of physicians provided some uncompensated care and that those physicians spent an average of 7.2 hours per week delivering that care.¹⁷ Using an estimate (from the same study) of physicians’ average gross earnings per hour of $105 and inflating to 2001 prices produces an estimate of $9.1 billion in uncompensated care delivered by physicians. Another study, using different survey methods, found that physicians spent 2.6 hours per week in uncompensated care in 1999.¹⁸ Based on gross hourly earnings of $105, we estimate the total value of uncompensated care in 2001 from this source to be approximately $4.5 billion.

Some of the difference between the two estimates may reflect a true decrease in physicians’ uncompensated care, and some may reflect differences in survey design. A midrange estimate based on both studies would be $6.8 billion. Excluding an estimate of the amount of uncompensated care provided by salaried physicians employed by hospitals and clinics, which would be included in the previous estimates of hospital and clinic care to the uninsured, reduces the estimate of physicians’ uncompensated care to $5.1 billion.¹⁹

Combining the estimates for hospitals, clinics, and physicians produces a total estimate of $35.8 billion, which is only $1.3 billion more than the estimate of uncompensated care generated from the MEPS data.²⁰ Hospitals spent approximately $23.6 billion; clinics and community health care providers (including VA and Indian Health Service hospitals), about $7.1 billion; and physicians, $5.1 billion (based on the value of their time spent providing uncompensated care).

How Much Do Governments Spend To Care For The Uninsured?

Given the very similar MEPS and provider-based estimates of uncompensated care received by the uninsured in 2001, it seems reasonable to conclude that the amount of uncompensated care was around $35 billion. To develop estimates of what governments spend on care for the uninsured through either direct care programs or appropriations, grants, and payments to providers, we rely primarily on information from providers on the sources of their revenues that can be attributed to or justified by care to the uninsured. These estimates “get behind” the undifferentiated estimates of uncompensated care from the MEPS data and overall hospital uncompensated care from the provider data.

In developing the provider-based estimate of uncompensated care, we identified two major pieces of the funding-sources puzzle: private physicians’ uncompensated care, which represents $5.1 billion in private funding; and government grants and appropriations for clinics, the VA, and the Indian Health Service, which account for almost all of the $7.1 billion in uncompensated care delivered by such providers. Together, these two sources make up 35 percent of the funding for uncompensated care.
The funding sources for the remaining 65 percent, which is provided by hospitals, are more difficult to identify and require making assumptions about purposes and allocations of funds that are potentially available to cover the cost of uncompensated care. In particular, Medicare and Medicaid make payments in several different ways that are intended to help hospitals care for the uninsured. But these payments usually serve other purposes as well, and the share devoted to actually providing care to the uninsured must be estimated. We also assess state and local governments’ payments to hospitals and private sources of support for hospitals’ uncompensated care.

Medicare and Medicaid. Medicare and Medicaid make substantial hospital payments arguably intended to offset some of the costs of uncompensated care. Under Medicare’s prospective payment system (PPS) for hospital inpatient services, a DSH adjustment is applied to the payment rate for hospitals that treat a large share of poor patients.²¹ Hospitals received an estimated $5.0 billion in DSH payments in fiscal year 2001.²² The justification for these payments has increasingly been the need to protect those hospitals that serve a disproportionate share of low-income patients and are financially stressed and at risk of closing. The DSH payments are a way of preserving access for Medicare beneficiaries.

Medicare hospital payments also are adjusted for the indirect costs of graduate medical education (GME) programs through the indirect medical education (IME) adjustment. However, it has been estimated that the IME adjustment overcompensates hospitals by about one-third to one-half for teaching costs and that these payments also support teaching hospitals’ care of the uninsured.²³ IME payments in FY 2001 were estimated to total $3.7 billion.²⁴ If these payments are indeed one-third to one-half greater than justified by teaching hospitals’ higher costs, the remaining $1.2–$1.9 billion may be viewed as additional subsidies for teaching hospitals’ social missions. (Medicare also paid an estimated $2.1 billion in FY 2001 to reimburse teaching hospitals for the direct costs of their GME programs, but because these payments are tied directly to physician training, we do not include them in our tabulation of hospital payments for care of the uninsured.) Overall, we estimate that Medicare’s payments (through the DSH and IME programs) to support hospitals that treat poor and uninsured patients are between $6.2 billion and $6.9 billion a year.

Medicaid also earmarks substantial funds to support hospitals that treat a large number of poor patients. In FY 2001 the federal government contributed an estimated $8.9 billion to Medicaid DSH payments, with states contributing approximately $6.7 billion more.²⁵ However, some of the federal payment either goes to mental hospitals or is returned by providers to state treasuries and thus does not contribute to acute care. In addition, an unknown portion of the state contribution represents intergovernmental transfers and other financial transactions that may result in no net increase in state spending. Based on a recent survey of states, we estimate that 75 percent, or $6.7 billion, of the federal share is paid to general hospitals and that 25 percent, or $1.7 billion, comes from state general funds.²⁶ Therefore, the amount of DSH spending that appears to be available for the uninsured from both federal and state payments is $8.4 billion.

The use of supplemental, or upper payment limit (UPL), mechanisms is a newer approach to targeting additional funds to selected hospitals by raising their payment rates for the services they provide. As in the case of Medicaid DSH payments, there is often no net increase in state spending because of the use of intergovernmental transfers.²⁷ Nursing homes and hospitals received an estimated $11.6 billion through this mechanism in 2001.²⁸ The estimated federal contribution to these mechanisms was $6.6 billion, and the nominal state contribution, about $5.0 billion. Based on a recent survey, we estimate that hospitals receive $0.9 billion in revenue through federal payments and an additional $0.3 billion in state general fund revenue, or a total of $1.2 billion.²⁹

Total federal spending on Medicaid DSH and UPL payments that arguably go toward uncompensated care is therefore about $7.6 billion after payments to mental hospitals and nursing homes are excluded. The state contribution to this spending is about $2.0 billion, which reflects our assumptions about the use of phantom payments in attracting federal revenues in these categories.³⁰

Payments by state and local governments. Hospitals receive payments from state and local governments in the form of tax appropriations. The Medicare Payment Advisory Commission (MedPAC) treats these funds as reimbursement for care provided to uninsured patients. In 1999 hospitals received $2.7 billion in tax appropriations from state and local governments.³¹ This amount probably exceeds actual state and local support for uncompensated care, because there is no information on the purpose for which these revenues were intended. However, we know that they tend to go to hospitals that are major providers of uncompensated care in their communities, so even if they are not earmarked as support for uncompensated care, they probably serve that purpose to a large extent. Applying the CMS’s projected 13.3 percent increase in hospital spending between 1999 and 2001 yields an estimate of $3.1 billion from state and local governments’ tax appropriations as support for hospitals’ uncompensated care costs in 2001.

Another component of state and local government spending for uninsured patients is local indigent care programs, which presumably would not be necessary if there were other sources of insurance for these people. Hospitals provided a total of $3.8 billion in health care to enrollees in public programs other than Medicare, Medicaid, and military health programs in 1999.³² Applying the CMS’s 13.3 percent projection factor yields an estimate of $4.3 billion for 2001.

Private funding of hospitals’ uncompensated care. Private funds to support hospitals’ uncompensated care costs can come from two sources: the financial surplus hospitals may earn from care to privately insured patients or from non-clinical activities, and private philanthropy. The latter represents a fairly small source of hospital revenues, about 1–3 percent, and is often directed to activities that benefit all patients, not just the uninsured.³³

To estimate the amount of philanthropic revenue potentially available to support the cost of uncompensated care, we assumed that 2 percent of community hospitals’ revenues in 2001 ($7.8 billion) came from philanthropy.³⁴ Hospitals’ uncompensated care represents about 6 percent of hospitals’ expenses. If we assume that a somewhat higher proportion of philanthropic revenues—say, 10–20 percent—is allocated to support uncompensated care, then the amount used to pay for uncompensated care in hospitals from this source would be $0.8–$1.6 billion in 2001.

Although declining over time, hospitals’ financial surplus from private insurance in 1999 was $17.4 billion.³⁵ However, it is difficult to determine how much of the private surplus contributes to uncompensated care. Some hospitals with large uncompensated-care burdens and small surpluses may allocate the entire surplus to offset uncompensated care, while others may have surpluses well in excess of their uncompensated-care load.

We also know that the surplus from privately insured patients is not distributed across hospitals in proportion to the amount of uncompensated care they provide. In fact, past studies suggest that the two are inversely related: Hospitals with high levels of uncompensated care tend to have low proportions of privately insured patients. For example, a recent study of urban hospitals in nine states in 1994–95 found that 10.3 percent of safety-net hospitals’ patients were either self-pay or charity cases and that 19.9 percent were privately insured, while only 4.1 percent of non-safety-net hospitals’ patients were self-pay or charity cases and 38.8 percent were privately insured.³⁶

Given these discrepancies between shares of uninsured and privately insured patients across hospitals, it seems reasonable to assume that only a portion of the private payer surplus is being used to pay for care to the uninsured. If we assume that 10–20 percent of the profit from care to the privately insured subsidizes the cost of care to the uninsured, then the approximate amount of this subsidy in 2001 would range from $1.5 billion to $3 billion.

Exhibit 4 summarizes the estimates of funds potentially available to pay for the cost of uncompensated care provided by hospitals. In total, we estimate that $25.9–$28.2 billion was potentially available to hospitals to offset the reported $23.6 billion they spent for uncompensated care. Given that government money is not perfectly targeted to hospitals in proportion to the amount of uncompensated care they provide, it seems reasonable to assume that $2.3–$4.6 billion in available private funds also supports the cost of uncompensated care.

The balance of the financing of uncompensated care would then come from government sources. Some share of this may be misdirected—that is, some hospitals get more in government support than they provide in uncompensated care. Nonetheless, governments allocate $23.6 billion to hospitals primarily with the intent of providing care to the uninsured. The federal government provides $14.2 billion of this total, and state governments, $9.4 billion.

Exhibit 5 summarizes the estimated sources of funding underlying the $35 billion in uncompensated care in 2001. Governments spent $30.6 billion. Physicians’ donated time and forgone profits, and hospitals’ philanthropy and profit margins, were responsible for $7.5–$9.8 billion (21–28 percent) in private funding of uncompensated care. The federal government provided $19.9 billion, primarily through DSH, UPL, and IME payments to hospitals through Medicare and Medicaid, and through the VA. State and local governments spent $10.7 billion, allocated primarily to hospitals through tax appropriations and indigent care programs.

These estimates of funding sources exceed the amount of uncompensated care we have estimated. Some government payments may be poorly targeted or, alternatively, our estimate of uncompensated care may be too low. Nonetheless, we estimate that $30.6 billion of government money is “in the system” in the name of the uninsured.

Discussion And Policy Implications

Using two independent sources of data, we estimated that uninsured people received $35 billion in uncompensated care in 2001, or about 2.8 percent of total personal health care spending. Including the amount of uncompensated care, people who were uninsured all year averaged $1,253 per person in medical care costs. In contrast, people with full-year private insurance coverage received almost twice as much care. Thus, uncompensated care does not fully substitute for the lack of insurance, a result that is highly consistent with recent literature reviews that document sizable differences in health outcomes between the uninsured and the insured.³⁷

We also estimated that governments finance most of the uncompensated care received by the uninsured, spending about $30.6 billion on payments and programs largely justified to serve the uninsured and covering possibly as much as 80–85 percent of uncompensated-care costs through a maze of grants, direct provision programs, tax appropriations, and Medicare and Medicaid payment add-ons. Most of this money comes from the federal government, primarily through Medicare and Medicaid, followed by state/local tax appropriations for hospitals, Medicaid DSH and UPL payments, and the VA’s direct care programs. To place these estimates in perspective, total government spending in the name of the uninsured is considerably less than government spending on Medicare ($247 billion), Medicaid ($226 billion), and tax subsidies for private insurance ($138 billion).³⁸

Who benefits from increased spending? Most of the money for uncompensated care goes to hospitals, which deliver about two-thirds of such care. If all people were fully insured, hospitals would be the biggest beneficiaries, after the uninsured themselves, since they provide the majority of uncompensated care. Physicians, however, account for more than half of the private subsidies that underwrite the cost of uncompensated care. They too would benefit substantially from expanding insurance coverage to all Americans.

Since most of the current subsidies for uncompensated care come through Medicare and Medicaid payments and state/local tax appropriations to hospitals, it should be possible to transfer much of these funds to a new program to subsidize the cost of providing coverage for the uninsured since these programs’ constituents are not primarily uninsured people. If care were financed primarily through federal revenues, then state and local governments would be relieved of a major source of countercyclical financial pressure on their budgets, since the demand for uncompensated care tends to go up during recessions, when state and local governments’ revenues tend to decline. Medicare and Medicaid would also benefit by being better able to rationalize their program payments to address the needs of their primary beneficiaries, rather than having to distort payments to certain providers to help pay for uncompensated care delivered to the uninsured.

Targeted insurance versus a patchwork of programs. Another reason to prefer insurance over a patchwork of indirect and hidden subsidies to pay for uncompensated care is that payments would move with people and would be much better targeted to the providers actually providing the care. Current methods for allocating subsidies to hospitals, while generally on target, still overpay some institutions and underpay others relative to the amounts of uncompensated care they provide.³⁹

Money that pays for uncompensated care provided by the VA and for clinics and special care programs supported by government grants would be harder to reallocate, since these programs serve populations with special needs that might very well persist even with full insurance coverage. In the case of the VA, which accounts for almost $4 billion of the $7 billion in care delivered by direct care programs, the issue is complicated by the long history of its health care program. However, if uninsured veterans were eligible for subsidized public insurance, they would have much more freedom to make their own decisions about whether to receive care from the VA or from other providers. Over time, their behavior would undoubtedly affect the VA budget process.

Estimating how much care the uninsured would use if they had coverage is beyond the scope of this analysis. This estimate depends on how the coverage expansion is structured and, in particular, how much current payers (including the uninsured themselves) would be expected to contribute. Our analysis demonstrates that a fair amount of money is already in the system and that a substantial portion of the cost of covering the uninsured is potentially available from existing government sources. Adding in $13.8 billion in public insurance spending for people who are uninsured for part of the year (Exhibit 1), about half of the existing money spent on the uninsured already comes from governments.

This research was supported by a grant from the Henry J. Kaiser Family Foundation under the Cost of Not Covering the Uninsured project. Diane Rowland, Catherine Hoffman, David Rousseau, and Barbara Lyons of the Kaiser Commission on Medicaid and the Uninsured as well as the members of this project’s advisory group provided useful comments on earlier drafts of this paper. We also gratefully acknowledge the contributions of Stuart Guterman, who participated in the preliminary phase of this work while with the Urban Institute, and Teresa Coughlin; and of Mary Pohl, Heidi Kapustka, Wei Tan, and Marc Rockmore, who provided research and statistical assistance.

NOTES

1. Uncompensated care is defined as medical care the uninsured receive but do not pay for fully themselves. It includes reduced-fee care; charity care, for which the uninsured do not pay anything; and bad debts incurred by the uninsured. From providers’ perspectives, the “cost” of uncompensated care is the difference between the cost of the resources used to provide the care and whatever the uninsured pay themselves. These costs are paid for, wholly or in part, from other sources of revenue.
2. For details on methods and data sources, see J. Hadley and J. Holahan, “Who Pays and How Much? The Cost of Caring for the Uninsured,” 12 February 2003, www.kff.org/content/2003/4088/4088.pdf.
3. See J.W. Cohen, Design and Methods of the Medical Expenditure Panel Survey Household Component, MEPS Methodology Report no. 1, Pub. no. 97-0026 (Rockville, Md.: Agency for Healthcare Research and Quality, 1997).
4. All expenditure estimates were inflated to 2001 dollar values using information on annual increases in national personal health care expenditures from the National Health Accounts. Basing the inflation factor on changes in total personal health expenditures adjusts for changes in medical care price, utilization patterns, and overall population. Data are reported through 2000. It was assumed that expenditures increased 6 percent between 2000 and 2001, which was the same increase as between 1999 and 2000. Data were from the Centers for Medicare and Medicaid Services, “Health Accounts,” 3 December 2002, cms.hhs.gov/statistics/nhe/historical/t4.asp (6 January 2003).
5. Nonelderly Medicare beneficiaries either are disabled or have end-stage renal disease (ESRD), and their medical care use is not likely to be typical of either the uninsured population or the population with private insurance coverage.
6. T.M. Selden et al., “Reconciling Medical Expenditure Estimates from the MEPS and the NHA, 1996,” Health Care Financing Review (Fall 2001): 161–178. See Hadley and Holahan, “Who Pays and How Much?” for details.
7. Although responses from those reporting any Medicare coverage were deleted from the sample populations, some payments were identified by providers as coming from Medicare. This could represent ESRD payments for people who retain their private insurance coverage. The amounts were very small and could also include some reporting error.
8. We exclude from the calculations the share of the uninsured’s (full- and part-year) charges attributable to private insurance sources, public insurance sources, and other public sources. Differences between charges and payments for these sources of payment should be treated as either bad debts (from private insurance) or contractual allowances between providers and insurers and public programs.
9. CMS, “National Health Care Expenditures Projections Tables,” Table 5, 17 July 2002, cms.hhs.gov/statistics/nhe/projections-2001/t5.asp (21 October 2002). Includes Medicare, nursing home care, and other categories of spending that are excluded from the MEPS calculations.
10. M.S. Marquis and S.H. Long, “The Uninsured Access Gap: Narrowing the Estimates,” Inquiry 31, no. 4 (1994): 405–414.
11. Institute of Medicine, Coverage Matters (Washington: National Academies Press, 2001); and American College of Physicians, No Health Insurance? It’s Enough to Make You Sick (Philadelphia: ACP, 2000).
12. Medicare Payment Advisory Committee, Report to the Congress (Washington: MedPAC, March 2001), 182. This estimate was derived from the American Hospital Association’s Annual Survey of Hospitals by taking the sum of the charges reported by each hospital as charity care or bad debt (the two amounts are reported separately on the survey) and multiplying by the hospital’s overall ratio of total expenses to total charges. It includes not only inpatient acute care but also all other hospital services (outpatient, hospital-based skilled nursing and home health, and so on).
13. This estimate reflects two assumptions: (1) that the CMS’s estimate of hospital spending increase is applicable to hospital resource costs, and (2) that the proportion of hospital resources devoted to nonpaying patients is unchanged from 1999.
14. M.E. Lewin and S. Altman, eds., America’s Health Care Safety Net: Intact but Endangered (Washington: National Academies Press, 2000).
15. Bureau of Primary Health Care, “Uniform Data System (UDS) National Rollup Report,” National Summary for 2001, ftp://ftp.hrsa.gov/bphc/pdf/uds (10 July 2002).
16. National Association of State Budget Officers, 1998–1999 State Health Care Expenditure Report (New York: Milbank Memorial Fund, 2001).
17. D.W. Emmons, “Uncompensated Physician Care,” in Socioeconomic Characteristics of Medical Practice, 1995, ed. M.L. Gonzalez (Chicago: American Medical Association, 1995), 11–14.
18. M.C. Reed, P.J. Cunningham, and J.J. Stoddard, “Physicians Pulling Back from Charity Care,” Issue Brief no. 42 (Washington: Center for Studying Health System Change, August 2001).
19. One-quarter of physicians’ uncompensated-care hours are provided by physicians who are salaried employees of hospitals or clinics. Marie Reed, Center for Studying Health System Change, personal communication, 23 October 2002.
20. Hospital data are based on MedPAC, Report to the Congress,182. Clinic and government direct care programs are derived from Exhibit 3. Estimate for physicians is derived from Emmons, “Uncompensated Physician Care”; and Reed et al., “Physicians Pulling Back from Charity Care.”
21. Prospective Payment Assessment Commission, Report and Recommendations to the Congress (Washington: ProPAC, March 1997), 31.
22. Congressional Budget Office, “CBO April 2001 Baseline: Medicare” (Washington: CBO, 18 May 2001).
23. ProPAC, Report and Recommendations to the Congress, 28–29.
24. CBO, “CBO April 2001 Baseline: Medicare.”
25. Federal share estimate from CBO, “CBO March 2002 Baseline: Medicaid and the State Children’s Health Insurance Program” (Washington: CBO, 18 March 2002); state share based on assumption of average 57 percent federal match.
26. T.A. Coughlin and B. Bruen, “State Use of Medicaid UPL and DSH Financing Mechanisms,” Report prepared for the Kaiser Commission on Medicaid and the Uninsured (Washington: Kaiser Family Foundation, forthcoming).
27. T.A. Coughlin, L. Ku, and J. Kim, “Reforming the Medicaid Disproportionate Share Hospital Program in the 1990s,” Discussion Paper no. 99-14 (Washington: Urban Institute, January 2000).
28. CBO, “CBO March 2002 Baseline: Medicaid and State Children’s Health Insurance Program.” A 12 January 2001 regulation tightened the limits on the use of this mechanism.
29. Coughlin and Bruen, “State Use of Medicaid UPL and DSH Financing Mechanisms.”
30. Another issue that must be addressed in assessing Medicaid’s contribution to care for the uninsured is that payments net of DSH payments are well below costs for many hospitals. If Medicaid DSH payments were discontinued, some of those payments might have to be restored to the Medicaid program to reconcile payments and costs. Alternatively, the Medicaid program might change substantially or be eliminated, depending on the avenue through which universal coverage was to be achieved. In any case, the level of payment under the new program would have to be addressed.
31. The AHA’s annual survey asks hospitals to report separately patient revenues, tax appropriations, other operating revenues (from services other than patient care, such as gift shop sales), and nonoperating revenues (such as investment income and philanthropy). Following MedPAC’s approach, tax appropriations are included as an offset to uncompensated care expenses subject to the condition that they do not exceed uncompensated care expenses for any hospital; in other words, no hospital can have positive net revenues from uncompensated care. If tax appropriations exceed uncompensated care costs, the difference is accounted for as other government payments and subsidies, for example, as payments for medical education expenses or research. MedPAC, Report to the Congress.
32. This category also includes payments for patients covered by state-subsidized insurance programs, such as New York’s Home Relief Program and Washington’s Basic Health Plan. Because most—but probably not all—of these patients otherwise would have been uninsured, most—but probably not all—of these payments may be viewed as a consequence of the lack of general availability of health insurance.
33. D. Davison, “Is Philanthropy Dying at the Hospital?” Philanthropy,August/September 2001, www.philanthropyroundtable.org/magazines/2001/august/davison.html (6 January 2003).
34. Community hospitals’ total expenses in 1999 were $335 billion. American Hospital Association, Hospital Statistics (Chicago: AHA, 2001), 3. Trending this forward to 2001 and assuming a profit margin of 4 percent results in an estimate of total revenues of $393.7 billion.
35. Estimates are calculated from MedPAC, Report to the Congress,180.
36. D. Gaskin and J. Hadley, “Are Subsidies Allocated to Urban Hospitals in Accordance with Their Safety-Net Roles?” (Washington: Institute for Health Care Research and Policy, Georgetown University, 1999).
37. IOM, Care without Coverage (Washington: National Academies Press, 2002); and J. Hadley, Sicker and Poorer: The Consequences of Being Uninsured, May 2002, www.kff.org/content/2002/20020510 (6 January 2003).
38. Medicare and Medicaid figures are from CMS, “National Health Care Expenditures Projections,” Table 5, cms.hhs.gov/statistics/nhe/projections-2001/t5.asp; tax subsidy for private insurance is from L.E. Burman et al., “Tax Incentives for Health Insurance,” Tax Policy Discussion Paper (Washington: Urban Institute, 2002).
39. Gaskin and Hadley, “Are Subsidies Allocated to Urban Hospitals?”

Jack Hadley is a principal research associate at the Urban Institute in Washington D.C., and a senior fellow at the Center for Studying Health System Change. John Holahan directs Urban's Health Policy Center.

©2003 Project HOPE–The People-to-People Health Foundation, Inc.