FAMILY OUT-OF-POCKET SPENDING FOR HEALTH SERVICES: A CONTINUING SOURCE OF FINANCIAL INSECURITY Mark Merlis June 2002 Support for this research was provided by The Commonwealth Fund through a grant to the Institute for Health Policy Solutions. The views presented here are those of the author and should not be attributed to The Commonwealth Fund or its directors, officers, or staff. Copies of this report are available from The Commonwealth Fund by calling our toll-free publications line at 1-888-777-2744 and ordering publication number 509. The report can also be found on the Fund’s website at www.cmwf.org. CONTENTS About the Author ........................................................................................................... vi Executive Summary....................................................................................................... vii Introduction .................................................................................................................... 1 Trends in Out-of-Pocket Spending, 1987–1996............................................................... 2 Out-of-Pocket Spending by Non-Medicare Families ....................................................... 4 Income and Health .................................................................................................... 5 Health Insurance ........................................................................................................ 7 Type of Insurance ...................................................................................................... 9 Spending by Type of Service.................................................................................... 12 Persistence of High Out-of-Pocket Spending ........................................................... 12 Factoring in Premiums: Consumer Expenditure Survey Data ......................................... 13 Policies to Address Underinsurance................................................................................ 15 Public Programs ....................................................................................................... 15 Private Insurance...................................................................................................... 18 Tax Treatment of Medical Expenses ........................................................................ 19 Scope of Services ..................................................................................................... 22 Appendix A. Comparing Employer and Nongroup Health Plans.................................... 25 Appendix B. Detailed Tables on Trends in Out-of-Pocket Spending, 1987–1996 .......... 28 References..................................................................................................................... 32 LIST OF FIGURES Figure ES-1 One of Six Families Spends More Than 5 Percent of Income on Out-of-Pocket Health Care Costs............................................... viii Figure ES-2 High Out-of-Pocket Costs Affect One of Six Individuals in Non-Medicare Families Over a Two-Year Period ....................... viii Figure ES-3 Low-Income Families Pay High Percentage of Income in Out-of-Pocket Expenses ................................................................ ix Figure ES-4 One of Four Low-Income Families with Health Problems Spends More Than 5 Percent of Income on Health Costs ................... x Figure ES-5 As Out-of-Pocket Costs Rise, Access to Care Becomes More Difficult...................................................................... x Figure 1 One of Six Families Spends More Than 5 Percent of Income on Out-of-Pocket Health Care Costs.................................................. 1 Figure 2 Per-Capita Annual Out-of-Pocket Expenses Declined in Constant 1996 Dollars..................................................................... 2 Figure 3 Out-of-Pocket Spending by Type of Service, 1987 and 1996.................... 2 iii Figure 4 Low-Income Families Pay High Percentage of Income in Out-of-Pocket Expenses ................................................................. 5 Figure 5 One of Four Low-Income Families with Health Problems Spends More Than 5 Percent of Income on Health Costs ................... 6 Figure 6 Families with Serious Health Problems Are Most at Risk for Costs Greater Than 5 Percent of Income, Even with Insurance .................. 10 Figure 7 High Out-of-Pocket Costs Affect One of Six Individuals in Non-Medicare Families Over a Two-Year Period ........................ 13 LIST OF TABLES Table 1 Percentage of Families with High Out-of-Pocket Expenses, 1996 ............. 4 Table 2 Prevalence of Reported Health Problems Among Non-Medicare Families, 1996............................................................. 6 Table 3 Prevalence of Specified Conditions and Effect on Family OOP Spending, 1996 .............................................................. 7 Table 4 Out-of-Pocket Expenses Among Non-Medicare Families as a Percentage of Family Income, by Insurance Coverage and Duration, 1996............................................................................. 8 Table 5 Out-of-Pocket Expenses Among Low-Income Non-Medicare Families as a Percentage of Family Income, by Insurance Coverage and Presence of Health Problem, 1996 ................................ 9 Table 6 Mean Total and OOP Expenses Among Low-Income Non-Medicare Families, by Insurance Coverage and Presence of Health Problem, 1996 ...................................................... 9 Table 7 Non-Medicare Families Insured for Entire Year, by Source of Coverage, 1996 ............................................................................ 10 Table 8 Mean Total and OOP Expense, by Source of Coverage and Presence of Health Problem, Non-Medicare Families with Full-Year Insurance, 1996......................................................... 11 Table 9 Mean Total and Out-of-Pocket Spending for Non-Medicare Families with Full-Year Employer Coverage, as a Share of Family Income and Service Type, 1996 ............................................ 12 Table 10 Individuals in Non-Medicare Families with High Out-of-Pocket Spending, 1996 and 1997.......................................... 13 Table 11 Percentage of Nonelderly Consumer Units with High Health Expenses, 1998 ................................................................................. 14 Table 12 Mean OOP and Premium Expense, 1998 ............................................... 15 iv Table 13 Percentage of Non-Medicare Families with Access Problems, by Out-of-Pocket Expense as a Percentage of Family Income, 1996 ........................................................................ 17 Table 14 Families with Medical Expenses Higher Than 7.5 Percent of Income vs. Tax Returns with Medical Expense Deduction, 1996 .................. 20 Table 15 Distribution by Income Class of All Returns, 2000 ................................. 21 Table 16 Hypothetical Distribution of an Income-Based Refundable Credit for Excess Health Spending, 1996 ..................................................... 22 Table 17 Percentage of Families with High Out-of-Pocket Expenses, 1996 ........... 23 Appendix Tables Table A-1 One-Person Deductible Under Employer Coverage and Nongroup Policies for MEPS Respondents, 1996 ............................. 26 Table A-2 Overall Coinsurance Rates Under Employer Coverage and Nongroup Policies, 1996 .................................................................. 26 Table A-3 One-Person Out-of-Pocket Limit, 1996 ................................................. 27 Table A-4 Prescription Drug Coverage, 1996.......................................................... 27 Table B-1 Components of Real Annual Out-of-Pocket Expenses for Health Services, by Type of Service, 1987 and 1996 ......................... 29 Table B-2 Percentage of Families with Out-of-Pocket Expense Greater Than 5 Percent and Greater Than 10 Percent of Family Income, 1987 and 1996..................................................... 30 v ABOUT THE AUTHOR Mark Merlis is an independent health policy consultant based in New Hope, Pennsylvania. His current or recent projects include: a study of options for improving private long-term care financing; an analysis of the impact of out-of-pocket medical spending on families; a report on possible reforms in the Medicare+Choice program; and technical assistance to the U.S. Department of Health and Human Services’ Regulatory Reform Initiative. Previously, Merlis was Senior Fellow at the Institute for Health Policy Solutions, a senior health policy analyst for the Congressional Research Service, and an administrator in Maryland’s Medicaid program. vi EXECUTIVE SUMMARY The shift to managed care plans in the 1990s brought changes in health insurance benefits that included less stringent cost-sharing requirements for most families with insurance. Despite growth in overall medical care spending, direct out-of-pocket (OOP) spending by families was the same in 1996 as in 1987, and average spending as a share of family income declined. Nonetheless, there remain millions of families who face high OOP costs. In 1996, 16.3 percent of families spent 5 percent or more of their incomes on direct payments for medical services; 7.4 percent spent 10 percent or more. This report uses data from the 1996 Medical Expenditure Panel Survey (MEPS) household component and several additional data sources to examine trends in OOP spending, the components of that spending, and the characteristics of families with high OOP costs. Throughout, “family” is defined to include single individuals as well as families of two or more persons. The report finds that OOP spending on health care services remains a major source of financial insecurity for people with inadequate health insurance coverage. Those most at risk include Medicare beneficiaries, whose poor health and limited Medicare benefits can impose heavy financial burdens. Working families are affected too—while the growth of managed care brought them lower levels of cost-sharing and better financial protection, the rising cost of prescription drugs and rising premiums increasingly threatens this protection. At the bottom of the economic ladder, some families may be forced to forgo spending on necessities to meet the out-of-pocket cost of health care. And, while the uninsured are especially at risk, even those with privately purchased individual health insurance can fall victim to burdensome outlays for health care or be forced to forgo needed care. KEY FINDINGS One of six families faces high out-of-pocket expenses. High OOP costs affect millions of Americans. In 1996, roughly 18 million families—16 percent of all U. S. families—spent more than 5 percent of their incomes on OOP expenses for health services. Seven percent of all families spent more than 10 percent of income on OOP expenses (Figure ES-1). This does not include premium payments for health insurance coverage. vii Families headed by Figure ES-1 One of Six Families Spends More Than 5 Percent elderly or disabled of Income on Out-of-Pocket Health Care Costs Medicare beneficiaries are Percentage of families more at risk for 40% 35% burdensome health care costs than working families. 20% 16% 18% 12% Such families tend to be 7% 6% sicker and poorer, to use 0% more health services, and to All families Family Family All families Family Family head has head head has head have more limited benefits Medicare without Medicare without Medicare Medicare than those covered under Spends more than Spends more than employer-based health 5% of income 10% of income insurance plans. The Source: Author’s analysis of data from MEPS 1996. disabled are especially at risk: one-third of families headed by a Medicare beneficiary younger than 65 spent more than 5 percent of income on health care services. Given the special issues that surround Medicare and the extensive literature on gaps in Medicare benefits, the following findings focus on “non-Medicare” families—those with no member covered by Medicare during the year. More families are Figure ES-2 High Out-of-Pocket Costs Affect One of Six Individuals potentially at risk over in Non-Medicare Families Over a Two-Year Period longer periods. One of Percentage of individuals in a non-Medicare family 10 people in non-Medicare with OOP expenses greater than 5% of income households in 1996 was in 20% 17% a family that spent more 15% than 5 percent of income. 10% However, one of six was in 9% 10% a family that incurred 5% expenses over this threshold in either 1996 or 1997 0% (Figure ES-2). Over this 1996 1997 1996 or 1997 two-year period, 17 Source: Author’s analysis of data from MEPS 1996. percent of individuals in non-Medicare families were in families with OOP costs higher than 5 percent of family income. These costs are not always the result of a chronic condition; they are sometimes related to a single, unexpected health event such as an injury or stroke. Thus, the spectrum of families at risk for high OOP expenses is much wider than those affected in any given year. viii One of three low- Figure ES-3 Low-Income Families Pay High Percentage income families faces of Income in Out-of-Pocket Expenses high out-of-pocket costs. Low-income families 30% 5% or more of income 10% or more of income 23% are even more likely than other families to face 20% 14% 15% relatively high OOP costs, 10% 10% 10% because the 5-percent-of- 6% 4% 3% 2% income threshold is lower. 1% 0% In 1996, almost one- Total Less than 100%– 200%– 400% or quarter of families (23%) 100% 199% 399% more with incomes below the Family income as a percentage of poverty poverty line spent more Source: Author’s analysis of data from MEPS 1996. than 5 percent of income on OOP health care expenses. In contrast, only 3 percent of families with incomes above 400 percent of the poverty line crossed this threshold (Figure ES-3). Employer group coverage provides better financial protection than individual private coverage. Among families with insurance, those who are most at risk for high OOP costs are those with individual private coverage. These plans tend to have higher deductibles and coinsurance payments, and are less likely to cap patient liability for health care expenses or to cover prescription drugs. One of five families with private nongroup plans spends more than 5 percent of income on health expenses, compared with one of 12 families with employer coverage. Serious health problems lead to serious out-of-pocket expenses. Families that have a member with any health problem are twice as likely as other families to spend a high portion of their incomes on health services. Chronic conditions place families at the highest risk. For instance, 25 percent of families reporting a member with heart disease also report spending more than 5 percent of their income on OOP expenses. Other conditions especially likely to lead to high OOP costs include diabetes, mental disorders, high cholesterol, and back problems. Poverty and poor health multiply exposure to high health costs. Among families with serious health problems, low-income families face an especially high hurdle when it comes to OOP expenses. One-quarter of families with any health problem and incomes below 200 percent of poverty spend 5 percent or more of their incomes on OOP ix costs. Higher-income Figure ES-4 One of Four Low-Income Families families with a health with Health Problems Spends More Than problem are about one- 5 Percent of Income on Health Costs third as likely as lower- Percentage of non-Medicare families spending more than 5% of income income families to incur Total No health problem Any health problem high OOP costs (Figure 30% 26% ES-4). 19% 15% 13% 12% 10% 9% One of four 7% 6% 5% families with a serious 0% health problem who are All incomes Less than 200% 200% of poverty covered by private of poverty or more individual insurance is Source: Author’s analysis of data from MEPS 1996. at risk for high outlays. Health insurance does not necessarily protect families against the high costs of getting sick, especially if the coverage is private individual insurance. Twenty-seven percent of families with serious health problems who are covered by private individual insurance spend 5 percent or more of their incomes on OOP costs, compared with 13 percent of families with serious health problems who are covered by employer group insurance. Families with Figure ES-5 As Out-of-Pocket Costs Rise, high out-of-pocket costs Access to Care Becomes More Difficult also face access barriers. Percentage of non-Medicare families, 1996 Families with high health 40% Went without neede d care care expenses also face Difficulty for financial / insurance reason greater difficulties than 21% 19% other families in obtaining 20% 17% 12% 10% needed services. Of those 7% who spend more than 10 0% percent of income on direct Less than 5% of 5% or more of 10% or more of income income income health care expenses, 17 percent report going Amount spent on out-of-pocket costs without needed care, and Source: Author’s analysis of data from MEPS 1996. one of five reports difficulty obtaining a service for financial or insurance reasons (Figure ES-5). As expenses rise, families may delay seeking care; if they are slow to pay bills, health care providers may be less willing to provide them with further care. x While some families experienced high costs in 1996, out-of-pocket costs for the average family declined in real dollars between 1987 and 1996. During a period of rapid growth in managed care, the comprehensiveness of benefits improved for those with insurance. Traditional insurance often has large deductibles and patient cost- sharing of 20 percent for hospital and physician services; managed care plans typically require modest copayments, such as $10 per physician office visit. This reduction in patient cost-sharing and improved benefits contributed to a decline in average OOP costs adjusted for inflation—even though the numbers of people without insurance increased. While per-capita OOP spending stayed nearly the same in nominal dollars between 1987 and 1996, OOP costs per person declined in constant 1996 dollars from $495 per person in 1987 to $361 in 1996—a 27-percent decline. Similarly, the percentage of families who spent more than 5 percent of their incomes on direct payments for medical services declined from 20.2 percent to 16.3 percent over this period. Out-of-pocket costs for prescription drugs soared. Between 1987 and 1996, prescription drugs emerged as the leading source of OOP costs, displacing inpatient hospital services. In 1987, prescription drugs accounted for 19 percent of all OOP health care costs. By 1996, they accounted for 30 percent. Per-capita spending for prescription drugs increased by 20 percent in constant dollars. On the other hand, the growth in managed care greatly reduced patients’ costs for hospitalization. Managed care’s replacement of traditional fee-for-service insurance with high hospital deductibles is reflected in the decline in inpatient hospital OOP costs, from one-fifth of all spending in 1987 to less than 4 percent in 1996. When families’ share of health insurance premiums are added, almost one-third of families devote more than 5 percent of household spending to health care. The 1998 Consumer Expenditure Survey (CES), which includes information on both family health insurance premium payments and OOP health care spending, allows insights into the combined financial burden of premiums and direct payments for health care. The addition of premium payments to direct OOP costs doubles the number of families that allocate a high level of household spending to health care services. Nearly one-third of nonelderly families spend more than 5 percent of their incomes on health care when both premiums and OOP expenses are counted. Eleven percent of families devote 10 percent or more of household spending to combined OOP and premium costs. xi Current tax subsidies for health care are poorly targeted. The current tax code exacerbates the burdens of high OOP expenses on low-income families. At this time, taxpayers may deduct most kinds of medical expenses only if those expenses exceed 7.5 percent of gross adjusted income and if they itemize deductions. But poorer families tend to take the standard deduction because they are less likely to have other itemized deductions. Also, low-income families may not have any tax liability to begin with, so the deduction cannot reduce their liability, or they are eligible for the Earned Income Tax Credit, and the deduction cannot increase the credit. Four-fifths of the current tax expenditure resulting from the medical expense deduction goes to filing units with adjusted gross incomes of more than $50,000 a year. An income-scaled refundable tax credit for excess health spending could better target the existing tax expenditure. One model would funnel 82 percent of the benefit to those with incomes below $15,000 a year. CONCLUSION High OOP costs are a persistent problem within the health care system. They tend to affect the poorest families and those with the most serious health problems, and they present a major obstacle to obtaining necessary health care. The current health insurance system does not protect families from these costs. Most insurance plans, especially private individual insurance plans, provide no guarantee that their members will not encounter significant OOP costs. Even Medicare leaves many Medicare beneficiaries with high OOP costs because it fails to include prescription drug coverage and has high OOP deductible and cost-sharing requirements. By contrast, employer-plan benefits improved in recent years. However, rising premiums and rising prescription drug costs now threaten those benefits. Improved benefits and standards for health insurance plans may be necessary in order to assure that those with insurance are adequately protected against the high costs of medical bills and have access to care. Better targeting of the personal income tax code’s current tax incentives could reach those at greatest risk. xii FAMILY OUT-OF-POCKET SPENDING FOR HEALTH SERVICES: A CONTINUING SOURCE OF FINANCIAL INSECURITY INTRODUCTION The shift to managed care plans in the 1990s brought changes in health insurance benefits that included less stringent cost-sharing requirements for most families with health insurance. Despite growth in overall spending for medical care, average direct out-of- pocket (OOP) spending by families was the same in 1996 as in 1987, while average spending as a share of Figure 1 One of Six Families Spends More Than 5 Percent family income declined. of Income on Out-of-Pocket Health Care Costs Nonetheless, millions of Percentage of families families still have high 40% 35% OOP costs. In 1996, 16.3 percent of all U.S. families 20% 16% 18% 12% spent 5 percent or more of 7% 6% their incomes on direct 0% payments for medical All families Family Family All families Family Family head has head head has head services, and 7.4 percent Medicare without Medicare without Medicare Medicare spent 10 percent or more Spends more than Spends more than (Figure 1). Of families with 5% of income 10% of income employer coverage Source: Author’s analysis of data from MEPS 1996. throughout the year, nearly 8 percent had OOP costs that exceeded 5 percent of their incomes; one million of these families had costs that exceeded 10 percent of income. Those who relied on private nongroup coverage were even more vulnerable, as were families—regardless of coverage source—with one or more members in fair or poor health or with functional limitations. This report uses data from the 1996 Medical Expenditure Panel Survey (MEPS) household component and published results from the 1987 National Medical Expenditures Survey (NMES) to examine trends in out-of-pocket spending, the components of that spending, and the characteristics of families with high out-of-pocket costs. Throughout, “family” is defined to include single individuals as well as families of two or more.1 1 MEPS families consist of persons living together in the same household who are related by blood, marriage, or adoption, as well as foster children. They also include unmarried persons living together who consider themselves a family unit. Both the NMES and MEPS are confined to the civilian noninstitutional population. This means that the very large OOP costs for nursing-home residents are not reflected here. 1 The MEPS and NMES data include direct payments for medical services but exclude family payments for health insurance premiums or contributions to group health plans. To show how the addition of premium payments affects the estimates, this report uses data from the 1998 Consumer Expenditures Survey (CES), which includes both types of payments. Finally, the report considers policy options that might improve protection for families that face the greatest expenditures. TRENDS IN OUT-OF-POCKET SPENDING, 1987–1996 On average, Americans Figure 2 Per Capita Annual Out-of-Pocket Expenses paid the same amount out- Declined in Constant 1996 Dollars of-pocket—about $360 per capita—for medical services $600 $495 in 1996 as in 1987, even as overall per-capita medical $400 $361 care spending rose by 86 percent.2 In constant dollars, $200 average OOP spending declined (Figure 2). $0 1987 1996 Figure 3 shows that the components of OOP Source: Author’s analysis of data from MEPS 1996. spending shifted Figure 3 dramatically between 1987 Out-of-Pocket Spending by Type of Service, 1987 and 1996 and 1996. Cost-sharing for inpatient care and associated All other All other physician services accounted 15% Ambulatory 14% Ambulatory 30% 29% for 21 percent of OOP Rx spending in 1987, but just drugs Rx 15% drugs 4 percent in 1996. Separate 30% Inpatient 4% inpatient deductibles have Dental Inpatient 19% 21% Dental become rare in private 23% health plans, and many 1987 1996 managed care plans now Sources: 1987–NMES data from Taylor & Banthin (1994); 1996–author’s analysis of data pay in full for inpatient from MEPS 1996. services from network physicians. This change was almost entirely offset by rising costs for prescription drugs. The average annual OOP payments of people who received any 2 OOP spending is based on NMES and MEPS, but overall per-capita spending is calculated from the National Health Expenditures series and annual change in the U.S. resident population. This is because published NMES data reflect charges rather than actual expenditures. 2 prescription medication grew nearly 6 percent a year, from $105 in 1987 to $175 in 1996. In all likelihood, this trend has continued since 1996, as drugs have grown more costly and many plans have increased OOP contribution requirements. Appendix B gives detailed estimates of changes in OOP spending and spending distribution by age, sex, race/ethnicity, income, and insurance coverage. In 1996, 16 percent of families spent 5 percent or more of their incomes on OOP costs, compared with 20 percent in 1987. About 7 percent of families spent 10 percent or more of their incomes on OOP costs in 1996, compared with 10 percent of families in 1987. The slight decline in the proportion of families with high OOP costs relative to income reflects rising family incomes during a period when OOP spending levels stayed more or less constant. Even so, nearly 18 million families spent 5 percent or more of their incomes on OOP costs in 1996; of these, 8 million spent 10 percent or more of income. Not surprisingly, the likelihood that a family will devote a high share of income to direct medical costs depends on age, income, insurance coverage, and health status. Table 1 shows that families whose head was elderly or near elderly were much more likely to have high relative OOP costs in 1996, as were low-income families. (In this and all subsequent tables, the group of families spending 10 percent or more of income is a subset of the group spending 5 percent or more of income.) The effects of insurance are different for the elderly and nonelderly.3 In the latter group, families whose head had employer coverage or Medicaid were much less likely to spend a high share of income on OOP costs than other families. Families whose head had nonemployer private coverage were almost as exposed to high costs as were those whose head was uninsured.4 Among families with an elderly head, having any form of insurance that supplemented Medicare only slightly reduced the likelihood that the family would spend a high share of income on OOP costs. One reason for this may be that although more than three-fourths of family heads with Medicare had supplemental coverage for some part of the year, many with Medigap or Medicaid lacked coverage for prescription 3 Types of insurance coverage are defined as follows. Employer: coverage through an employer or union plan, from one’s own work or that of a family member. Other private: private nongroup coverage (including coverage purchased by self-employed people), nonemployer group coverage (such as association plans), coverage obtained through someone outside the household and private coverage whose source was not identified. Other public: public coverage other than Medicare, including Medicaid, other state programs, and CHAMPUS/CHAMPVA. 4 Categorizing families according to the characteristics of the family head may sometimes be misleading—for example, when the family head has insurance and other family members do not. Other ways of categorizing families are explored in the next section. 3 drugs. One study found that only 53 percent of beneficiaries had drug coverage throughout 1996 (Stuart, Shea, and Briesacher 2000). Table 1. Percentage of Families with High Out-of-Pocket Expenses, 1996 Percentage of Families with OOP Expense of: Persons in 5 Percent 10 Percent Families Families or More or More Characteristics of Family Head (thousands) (thousands) of Income of Income Total 109,482 268,905 16.3% 7.4% Age of family head 19 to 24 6,133 10,219 11.7 6.1 25 to 54 65,453 182,541 9.2 3.5 55 to 64 13,049 29,617 18.7 8.2 65 or older 24,321 45,472 35.0 17.6 Family income Less than $20,000 34,219 64,748 30.2 17.2 $20,000 to $39,999 30,327 72,085 15.9 5.1 $40,000 or more 44,936 132,073 6.0 1.4 Insurance coverage of family head Under age 65 Employer 58,337 157,662 8.2 2.6 Other private 5,413 10,960 19.2 9.4 Medicare 2,509 5,012 33.4 19.3 Medicaid 6,114 18,316 12.1 5.3 Uninsured 12,789 31,484 15.0 7.5 Age 65 or older Medicare only 5,522 10,642 38.3 20.1 Medicare and other public insurance only 3,217 6,343 26.1 14.8 Medicare and private insurance 15,244 27,398 36.1 17.6 Note: If the head of the family’s insurance coverage changed over the course of the year, the family is assigned to a coverage category in the following sequence: Medicare, employer, other private, Medicaid and other public. Uninsured families are those without coverage at any time during the year. Source: Author’s analysis of data from MEPS 1996. OUT-OF-POCKET SPENDING BY NON-MEDICARE FAMILIES Of late, public attention has focused on the gaps in Medicare coverage that can expose elderly and disabled people to high OOP costs, especially for prescription drugs. This problem has been extensively studied, often using the Medicare Current Beneficiary Survey, which has a much larger sample of Medicare beneficiaries than the MEPS.5 Here, however, the focus is on the experience of families who have other public or private 5 See, for example, MedPAC; Stuart, Shea, and Briesacher. 4 coverage or who are uninsured for part or all of the year. To do this, it uses the concept of a “non-Medicare family,” defined as a single person who was not a Medicare beneficiary, or a family, none of whose members was a Medicare beneficiary at any time during 1996. This way of classifying families drops from the analysis all elderly with Medicare and more than 9 million families that included both Medicare and non-Medicare members, or families with a member who shifted to Medicare during the year. More than a million of these families had OOP costs greater than 10 percent of income. In most of these cases, however, it was OOP spending for the Medicare beneficiary or beneficiaries that drove family spending over this threshold. Omitting these families is the clearest way of factoring out the well-known shortcomings of Medicare benefits and zeroing in on potential gaps in other forms of coverage. In 1996, there were 82.1 million non-Medicare family units. Of these, 10 percent, or 8.2 million, spent 5 percent or more of family income on medical care. Almost 4 percent, or 3.2 million families, had OOP expenses greater than 10 percent of family income. The remainder of this section examines the characteristics of these high-spending families. While the number of such families is sizeable, it should be emphasized that it still represents a relatively small fraction of all families, and attempts to characterize this group using a sample as small as the MEPS are subject to considerable error. Income and Health The likelihood of spending Figure 4 Low-Income Families Pay High Percentage a high share of income on of Income in Out-of-Pocket Expenses OOP costs drops as income rises (Figure 4). This is not 30% 5% or more of income 10% or more of income 23% surprising: If two families 20% with different incomes 14% 15% spend the same amount for 10% 10% 10% medical care, the family 4% 6% 3% 2% 1% with the lower income will 0% have spent a higher share. Total Less than 100%– 200%– 400% or However, it is also the case 100% 199% 399% more Family income as a percentage of poverty that low-income families are more likely to have Source: Author’s analysis of data from MEPS 1996. health problems—either because poverty contributes to poor health or because poor health reduces income. Table 2 classifies a family as having a health problem if any family member reported fair or poor health, fair or poor mental health, or any functional 5 limitation (defined as requiring assistance with any activities of daily living [ADL] or instrumental activities of daily living [IADL] or having other specified functional limitations).6 Families with incomes below 100 percent of the federal poverty level were 55 percent more likely to report a health problem than families with incomes above 400 percent of poverty. Table 2. Prevalence of Reported Health Problems Among Non-Medicare Families, 1996 Family Income Percentage of Families as a Percentage of Poverty Reporting Any Health Problem Less than 100% 52.1% 100%–199% 44.8 200%–399% 37.1 400% or more 33.6 Total 39.1 Source: Author’s analysis of data from MEPS 1996. Both income and Figure 5 One of Four Low-Income Families health status play roles in with Health Problems Spends More Than determining whether a 5 Percent of Income on Health Costs family will incur high OOP Percentage of non-Medicare families spending more than 5% of income costs relative to income. Total No health problem Any health problem 30% Overall, families with any 26% health problem, regardless 19% of income, were more than 15% 13% 12% 10% 9% twice as likely as other 7% 6% 5% families to spend a high 0% share of income for OOP All incomes Less than 200% 200% of poverty of poverty or more expenses (Figure 5). At the same time, lower-income Source: Author’s analysis of data from MEPS 1996. families, regardless of health status, were much more likely to incur high relative OOP costs. Which is the more important factor? Health is more important in the determination of whether a family will spend 5 percent or more of income. The greater prevalence of health problems among lower-income families accounts for 65 percent of the difference in the likelihood that lower- and higher-income families will incur expenses at this level. However, the presence of health problems and income class are about equally important in the determination of whether a family will spend 10 percent or more of income. 6 For each family member, these measures are reported for each of the first two of the three 1996 interview rounds. The table uses the poorest health or mental health status reported for any family member in either round 1 or round 2. 6 Having a family member with certain specific medical conditions significantly increases the likelihood that a family will have high expenses. The conditions listed in Table 3 are drawn from a list of fifteen “priority conditions” established for the MEPS. (Conditions not listed either made no significant difference in the likelihood of high expenses or occurred too infrequently for analysis.)7 Table 3. Prevalence of Specified Conditions and Effect on Family OOP Spending, 1996 Percentage of Percentage of Families Non-Medicare with OOP Expense of: Families Reporting 5 Percent or 10 Percent or Disease or Condition Condition More of Income More of Income Heart disease 4% 25% 12% Diabetes 4 22 8 Mental disorder 13 21 9 High cholesterol 4 18 7 Back problem 11 18 6 Hypertension 12 17 5 Cancer 3 17 7 Injury 7 16 6 Pulmonary disease 4 15 5 Asthma 7 13 5 All families (n = 82,148,298) — 10 4 Source: Author’s analysis of data from MEPS 1996. Health Insurance Whether or not a family has health insurance, and what type of insurance it has, would be expected to affect its level of OOP expenditures. While a relationship exists, it is not necessarily a straightforward one. Table 4 groups non-Medicare families into three categories: those in which all family members had some form of insurance throughout 1996; those in which no family member had insurance at any time during the year; and a middle group consisting of families that had some uninsured and some insured members or members who had insurance for only part of the year.8 7 In addition to the conditions cited, the list of priority conditions included arthritis, HIV/AIDS, stroke, gall bladder disease, stomach ulcers, and Alzheimer’s disease. The table uses 3-digit ICD-9 codes, rather than the MEPS “priority condition” marker, to identify families with the condition; in addition, the definitions are somewhat broader than those used in MEPS. 8 For part-year survey participants, the “year” is the portion of the year for which data are available. Thus someone who participated for nine months was fully insured if he or she had coverage for all nine months, or partially insured if he or she had coverage during only some of those months. 7 Table 4. Out-of-Pocket Expenses Among Non-Medicare Families as a Percentage of Family Income, by Insurance Coverage and Duration, 1996 Families Families with OOP Expense of: Total 5 Percent or More 10 Percent or More Families of Income of Income Family type (thousands) N % N % All members insured for entire year 50,621 4,464 8.8 1,514 2.99 Some members uninsured and/or 23,245 2,478 10.7 1,066 4.58 some period without insurance All members uninsured for entire year 8,282 1,272 15.4 666 8.04 All families 82,148 8,214 10.0 3,245 3.95 Individuals in Families Individuals in Families with OOP Expense of: Total 5 Percent or More 10 Percent or More Individuals of Income of Income Family type (thousands) N % N % All members insured for entire year 131,815 11,950 9.1 3,959 3.00 Some members uninsured and/or 69,582 7,763 11.2 3,432 4.93 some period without insurance All members uninsured for entire year 15,424 2,201 14.3 1,031 6.68 All families 216,821 21,914 10.1 8,422 3.88 Source: Author’s analysis of data from MEPS 1996. Full-year uninsured families were much more likely than full-year insured families to spend a high share of income on OOP costs. The experience of the middle, partially insured group falls midway between that of the fully insured and the fully uninsured, although the difference from either of the other two groups is not very significant.9 Do these differences persist when income and health status are factored in? Table 5 looks at lower-income families with and without health problems.10 Among low-income families with no health problem, those uninsured for the entire year were slightly less likely than those insured for the entire year to incur high relative costs. The reverse was true for families with a reported health problem. In both groups, the differences between the full-year insured and the full-year uninsured were not statistically significant. 9 At 90 percent confidence, the partially insured are more likely to spend 10 percent or more of income than the fully insured and less likely than the fully uninsured. They are less likely than the fully uninsured to spend 5 percent or more of income, but not more likely than the fully insured. 10 Results for low-income families with some uninsured members or some period of uninsurance are not shown; these results may be misleading, because these families may qualify for Medicaid for part of a year precisely because they had high OOP costs. 8 Table 5. Out-of-Pocket Expenses Among Low-Income Non-Medicare Families as a Percentage of Family Income, by Insurance Coverage and Presence of Health Problem, 1996 Families Percentage of Families with OOP Expense (thousands) Greater Than 5 Percent of Income No health problem Full-year insured 5,110 11.1% Full-year uninsured 2,756 10.2 Family has health problem Full-year insured 4,565 26.3 Full-year uninsured 2,264 32.9 Source: Author’s analysis of data from MEPS 1996. Table 6, which shows average total and OOP expenditures for low-income families with and without insurance, suggests a possible explanation. Fully insured families, with or without health problems, pay a much lower share of their own costs, but their total expenses are higher. This may be because insurance enables them to obtain more care, or because families who expect to need care are more likely to seek insurance. Both factors probably play a role. Table 6. Mean Total and OOP Expenses Among Low-Income Non-Medicare Families, by Insurance Coverage and Presence of Health Problem, 1996 OOP as a Mean Total Mean OOP Percentage Health Spending Expense of Total No health problem Full-year insured $2,057 $327 16% Full-year uninsured 414 250 60 Family has health problem Full-year insured 5,380 636 12 Full-year uninsured 1,396 541 39 Source: Author’s analysis of data from MEPS 1996. Type of Insurance A family’s exposure to high OOP costs depends not only on whether the family has health insurance, but on the type of insurance it has. Employer-sponsored insurance is generally more comprehensive than private nongroup coverage. Medicaid, the major source of non-Medicare public insurance, generally requires little or no cost-sharing for the services it covers, although Medicaid for adults may provide no coverage at all for some services, such as dental care. 9 To assess the effects of different kinds of insurance for entire families, it is necessary to focus on families in which all members were insured for the entire year and who had only one source of coverage—employer, other private, or non-Medicare public.11 Table 7 shows that the number of families all of whose members had nonemployer private coverage or public coverage for the entire year is quite small. Many fully insured families have mixed coverage sources: for example, a family with employer coverage may buy nongroup coverage for a child too old to be covered as a dependent under the group plan, or a parent may have employer coverage while a child is enrolled in Medicaid. Table 7. Non-Medicare Families Insured for Entire Year, by Source of Coverage, 1996 Coverage Source Families (thousands) Percentage of Families Employer 40,303 79.6% Other private 3,196 6.3 Public 3,163 6.2 Mixed coverage sources 3,959 7.8 Total 50,621 100.0 Source: Author’s analysis of data from MEPS 1996. Given the sample Figure 6 Families with Serious Health Problems limitations, Figure 6 shows Are Most at Risk for Costs Greater Than only the proportion of 5 Percent of Income, Even with Insurance families spending 5 percent Percentage of non-Medicare families spending more than 5% of income or more of income on 40% No health problem Health problem OOP expenses. Families 28% with nonemployer private 20% 14% 15% 13% 13% coverage are much more 5% 7% likely to have high costs 0% * than are those with Employer Other private Medicaid or All families other public employer or public Note: Includes only families all of whose members had the same insurance status throughout the year. coverage. The employer * Too few cases for reliable estimate. and “other public” groups Source: Author’s analysis of data from MEPS 1996. do not differ significantly. Nongroup policies vary enormously, and there is little hard data on what people are actually buying. Appendix A provides some very limited evidence, from the 1996 11 Some individuals had multiple sources of coverage during a single month. In this case the individual is assigned to a primary coverage source in the following sequence: employer, other private, or non-Medicare public. 10 MEPS insurance component, that many nongroup plans are inferior to employer plans in a number of key respects. Nongroup plans tend to have higher deductibles, require higher coinsurance payments, are less likely to have limits on OOP spending, and are less likely to cover prescription drugs. Table 8 shows the results: on average, families who rely on nongroup coverage throughout the year pay a much higher share of their expenses out of pocket. Table 8. Mean Total and OOP Expense, by Source of Coverage and Presence of Health Problem, Non-Medicare Families with Full-Year Insurance, 1996 OOP as a Mean Total Mean OOP Percentage Health Spending Expense of Total No health problem Employer $2,779 $ 677 24% Other private 1,859 842 45 Medicaid or other public 3,072 207 7 Family has health problem Employer 7,093 1,107 16 Other private 4,238 1,245 29 Medicaid or other public 4,296 300 7 Total 4,089 794 19 Source: Author’s analysis of data from MEPS 1996. Among families with employer coverage, the type of health plan makes some difference in the likelihood that a family will incur OOP costs greater than 5 percent of income: • Families in conventional plans are twice as likely to incur this level of costs than are those in managed care plans (10.8 versus 5.8 percent, significant at p ≤ .05); • Families with coverage from very small establishments (2–49 workers) are more likely than those with coverage from establishments with 250 or more workers (8.1 percent versus 4.6 percent, significant at p ≤ .1); • Families with nonunion coverage are more likely than those with union plans (7.0 percent versus 4.5 percent, significant at p ≤ .1). 11 Spending by Type of Service Only the group of employer-covered families provides a large enough sample to permit analysis of OOP costs by type of service. Table 9 shows, by type of service, mean total and OOP expenditures for families spending 5 percent or more of income and for other employer-covered families. Higher-spending families had higher total expenditures in every service category. The distribution of spending across service types was roughly the same as for other families. Table 9. Mean Total and Out-of-Pocket Spending for Non-Medicare Families with Full-Year Employer Coverage, as a Share of Family Income and Service Type, 1996 Share of OOP as a Share of Total Total OOP Percentage OOP Expenses Expenses Expenses of Total Expenses OOP less than 5% Ambulatory $1,554 41% $ 224 14% 36% Inpatient 1,274 33 16 1 3 Dental and vision 531 14 228 43 37 Prescription drugs 397 10 141 36 23 Total 3,812 100 619 16 100 OOP 5% or more Ambulatory 3,505 38 1,117 32 36 Inpatient 2,564 28 162 6 5 Dental and vision 1,957 21 1,265 65 41 Prescription drugs 1,103 12 458 42 15 Total 9,286 100 3,094 33 100 Source: Author’s analysis of data from MEPS 1996. Higher-spending families paid more of the total cost out of pocket for each type of service. The disparity in OOP share is proportionately larger for ambulatory and inpatient services, and smaller for prescription drugs. This may be in part because, as noted above, high-spending families are more likely to be in conventional plans that impose higher cost-sharing for ambulatory and inpatient care. Cost-sharing for prescription drugs may not vary as much by type of plan. Persistence of High Out-of-Pocket Spending Most respondents in the 1996 MEPS household component participated in further rounds of interviews in 1997. Although the 1997 data have been released, the federal Agency for Healthcare Research and Quality has not released weighting factors that would allow national projections of the experience of entire families (as opposed to individuals) over the full two-year period. For this reason, Table 10 gives counts of individuals within 12 families. The table omits individuals in families any of whose members had Medicare in either year. Table 10. Individuals in Non-Medicare Families with High Out-of-Pocket Spending, 1996 and 1997 Family OOP Spending Equal to: 5 Percent or More 10 Percent or More of Family Income of Family Income Percentage of Percentage of All Individuals in All Individuals in Individuals Non-Medicare Individuals Non-Medicare (thousands) Families (thousands) Families 1996 20,854 9.9% 7,837 3.7% 1997 19,017 9.1 7,875 3.8 Either year 36,448 17.4 13,934 6.6 Both years 6,220 3.0 1,616 0.8 Source: Author’s analysis of data from MEPS 1996 and 1997. About the same number of people were in families with high OOP spending in 1996 as in 1997 (the differences are not statistically significant).12 The number of people in families with high OOP Figure 7 High Out-of-Pocket Costs Affect One of Six Individuals costs in both years is in Non-Medicare Families Over a Two-Year Period considerably smaller; for Percentage of individuals in a non-Medicare family with OOP expenses greater than 5% of income many families high OOP costs are truly a one-time 20% 17% catastrophe. On the other 15% hand, the likelihood that a 10% 9% family will experience high 10% costs is much greater over 5% the two-year period than in any one of the years. 0% (Figure 7). 1996 1997 1996 or 1997 Source: Author’s analysis of data from MEPS 1996. FACTORING IN PREMIUMS: CONSUMER EXPENDITURE SURVEY DATA In order to get some picture of combined premium and OOP spending, the following analysis uses an alternate data source, the Consumer Expenditure Survey (CES) conducted 12 The counts of individuals for 1996 are slightly different from those shown in Table 6, because this table omits families with a Medicare member in either 1996 or 1997, and because the two-year weights are different from those for 1996 alone. 13 by the Bureau of Labor Statistics. Based on quarterly interviews and expenditure diaries maintained by a rotating national sample of urban “consumer units” (or households), the CES allows estimates of spending patterns by different kinds of families. Figures here reflect spending during calendar year 1998 (collected in interviews in 1998 and the first quarter of 1999). Table 11 shows the proportion of consumer units that devoted a large share of total spending to health care, with and without premiums included. It is limited to consumer units with no members aged 65 or over. Note that the percentages shown are percentages of expenditures, not of income. Lower-income consumer units on the CES actually have mean expenditures totaling more than mean income—partly because income is underreported on the survey and partly because they buy on credit, use up savings during periods of unemployment, or finance their spending in other ways. Only in the $50,000-or-more group is mean reported spending less than mean income; that is, consumer units in this income group are saving and investing. Table 11. Percentage of Nonelderly Consumer Units with High Health Expenses, 1998 (with and without health insurance premiums) Percentage of Consumer Units with Health Expenses Totaling: 5 Percent or More of 10 Percent or More of Total Spending Total Spending Units Premiums Premiums Income (thousands) OOP Only and OOP OOP Only and OOP Less than $15,000 13,017 11.8% 24.6% 5.3% 12.1% $15,000–$29,999 13,403 12.5 30.8 4.6 13.3 $30,000–$49,999 14,958 11.9 33.1 4.1 12.4 $50,000 or more 23,314 11.3 28.8 3.6 8.7 Total 64,693 11.8 29.4 4.3 11.2 Source: 1998 Consumer Expenditure Survey. At all income levels, adding in premiums more than doubles the proportion of units that devote a high share of spending to health care.13 As Table 12 shows, there are actually two quite different populations with high health expenses. OOP costs make up 80 percent of total health spending for the population that reaches the 5-percent threshold on the basis of OOP costs alone. OOP costs make up just 24 percent of total spending for 13 With or without premiums included, the proportion of units with high spending does not generally diminish with increasing income. Of course, a higher-income family must spend a larger dollar amount on health care before this amount becomes a large share of its total spending. 14 the population that reaches the threshold only when premiums are added in. This pattern is true at all income levels. It also holds for the subset of high spenders for whom health spending is 10 percent or more of total spending. Table 12. Mean OOP and Premium Expense, 1998 (nonelderly consumer units with high health expenses) Units with Health Expenses Greater Than 5 Percent of Total Spending Counting OOP Adding Premiums Expense Alone to OOP Expense All Units Units (thousands) 7,622 11,366 18,988 Mean OOP expense $2,879 $502 $1,456 Mean premium expense $740 $1,624 $1,269 Mean total health expense $3,619 $2,126 $2,725 Premiums as percentage of total 20% 76% 47% Source: 1998 Consumer Expenditure Survey. POLICIES TO ADDRESS UNDERINSURANCE Federal and state policymakers have shown increasing interest in measures to extend health coverage to the uninsured. However, the data in this report suggest that the simple dichotomy “insured/uninsured” does not capture the nature of the varying burdens of different families, especially those with modest incomes. It is important to think about what health insurance is for, both in designing new policies for the uninsured and in evaluating how well health insurance works for people who already have it. Public policy could promote adequate protection for families in at least three basic ways. First, public programs that provide direct health insurance coverage can include limits on cost-sharing and comprehensive coverage. Second, minimum benefit standards for private health plans can be established, either through direct regulation or as a condition of current or future tax preferences for health insurance. Third, direct assistance could be provided—e.g., through the tax code—for families incurring very high costs. Public Programs With the conspicuous exception of Medicare, the major public insurance programs already include protections against high OOP costs. Federal Medicaid and State Children’s Health Insurance Plan (SCHIP) rules already limit the amount of premiums and cost-sharing that state plans may impose. Both programs may exclude or limit benefits for services not thought of as basic medical care and families in need of these services may have difficulty obtaining them (the issue of scope of services is considered at the end of this paper). Studies suggest that even the very small copayments that may be imposed for covered 15 services may deter use of necessary care.14 In general, however, the public programs provide more comprehensive protection against high OOP expenses than all but the most generous private plans. Some proposals to expand public coverage would continue this principle. For example, proposals to extend SCHIP to adults, such as S. 574 (Sen. Feinstein), would extend to all family members the current requirement that premiums and cost-sharing may not exceed 5 percent of income. On the other hand, some proposals have the potential to reach uninsured people by narrowing the scope of coverage. In August 2001, the Bush Administration announced that it would allow states to extend Medicaid or SCHIP coverage to more individuals while reducing benefits for some current participants. It granted the first such waiver to Utah in February 2002. The trade-off between expanding coverage and assuring adequacy is a difficult one. Health insurance serves two functions: it can promote access to medical care and it can protect families against catastrophic financial losses. These two functions, if not actually in conflict, are at least distinct. A private insurance plan or public program can encourage entry into the medical care system by providing first-dollar coverage for physician office visits and other specified services. Or it can emphasize financial protection by covering costs in excess of some fixed deductible or other threshold. For many years, public policy has focused on health insurance’s role in encouraging access. Beginning with the expansions of Medicaid eligibility for pregnant women and children in the 1980s, and continuing with the enactment of SCHIP in 1997, the emphasis has been on first-dollar coverage for preventive services, early diagnosis and treatment of childhood conditions, and prompt initiation of prenatal care. Some initiatives, undertaken with limited resources, have explicitly chosen to promote these goals instead of providing payment for very high-cost services. For example, some state- level precursors of SCHIP provided low-income children with coverage for preventive and other ambulatory services, while offering little or no payment for inpatient care. To some extent, Medicaid itself is structured as a “front-end” program: states may not require cost-sharing for services to pregnant women and children but may, for example, set an annual limit on the number of inpatient hospital days Medicaid will pay for. Generally, the rationale for this has been that families who meet the financial eligibility standards for Medicaid have few assets. Therefore, they are unlikely to experience real losses even in the event of a catastrophic episode; instead, hospitals will simply write off the bills. 14 See, for example, Soumerai (1987). 16 In fact, the goals of promoting access and providing financial protection are not so easily separable. Table 13 shows that families who are unprotected from high OOP costs also have greater difficulty obtaining necessary services.15 So providing substandard insurance to a broader population might reduce the count of uninsured people, but it will not necessarily achieve either of the two basic objectives of expanded health coverage. Table 13. Percentage of Non-Medicare Families with Access Problems, by Out-of-Pocket Expense as a Percentage of Family Income, 1996 OOP Expense as a Percentage of Income Less Than 5 Percent 10 Percent 5 Percent or More or More Family member: Went without needed care 6.9% 12.3% 16.6% Had difficulty obtaining a service 11.9 22.5 22.7 Difficulty for financial/insurance reason 9.9 18.6 20.5 Source: 1996 MEPS Access to Care supplement. An intermediate policy option that might promote both objectives would provide a limited amount of first-dollar protection—to encourage the initiation of care—and protection against catastrophic costs, while leaving families with some degree of exposure for costs in the middle range. For example, a family at 200 percent of poverty ($35,300 for a family of four in 2001) might have full coverage for services up to a given dollar limit— say $500. It might then be liable for 25 percent of costs between $500 and $7,500, after which full coverage would resume. The maximum potential OOP liability in the middle range would then be $1,750, or about 5 percent of family income. This is merely an illustration. The dollar amounts could certainly be varied, or the front-end coverage might be focused on certain types of services, such as primary and preventive care. The point is merely that public programs can be designed in any number of ways that may not resemble any existing insurance plan; the choice is not simply between first-dollar coverage and catastrophic protection. 15 This table is based on the MEPS Access to Care supplement, which gathers information on barriers to health care for the family. Participants are asked if any family members have recently gone without needed health care because the family needed money to buy food, clothing, or pay for housing. They are also asked whether any family members experienced difficulty in obtaining any type of health care, delayed obtaining care, or did not receive health care they thought they needed, and why. Possible reasons could include financial/insurance problems; communication, transportation, or physical problems; or other problems. Financial/insurance problems included: insurance company wouldn’t approve, cover, or pay for care; pre- existing condition; insurance required a referral, but couldn’t get one; doctor refused to accept family’s insurance plan. 17 Private Insurance The federal and state governments regulate private health plans in a variety of ways, but, generally, this regulation has not focused specifically on adequacy of financial protection. Current proposals to encourage expansion of private coverage—e.g., through tax credits for nongroup insurance or incentives for employer-sponsored coverage—often have no standards for the content of the coverage. Under the Employee Retirement Income Security Act (ERISA), states may not regulate employee benefit plans but may regulate health insurers. Over the years, states have imposed numerous “mandated benefit” laws, usually requiring that private insurance plans—both individual and group—cover specified services or particular types of providers, or offer coverage to particular groups or individuals. It does not appear that any state regulates cost-sharing requirements of private insurance policies or requires that policies specify any limit on OOP costs.16 The absence of these kinds of mandates may reflect the fact that no organized provider or patient group is pressing for them. In addition, any rule that actually dictated the dollar amounts of coverage insurers may sell to employer groups might, more than other state mandates, be seen as direct regulation of employee benefit plans, and hence preempted under ERISA. The federal government directly regulates only one kind of health insurer, HMOs that seek federal qualification.17 Current rules for federally qualified HMOs do have some limits on cost-sharing, although they are archaic and could potentially expose families to very high costs. In any event, HMO qualification is voluntary, and there is no federal certification of other types of plans. Most federal standards for health plans are in the form of rules for employer benefit plans under ERISA, along with some standards in the Internal Revenue Code for plans that receive tax preferences. Almost none of these standards relate directly to the kinds of benefits plans must provide. There are some exceptions, such as the Mental Health Parity Act of 1996, which forbids larger employer plans that include mental health benefits from applying annual or lifetime dollar limits that do not also apply to medical and surgical services. However, there are no general rules that govern cost-sharing or OOP liability. The tax deduction for health insurance premium payments by the self-employed is available for any health insurance plan; no minimum standards are established. 16 If any such mandate exists, it is not included in the Blue Cross Blue Shield Association’s annual compendium of state mandated benefit laws. For the December 2001 listing, see http://bcbshealthissues.com/state/. 17 This is to distinguish between regulation and the federal government as purchaser. When the federal government buys coverage, for example for federal employees or for Medicare beneficiaries in Medicare+Choice plans, it does specify some benefit requirements. 18 It is possible that some federal standards for private insurance plans would be considered as part of current proposals to provide tax credits for the purchase of health insurance. President Bush’s fiscal year 2003 budget proposal would provide a refundable health insurance tax credit of up to $1,000 for individuals and $3,000 for families. The credit could be used only for plans that met “minimum coverage standards, including coverage for high medical expenses.” (U.S. Treasury 2002.) However, the proposed credit amounts are quite modest; some analysts have questioned whether they would be enough to pay for a basic nongroup policy in the current market. The tendency of federal policy in recent years has been in the direction of maximizing consumer choice, rather than standardization. The prevailing view is that buyers should be able to decide whether they want catastrophic coverage, expanded coverage for specific services such as dental care or alternative therapies, or some other mix of benefits. There is a possible counter-argument: Because tax expenditures for health insurance are nearing $80 billion a year (see below) and further expenditures are contemplated to reach the uninsured, the government has some interest in assuring that the products being purchased are designed to achieve the policy goals of the expenditure. Currently, however, the policy goal is basically to assure that more people have something called “health insurance.” Tax Treatment of Medical Expenses As an alternative to federal regulation of insurance, the government itself could provide some form of protection for families with catastrophic expenses. The Internal Revenue Code already does this to some extent through the medical expense deduction. Taxpayers who itemize deductions may deduct most kinds of medical expenses, including directly paid health insurance premiums, but only to the extent that these expenses exceed 7.5 percent of adjusted gross income (AGI). In fiscal years 2001–2005, this deduction is projected to be the second largest tax expenditure for health care, totaling $28.3 billion.18 A minority of families whose medical expenses exceed the 7.5 percent of AGI threshold actually take the deduction. Table 14 compares MEPS data on the percentage of families with OOP costs that exceed 7.5 percent of gross income with IRS data on families taking the deduction by adjusted gross income. Besides the difference in income measure, the MEPS and IRS figures are not exactly comparable in other ways. MEPS families do not necessarily coincide with tax-filing units; in addition, the MEPS data omit insurance premiums and payments for nursing home care, both of which are deductible. 18 The exclusion of employer contributions from employees’ taxable income will cost $392 billion over this period. Joint Committee on Taxation (2001). 19 Still, it is clear that the low-income families who are most likely to incur catastrophic costs are least likely to take the deduction. There are several reasons. First, the deduction is available only to taxpayers who itemize their returns. Low- income families are less likely to have other deductions (e.g., mortgage interest and state income and property tax deductions) of a value sufficient to make itemization worthwhile. Suppose a family making $20,000 and another making $50,000 each incur medical expenses of $5,000. The lower-income family’s expenses exceed the 7.5 percent threshold by $3,500, compared with $1,250 for the higher-income family. But the 1996 standard deduction for a joint return was $6,700. The lower-income family would need $3,200 in other deductions before it would benefit from the medical expense deduction. Second, many low-income families already have zero or negative income tax liability, because of exemptions and the earned income tax credit (EITC). The deduction cannot reduce their liability, nor can it increase the amount of the EITC. Table 14. Families with Medical Expenses Higher Than 7.5 Percent of Income vs. Tax Returns with Medical Expense Deduction, 1996 Percentage with OOP Expense Percentage of Returns Greater Than 7.5 Percent of with Medical Expense Family Income Income, 1996 (MEPS) Deduction, 1996 Less than $15,000 26.30% 2.20% $15,000–$29,999 12.10 5.70 $30,000–$59,999 5.50 6.80 $60,000 or more 1.20 4.00 Total 10.60 4.50 Note: MEPS analysis uses gross family income; IRS data uses adjusted gross income (AGI). Sources: Author’s analysis of data from MEPS 1996; and Internal Revenue Service, Statistics of Income— 1996, Individual Income Tax Returns, Publication 1304 (Rev. 3/99). Because they are in a lower tax bracket, even low-income families who do take the deduction benefit less than higher-income families. The result is that the deduction overwhelmingly benefits the wealthiest taxpayers (Table 15). For the year 2000, 54 percent of the total tax expenditure is projected to benefit taxpayers with incomes that exceed $75,000. While taxpayers with more than $200,000 in income account for less than one percent of those who claim the deduction, they will receive more than 10 percent of the benefits. The 100,000 families with incomes less than $10,000 who claim the deduction will receive an average of $20 each. 20 Table 15. Distribution by Income Class of All Returns, 2000 (returns with medical expense deduction and tax expenditure) Expenditure Returns with Tax Percentage per Return Income Class Returns Deduction Expenditure of Total with (thousands) (thousands) (thousands) (thousands) Expenditure Deduction Less than $10 19,818 100 $ 2,000 0.00% $ 20 $10–$20 23,803 155 52,000 1.10 335 $20–$30 19,493 412 138,000 2.90 335 $30–$40 16,210 767 372,000 7.70 485 $40–$50 13,054 860 429,000 8.90 499 $50–$75 21,557 1,546 1,219,000 25.30 788 $75–$100 11,924 696 856,000 17.70 1,230 $100–$200 11,253 423 1,243,000 25.80 2,939 $200 or more 3,101 47 515,000 10.70 10,957 140,213 4,916 4,825,000 100.00 981 Note: Expenditure estimates exclude costs for long-term care and long-term care insurance. Source: U.S. Congress, Joint Committee on Taxation. There are a number of ways in which this tax expenditure could be better targeted toward needier families. The deduction could be an “above the line” deduction (as is the health insurance deduction for the self-employed) available to families that do not itemize. It could be replaced with a tax credit. A deduction, which reduces taxable income, is more valuable to taxpayers in higher tax brackets; a credit directly reduces tax liability. Or there could be a refundable credit for excess medical expenses. A refundable credit (like the EITC) may exceed tax liability; the excess is refundable to the taxpayer. This option would provide the greatest assistance to people with little or no tax liability. It is possible to provide much better targeted assistance for low-income families at approximately the same cost. Table 16 shows the possible effects of a refundable tax credit that would cover 80 percent of costs above 7.5 percent of income for the lowest-income families. The percentage of costs that the credit would cover would decrease as income rose, and the percentage-of-income threshold would increase. This dual phase-down would produce a zero available credit for single people with incomes greater than $50,000 and for families of two or more people with incomes above $75,000. The table, based on MEPS data, should not be taken as an actual simulation of tax policy. While the MEPS includes variables intended to allow estimation of family AGI and tax liability, these are unedited and of questionable reliability. This table is therefore based simply on reported gross income and OOP expenses in 1996. 21 Table 16. Hypothetical Distribution of an Income-Based Refundable Credit for Excess Health Spending, 1996 Families Receiving Credit Credit Percentage of Family Income (thousands) (millions) Total Credit Less than $15,000 4,592 $3,937 82% $15,000–$29,999 692 702 15 $30,000 or more 123 187 4 Total 5,407 4,826 100 Source: Author’s analysis of data from MEPS 1996. The cost of the credit in 1996 would have been $4.8 billion, the same as the estimated cost (excluding long-term care) of the medical expense deduction in 2000. But a much larger proportion of the funds would have flowed to the families in greatest need. Obviously the design could be improved; this is merely one illustration of how the existing funds could be re-targeted.19 One potential drawback of this form of credit is that it could in theory lead to erosion of insurance coverage. Policies with high cost-sharing and no OOP cap might conceivably be sold to low-income families (or their employers), because the federal government would be providing the catastrophic protection. However, it is improbable that such a product would be very marketable. Scope of Services Health insurance plans have always distinguished between a defined group of covered medical services and other types of care for which coverage is limited or excluded. For example, many people with private coverage have little or no coverage for dental care or routine vision care, and sharply limited coverage of mental health services. State Medicaid programs must cover all necessary services for children, but coverage for adults may or may not include dental services, vision care, and a variety of other optional scope of services benefits. SCHIP programs may limit coverage of prescription drugs, mental health, and vision and hearing services, and may exclude dental care altogether. All of the OOP spending estimates in this report include amounts paid for any form of health care. It could at least be argued that some forms of family spending are more discretionary than others and should be treated differently in the assessment of OOP 19 Stuart Butler of the Heritage Foundation has offered a more sweeping proposal that would replace all the current tax preferences for health care (the employer exclusion and the deductions for medical expenses and for health insurance for the self-employed) with a single refundable tax credit. 22 costs. For example, while many dental and vision services are essential, there are undoubtedly families who spend large amounts for cosmetic dentistry or designer eyewear. The MEPS data do not allow one to distinguish between necessary and “frivolous” expenditures, even if some boundary could be defined. However, it is possible to see what it would mean to exclude dental and vision expenses altogether from the concept of catastrophic spending. Table 17 shows the effects for families all of whose members had one source of coverage for the entire year, or all of whose members were uninsured for the entire year. Table 17. Percentage of Families with High Out-of-Pocket Expenses, 1996 (with and without expenses for dental and vision care) OOP as a Percentage of Family Income 5 Percent or More, 10 Percent or More, with Dental and Vision Care: with Dental and Vision Care: Full-Year Coverage Type Included Excluded Included Excluded Employer 7.8% 4.3% 2.5% 1.4% Other private 18.4 12.2 8.0 6.7 Public 11.1 10.0 4.5 3.8 Uninsured 15.4 10.9 8.0 5.7 Source: Author’s analysis of data from MEPS 1996. Complete exclusion of these services does significantly reduce the proportion of families spending a high share of income on OOP expenses. However, this does not mean that higher dental and vision expenses are the key difference between families that do and do not spend a high share of income. Table 9 showed that employer-covered families with higher OOP costs spend more in every major service category. Complete exclusion of dental and vision services from that table would have reduced mean OOP expense to $391 for the low-spending families and $1,829 for the high-spending families. The latter figure would still have been more than 5 percent of mean income for these families. Discussions of health insurance rarely consider services thought of as “ancillary.” It is probably fair to say, for example, that promoting universal dental insurance has not been high on the health policy agenda. This is true even though there is evidence of inadequate access for low-income families and a high incidence of serious oral health problems. A recent Surgeon General’s report (2000) points to research findings that indicate possible associations between chronic oral infections and diabetes, heart and lung disease, stroke, and low-birthweight premature births. Conversely, serious medical problems can affect 23 oral health. Thus, the distinction between basic medical care and other health services is to some extent an artificial one. Concepts of what services should be included in “basic” health insurance have shifted over time. Employer plans once commonly limited coverage of pregnancy-related services; under federal law, all but the smallest employer plans must now treat these services as they do other medical care. The Mental Health Parity Act of 1996 forbids larger employer plans that include mental health benefits from applying annual or lifetime dollar limits that do not apply to medical and surgical services. However, a plan remains free to impose limits on numbers of covered visits or inpatient days, or to exclude mental health coverage altogether. (Several proposals in the 107th Congress would move closer to full parity.) It is not inconceivable that the idea of what constitutes basic protection for families will evolve over time to include coverage of other forms of care. For the present it can at least be said that needed ancillary services represent a significant expense for some families. These costs cannot be separated lightly from the overall cost of out-of-pocket care; policy discussion should give fuller attention to problems with access to the full spectrum of health services. 24 APPENDIX A. COMPARING EMPLOYER AND NONGROUP HEALTH PLANS It is commonplace to observe that nongroup health insurance policies tend to be less comprehensive than employer group plans. However, benefits under nongroup plans vary widely, and there is little information on what kinds of plans individuals and families are actually buying. There are no surveys of the nongroup market that are comparable to the periodic employee benefit surveys conducted by the Bureau of Labor Statistics, the Henry J. Kaiser Family Foundation/Health Research and Education Trust, or private consultants such as Mercer/Foster Higgins or the Hay Group. The household component of the 1996 MEPS included a Health Insurance Plan Abstraction (HIPA) survey, which collected information on coverage features of private health insurance policies that household respondents held. Policy provisions were abstracted from plan booklets collected from households and employers. The HIPA was able to collect information for only 54 percent of the potential population (that is, households with private insurance in round 1). Accordingly, the Agency for Healthcare Research and Quality cautions that the results cannot be used to develop national estimates. The following unweighted figures reflect the provisions of the policies held by particular survey respondents and cannot be generalized to the total population of privately insured individuals. With this caveat, however, the MEPS results do conform to the general understanding that nongroup policies tend to be less comprehensive than group plans. At the same time, they highlight just how variable both nongroup and group coverage is. Table A-1 compares the one-person deductibles of policies held by MEPS respondent policyholders who had coverage through an employer plan or nongroup insurance (other than Medigap insurance).20 Nearly 28 percent of the abstracted nongroup policies had deductibles higher than $500, compared with a very small share of the employer plans. At the same time, nongroup policies were only a little less likely than employer plans to have no deductible. 20 The HIPA public use file does not identify the source of the abstracted coverage; policies were categorized using other household component data on round 1 coverage and employment. 25 Table A-1. One-Person Deductible Under Employer Coverage and Nongroup Policies for MEPS Respondents, 1996 Employer Nongroup No deductible 1,810 70 $1–$100 764 7 $101–$250 2,042 32 $251–$500 653 68 $501–$1,000 99 48 More than $1,000 30 23 Total 5,398 248 Source: 1996 MEPS Health Insurance Plan Abstraction survey. Table A-2 shows the overall coinsurance rate under the different types of policies. The coinsurance rate is defined as the percentage of total costs that are covered by the insurance plan after any deductibles have been met and before any maximums take effect. It assumes that policyholders will use services in whatever way will maximize insurance reimbursement—e.g., by staying in-network in a PPO or POS plan. Nongroup policies do not appear to be markedly different from employer plans. Coinsurance rates for both abstracted nongroup and employer policies were most commonly in the 80 to 89 percent range. Table A-2. Overall Coinsurance Rates Under Employer Coverage and Nongroup Policies, 1996 Employer Nongroup Less than 70% 66 9 70%–79% 145 10 80%–89% 2,824 165 90%–99% 1,861 32 100% 502 32 Total 5,398 248 Source: 1996 MEPS Health Insurance Plan Abstraction survey. Table A-3 shows the one-person out-of-pocket limits under the different types of policies. Employer plans were more likely to have a very low OOP limit; nongroup plans were more likely to have a very high limit or no limit at all. The most commonly reported limits were between $1,000 and $2,000 under both types of coverage. 26 Table A-3. One-Person Out-of-Pocket Limit, 1996 Employer Nongroup Coverage Coverage OOP Limit N N Less than $500 530 9 $500–$749 669 6 $750–$999 342 8 $1,000–$1,499 1,413 61 $1,500–$1,999 921 117 $2,000 or more 897 82 No limit 669 47 Total 5,441 330 Source: 1996 MEPS Health Insurance Plan Abstraction survey. Finally, nongroup policies appear to be much less likely to include prescription drug coverage, although a surprisingly large number of employer plans also lacked coverage (Table A-4). Table A-4. Prescription Drug Coverage, 1996 Employer Nongroup Policies with: N N Prescription drug coverage 5,727 205 No drug coverage 1,519 130 Total 7,246 335 Source: 1996 MEPS Health Insurance Plan Abstraction survey. 27 APPENDIX B. DETAILED TABLES ON TRENDS IN OUT-OF-POCKET SPENDING, 1987–1996 Table B-1 shows changes in the components of OOP spending between 1987 and 1996. In nominal dollars, per-capita OOP spending for the entire population was virtually unchanged, while spending per person with any expense actually dropped slightly. The most conspicuous change was in spending for inpatient care and associated physician services. Separate inpatient deductibles have become rare in private health plans, and many managed care plans now pay in full for inpatient services from network physicians. In constant 1987 dollars, spending dropped in nearly every category.21 (The apparent drop in spending for ambulatory services is not statistically significant.) The major exception was in the prescription drug category, for which per-capita spending increased by two-thirds in nominal dollars and by more than 20 percent in constant dollars. This trend is likely to have continued since 1996, as drugs have grown more costly and many plans have increased out-of-pocket contribution requirements. 21 The CPI-U for all items is used as the deflator, rather than the medical care component, because the focus here is on the change in out-of-pocket spending relative to overall family budgets. 28 Table B-1. Components of Real Annual Out-of-Pocket Expenses for Health Services, by Type of Service, 1987 and 1996 (mean per person and mean per person with expense) 1987 1996 Percentage Percentage Mean OOP Mean OOP Change, Mean Change in Persons Percentage Expense per Persons Percentage Expense per per Person, Constant (thousands) of Persons Person a (thousands) of Persons Person 1987–1996 1987 Dollars Total population 239,393 100.0% $360 268,905 100.0% $361 0.1% –27.1% With out-of-pocket expense for: Any serviceb 181,285 75.7 476 211,346 78.6 459 –3.5 –29.8 Ambulatory physician 126,251 52.7 153 144,512 53.7 147 –3.9 –30.1 Ambulatory nonphysician 39,570 16.5 173 45,193 16.8 160 –7.3 –32.6 Inpatient hospital 9,407 3.9 1,360 5,502 2.0 553 –59.3 –70.4 Inpatient hospital, physician 7,916 3.3 670 4,283 1.6 236 –64.8 –74.4 Dental care 82,251 34.4 203 80,580 30.0 276 35.8 –1.2 Prescription drugs 120,784 50.5 105 165,711 61.6 175 66.7 21.3 Vision care 33,736 14.1 119 38,141 14.2 157 31.7 –4.2 Medical equipment 11,926 5.0 174 8,801 3.3 252 44.8 5.4 a Adjusted by the Consumer Price Index, all items in 1996 dollars. b Includes spending for home health care and other services not separately listed. In 1996, these include some ambulatory services not classified as physician or nonphysician. Sources: Author’s analysis of data from MEPS 1996; and Taylor and Banthin. 29 Table B-2 is at the level of entire families and shows changes in the distribution of OOP costs as a share of family income. For the purposes of comparison, this table follows the published NMES data in categorizing families according to the characteristics of the family head.22 As the family head is not necessarily the family member incurring the largest costs, this may sometimes be misleading—for example, when the family head has insurance and other family members do not. Table B-2. Percentage of Families with Out-of-Pocket Expense Greater Than 5 Percent and Greater Than 10 Percent of Family Income, 1987 and 1996 Percentage of Families with OOP Expense of: 5 Percent or More 10 Percent or More of Income of Income Characteristics of Family Head 1987 1996 1987 1996 Total 20.2% 16.3% 10.0% 7.4% Age of family head 19 to 24 13.6 11.7 8.4 6.1 25 to 54 13.5 9.2 5.8 3.5 55 to 64 23.9 18.7 10.5 8.2 65 or older 40.2 35.0 22.9 17.6 Sex Male 17.1 14.3 7.7 5.8 Female 25.6 19.2 14.1 9.8 Ethnic/racial background White (1996 includes other) 21.2 17.3 10.5 7.8 Black 16.5 12.8 9.3 6.4 Hispanic 17.0 11.7 7.1 5.2 Family income Less than $20,000 33.1 30.2 19.1 17.2 $20,000 to $39,999 15.1 15.9 5.2 5.1 $40,000 or more 6.4 6.0 1.4 1.4 Insurance coverage of family head Under age 65 Any private insurance 13.7 9.3 5.5 3.3 Public insurance only 16.8 17.4 10.6 8.8 Uninsured 24.5 15.0 13.2 7.5 Age 65 or older Medicare only 44.0 38.3 29.0 20.1 Medicare and other public insurance only 18.9 26.1 11.5 14.8 Medicare and private insurance 42.0 36.1 23.3 17.6 Note 1: If the head of the family’s insurance coverage changed over the course of the year the family is assigned to a coverage category in the following sequence: Medicare, employer, other private, Medicaid and other public. The uninsured are those without coverage at any time during the year. Note 2: 1996 estimates in bold significantly different from 1987 estimate at the p ≤ .05 level. Sources: Author’s analysis of data from MEPS 1996; and Taylor and Banthin. 22 “Family head” is not defined in Taylor and Banthin, nor is any “head” identified in the MEPS data. For the purpose of Table B-2, the head is simply the oldest person in the MEPS-defined family unit. 30 The share of families with OOP costs in excess of 5 percent of family income, or in excess of 10 percent, dropped between 1987 and 1996. This is true for the population as a whole and for most subgroups shown in Appendix Table B-2. As the table indicates, the change was not always significant, and some groups may have seen an increase in the proportion of families with high costs. About three-fourths of the total change is attributable to growth in family incomes during the period. The share of families with incomes below $20,000 went from 42 percent in 1987 to 31 percent in 1996; 27 percent of families had incomes above $40,000 in 1987 and 41 percent in 1996. This change, in the absence of any change in the distribution of OOP costs, would have reduced the proportion of families with costs in excess of 10 percent of family income from 10.0 percent to 7.9 percent. Within each income group, changes in proportions of families with high costs were not significant. The table shows, not surprisingly, that low-income families and those whose head is aged 65 or older spend the largest share of income on medical care. It appears to show that, among families with younger heads, those relying on public insurance are even more likely than the uninsured to incur high costs. This may be in part an artifact of the way family heads are assigned to insurance categories. Someone who was uninsured for 10 months and then qualified for Medicaid after spending down income and assets would be in the “public insurance only” line. As was shown in Table 1, families with public insurance for the entire year are about as well protected as those with employer coverage. 31 REFERENCES Butler, S., Expanding Health Insurance Through Tax Reform (Washington, D.C.: Henry J. Kaiser Family Foundation, October 1999). Medicare Payment Advisory Commission (MedPAC), Report to the Congress: Medicare Payment Policy (Washington, D.C.: MedPAC, March 2000). Soumerai, S. B. et al., “Payment Restrictions for Prescription Drugs Under Medicaid: Effects on Therapy, Cost, and Equity,” New England Journal of Medicine 317 (Aug. 27, 1987): 550–56. Stuart, B., D. Shea, and B. Briesacher, Prescription Drug Costs for Medicare Beneficiaries: Coverage and Health Status Matter (New York: The Commonwealth Fund, Issue Brief, January 2000). Taylor, A., and J. Banthin, Changes in Out-of-Pocket Expenditures for Personal Health Services: 1977 and 1987, National Medical Expenditure Survey Findings 21 (Rockville, Md.: Agency for Health Care Policy and Research, U.S. Public Health Service, 1994). U.S. Congress, Joint Committee on Taxation, Estimates of Federal Tax Expenditures for Fiscal Years 2001–2005 (Washington, D.C.: U.S. Government Printing Office, April 6, 2001, JCS-1-01). U.S. Department of Health and Human Services, “HHS to Give States New Options for Expanding Health Coverage,” press release (Washington, D.C.: August 4, 2001). U.S. Department of the Treasury, General Explanations of the Administration’s Fiscal Year 2003 Revenue Proposals (Washington, D.C.: February 2002). U.S. Surgeon General, Surgeon General’s Report on Oral Health (Washington, D.C.: May 2000). 32 RELATED PUBLICATIONS In the list below, items that begin with a publication number are available from The Commonwealth Fund by calling our toll-free publications line at 1-888-777-2744 and ordering by number. These items can also be found on the Fund’s website at www.cmwf.org. Other items are available from the authors and/or publishers. #527 Are Tax Credits Alone the Solution to Affordable Health Insurance? Comparing Individual and Group Insurance Costs in 17 U.S. Markets (May 2002). Jon R. Gabel, Kelley Dhont, and Jeremy Pickreign, Health Research and Educational Trust. This report identifies solutions that might make tax credits and the individual insurance market work, including raising the amount of the tax credits; adjusting the credit according to age, sex, and health status; and combining tax credits with new access to health coverage through existing public or private group insurance programs. #518 Bare-Bones Health Plans: Are They Worth the Money? (May 2002). Sherry Glied, Cathi Callahan, James Mays, and Jennifer N. Edwards. This issue brief finds that a less-expensive health insurance product would leave low-income adults at risk for high out-of-pocket costs that could exceed their annual income. #540 Individual Insurance: How Much Financial Protection Does It Provide? (April 17, 2002). Jon R. Gabel, Kelley Dhont, Heidi Whitmore, and Jeremy Pickreign, Health Research and Educational Trust. Health Affairs web exclusive. This article demonstrates that a $1,000 tax credit would be more than adequate to buy individual coverage for healthy, young, single males, but it would not even come close for their middle-aged peers. Available online only at http://www.healthaffairs.org/ WebExclusives/Gabel_Web_Excl_041702.htm. #533 Medicare+Choice: Beneficiaries Will Face Higher Cost-Sharing in 2002 (March 2002). Lori Achman and Marsha Gold, Mathematica Policy Research, Inc. In this report (available on the Fund’s website only), the authors note that while increases in monthly premiums will affect all enrollees in 2002, sicker beneficiaries will bear the brunt of changes in the structure of prescription drug benefits and cost-sharing requirements as more plans restrict drug coverage to generics only and raise cost-sharing requirements for services such as inpatient and outpatient hospital care. #494 Out-of-Pocket Health Care Expenses for Medicare HMO Beneficiaries: Estimates by Health Status, 1999–2001 (February 2002). Lori Achman and Marsha Gold, Mathematica Policy Research, Inc. Analysis by the authors of Medicare Compare found that out-of-pocket spending for Medicare+ Choice enrollees can be substantial and varies significantly with health status. In 2001, the average enrollee in good health spent $1,195 annually out-of-pocket on health care, while an enrollee in poor health spent $3,578, or about three times as much. #512 Security Matters: How Instability in Health Insurance Puts U.S. Workers at Risk (December 2001). Lisa Duchon, Cathy Schoen, Michelle M. Doty, Karen Davis, Erin Strumpf, and Stephanie Bruegman. This report, based on The Commonwealth Fund 2001 Health Insurance Survey, finds that in the past year one of four Americans ages 19 to 64, some 38 million adults, was uninsured for all or part of the time. Lapses in coverage often restrict people's access to medical care, cause problems in paying medical bills, and even make it difficult to afford basic living costs such as food and rent. 33 #514 Experiences of Working-Age Adults in the Individual Insurance Market (December 2001). Lisa Duchon and Cathy Schoen. This issue brief, based on The Commonwealth Fund 2001 Health Insurance Survey, describes the difficulties faced by those without access to group health coverage in obtaining adequate, affordable individual health insurance. #511 How the Slowing U.S. Economy Threatens Employer-Based Health Insurance (November 2001). Jeanne M. Lambrew, George Washington University. This report documents the link between loss of health insurance and unemployment, estimating that 37 percent of unemployed people are uninsured—nearly three times as high as the uninsured rate for all Americans (14%). The jobless uninsured are at great financial risk should they become ill or injured. #474 One-Third at Risk: The Special Circumstances of Medicare Beneficiaries with Health Problems (September 2001). Marilyn Moon and Matthew Storeygard, The Urban Institute. The authors argue that any major change to the Medicare program—such as requiring coinsurance for home health care—must take into account the steep costs seriously ill beneficiaries already pay for health services. #502 Gaps in Health Coverage Among Working-Age Americans and the Consequences (August 2001). Catherine Hoffman, Cathy Schoen, Diane Rowland, and Karen Davis. Journal of Health Care for the Poor and Underserved, vol. 12, no. 3. In this article, the authors examine health coverage and access to care among working-age adults using the Kaiser/Commonwealth 1997 National Survey of Health Insurance, and report that having even a temporary gap in health coverage made a significant difference in access to care for working-age adults. #472 Insuring the Uninsurable: An Overview of State High-Risk Health Insurance Pools (August 2001). Lori Achman and Deborah Chollet, Mathematica Policy Research, Inc. The authors argue that high premiums, deductibles, and copayments make high-risk pools unaffordable for people with serious medical conditions, and suggest that by lifting the tax exemption granted to self-insured plans, states could provide their high-risk pools with some much-needed financing. #468 Market Failure? Individual Insurance Markets for Older Americans (July/August 2001). Elisabeth Simantov, Cathy Schoen, and Stephanie Bruegman. Health Affairs, vol. 20, no. 4. This new study shows that adults ages 50 to 64 who buy individual coverage are likely to pay much more out-of- pocket for a limited package of benefits than their counterparts who are covered via their employers. #464 Health Insurance: A Family Affair—A National Profile and State-by-State Analysis of Uninsured Parents and Their Children (May 2001). Jeanne M. Lambrew, George Washington University. This report suggests that expanding Medicaid and State Children’s Health Insurance Program (CHIP) coverage to parents as well as children may not only decrease the number of uninsured Americans but may be the best way to cover more uninsured children. #463 Strengthening Medicare: Modernizing Beneficiary Cost-Sharing (May 2001). Karen Davis. In invited testimony before a House Ways and Means Health Subcommittee hearing, the Fund’s president cautioned that any effort to reform Medicare’s benefit package must take into account the circumstances of all beneficiaries, including those who are older, low-income, and chronically ill. #461 Reforming Medicare’s Benefit Package: Impact on Beneficiary Expenditures (May 2001). Stephanie Maxwell, Marilyn Moon, and Matthew Storeygard, The Urban Institute. This report presents four possible options for modernizing Medicare that would reverse spiraling costs for beneficiaries and reduce or eliminate the need for private supplemental insurance. 34 #462 Expanding Public Programs to Cover the Sick and Poor Uninsured (March 2001). Karen Davis. In invited testimony before the Senate Finance Committee, the Fund’s president presented a compelling case for expanding existing public health insurance programs to provide coverage for the most vulnerable segments of the nation’s 42.6 million uninsured. She stressed the importance of expanding Medicaid and the Children’s Health Insurance Program (CHIP) to cover parents of covered children. #415 Challenges and Options for Increasing the Number of Americans with Health Insurance (January 2001). Sherry A. Glied, Joseph A. Mailman School of Public Health, Columbia University. This overview paper summarizes the 10 option papers written as part of the series Strategies to Expand Health Insurance for Working Americans. #442 Incremental Coverage Expansion Options: Detailed Table Summaries to Accompany Option Papers Commissioned by The Commonwealth Fund Task Force on the Future of Health Insurance (January 2001). Sherry A. Glied and Danielle H. Ferry, Joseph L. Mailman School of Public Health, Columbia University. This paper, a companion to publication #415, presents a detailed side-by-side look at the 10 option papers in the series Strategies to Expand Health Insurance for Working Americans. #430 Growth in Medicare and Out-of-Pocket Spending: Impact on Vulnerable Beneficiaries (January 2001). Stephanie Maxwell, Marilyn Moon, and Misha Segal, The Urban Institute. Medicare beneficiaries will have to pay substantially more out of their own pockets for health care in the future, according to this new report. The authors find that those with low incomes and health problems will be at even greater risk than average beneficiaries for costs such as Medicare premiums, medical services, and prescription drugs. #421 Markets for Individual Health Insurance: Can We Make Them Work with Incentives to Purchase Insurance? (December 2000). Katherine Swartz, Harvard School of Public Health. Efforts to improve the functioning of individual insurance markets require policy makers to trade off access for the highest-risk groups against keeping access for the lowest risk-groups. This paper, part of the series Strategies to Expand Health Insurance for Working Americans, discusses how individual insurance markets might best be designed in view of this trade-off. Available online only at www.cmwf.org. #393 What Do Medicare HMO Enrollees Spend Out-of-Pocket? (August 2000). Jessica Kasten, Marilyn Moon, and Misha Segal, The Urban Institute. Medicare+Choice plans are scaling back benefits and shifting costs to enrollees through increases in service copayments and decreases in the value of prescription drug benefits. This report examines the financial effects of these actions on Medicare managed care enrollees. Inadequate Health Insurance: Costs and Consequences (August 11, 2000). Karen Donelan, Catherine M. DesRoches, and Cathy Schoen. Medscape General Medicine. Available online at www.medscape.com/Medscape/GeneralMedicine/journal/public/mgm.journal.html. #405 Counting on Medicare: Perspectives and Concerns of Americans Ages 50 to 70 (July 2000). Cathy Schoen, Elisabeth Simantov, Lisa Duchon, and Karen Davis. This summary report, based on The Commonwealth Fund 1999 Health Care Survey of Adults Ages 50 to 70, reveals that those nearing the age of Medicare eligibility and those who recently enrolled in the program place high value on Medicare. At the same time, many people in this age group are struggling to pay for prescription drugs, which Medicare doesn’t cover. #429 Role of Insurance in Promoting Access to Care—Uninsured and Unstably Insured: The Importance of Continuous Coverage (April 2000). Cathy Schoen and Catherine M. DesRoches. HSR: Health 35 Services Research, vol. 35, part II. Using data from three different survey databases, the authors report that, compared with those continuously insured, those insured but with a recent time uninsured are two to three times as likely to report access problems. #361 Listening to Workers: Challenges for Employer-Sponsored Coverage in the 21st Century (January 2000). Lisa Duchon, Cathy Schoen, Elisabeth Simantov, Karen Davis, and Christina An. Based on The Commonwealth Fund 1999 National Survey of Workers’ Health Insurance, this short report shows that although most working Americans with employer-sponsored health insurance are satisfied with their plans, too many middle- and low-income workers cannot afford health coverage or are not offered it. #347 Can’t Afford to Get Sick: A Reality for Millions of Working Americans (September 1999). John Budetti, Lisa Duchon, Cathy Schoen, and Janet Shikles. This report from The Commonwealth Fund 1999 National Survey of Workers’ Health Insurance finds that millions of working Americans are struggling to get the health care they need because they lack insurance or experience gaps in coverage. #262 Working Families at Risk: Coverage, Access, Costs, and Worries—The Kaiser/Commonwealth 1997 National Survey of Health Insurance (April 1998). This survey of more than 4,000 adults age 18 and older, conducted by Louis Harris and Associates, Inc., found that affordability was the most frequent reason given for not having health insurance, and that lack of insurance undermined access to health care and exposed families to financial burdens. 36