AUTOMATICALLY ENROLLING ELIGIBLE CHILDREN AND FAMILIES INTO MEDICAID AND SCHIP: OPPORTUNITIES, OBSTACLES, AND OPTIONS FOR FEDERAL POLICYMAKERS Stan Dorn and Genevieve M. Kenney June 2006 ABSTRACT: Sixty-two percent of uninsured children and two-thirds of uninsured, poor parents qualify for publicly funded health coverage programs but are not enrolled. This study assesses the potential impact of automatically enrolling children and parents in Medicaid and the State Children’s Health Insurance Program (SCHIP) based on determinations of other means-tested programs. Current law permits states to cover some uninsured parents based on information in their children’s Medicaid case files. However, current federal law forbids states from providing Medicaid or SCHIP based on the final income determinations of non-health agencies—the type of auto-enrollment that could reach eligible children. For such auto-enrollment to succeed, federal policymakers need to provide states with additional flexibility in determining eligibility and new resources for investing in information technology. Support for this research was provided by The Commonwealth Fund. The views presented here are those of the authors and not necessarily those of The Commonwealth Fund or its directors, officers, or staff. This report and other Fund publications are available online at www.cmwf.org. To learn more about new publications when they become available, visit the Fund’s Web site and register to receive e-mail alerts. Commonwealth Fund pub. no. 931. CONTENTS List of Figures and Tables................................................................................................ iv About the Authors........................................................................................................... v About the Economic and Social Research Institute .......................................................... v About the Urban Institute................................................................................................ v Acknowledgments .......................................................................................................... vi Executive Summary....................................................................................................... vii Introduction .................................................................................................................... 1 Background Information About the Uninsured ................................................................ 2 Auto-Enrollment in Other Contexts................................................................................ 5 Findings from the 2002 National Survey of America’s Families ........................................ 8 Obstacles to Auto-Enrollment........................................................................................ 12 State and Federal Policy Options.................................................................................... 20 Conclusion .................................................................................................................... 30 Notes............................................................................................................................. 31 Appendix A. Methodology ............................................................................................ 38 Appendix B. Additional Tables ...................................................................................... 40 iii LIST OF FIGURES AND TABLES Figure 1 Distribution of Uninsured by Relationship to Children and Income, 2004........................................................................... 2 Figure 2 Distribution of Uninsured Children, by Eligibility for Medicaid and SCHIP, 2002........................................................................... 3 Figure 3 Distribution of Uninsured, Low-Income Adults, by Income, Relationship to Children, and Eligibility for Medicaid, 2002 ......................... 4 Figure 4 Percentage of Low-Income, Uninsured Children Whose Families Participated in Means-Tested Nutrition Programs, 2002................................ 9 Figure 5 Health Insurance Distribution Among Low-Income Children Whose Families Participated in Means-Tested Nutrition Programs, 2002 ................ 10 Figure 6 Percentage of Uninsured, Poor Parents Whose Families Participated in Means-Tested Nutrition Programs, 2002................................................. 11 Figure 7 Health Coverage Among Poor Parents Whose Families Participated in Means-Tested Nutrition Programs or Whose Children Received Medicaid, 2002 ........................................................................................... 11 Figure 8 Typical Income Eligibility Limits for Children, Various Means-Tested Programs.................................................................. 19 Appendix Percentage of Uninsured Children Whose Families Participated Table 1 in Means-Tested Nutrition Programs, by Income, 2002 .............................. 40 Appendix Percentage of Uninsured Children with Incomes Below Table 2 200 Percent of FPL Whose Families Participated in Means-Tested Nutrition Programs, by Citizenship Status, 2002.......................................... 40 Appendix Percentage of Uninsured, Nonelderly Parents Whose Families Table 3 Participated in Means-Tested Nutrition Programs, by Income, 2002 ........... 41 Appendix Health Insurance Coverage Distribution of Children, by Income Table 4 and Family Participation in Nutrition Programs, 2002 ................................. 41 Appendix Health Insurance Coverage Distribution of Children with Table 5 Incomes Below 200 Percent of FPL, by Citizenship Status and Family Participation in Nutrition Programs, 2002 ................................. 42 Appendix Health Insurance Coverage Distribution of Nonelderly Parents, Table 6 by Income and Family Participation in Nutrition Programs, 2002................ 42 iv ABOUT THE AUTHORS Stan Dorn, J.D., is a senior policy analyst at the Economic and Social Research Institute (ESRI). He has been involved in health policy at the state and national levels for more than 20 years, focusing on low-income consumers, Medicaid, the State Children’s Health Insurance Program (SCHIP), and the uninsured. Previously, Dorn served as director of the Health Consumer Alliance, a consortium of legal services groups in California that help low-income consumers obtain necessary health care. He also directed the Health Division of the Children’s Defense Fund (CDF), where he led the health policy team in a campaign that helped enact SCHIP in 1997. Before his work at CDF, Dorn directed the Washington, D.C., office of the National Health Law Program and served as a staff attorney in its Los Angeles headquarters. He is a graduate of Harvard College and the Boalt Hall School of Law at the University of California, Berkeley. He can be reached at sdorn@esresearch.org. Genevieve M. Kenney, Ph.D., is a principal research associate and health economist at The Urban Institute, with over 20 years of experience conducting research. She is one of the nation’s leading experts on SCHIP. She has examined a range of issues related to SCHIP, including: family coverage policies and the structure of program financing; participation and barriers to enrollment; access and use differentials among low-income children; the effects of premium increases on enrollment; and the impacts of SCHIP on insurance coverage, crowd-out, and access to care. She also has conducted research on a number of Medicaid and Medicare topics, including the impacts of Medicaid eligibility expansions for pregnant women and children, the adoption of managed care in Medicaid, and the use of home health services among the dual eligible population. ABOUT THE ECONOMIC AND SOCIAL RESEARCH INSTITUTE Founded in 1987, ESRI is a nonpartisan, nonprofit research organization headquartered in Washington, D.C. Specializing in health and social policy research, ESRI conducts studies aimed at enhancing the effectiveness of social programs, improving the ways in which health care services are organized and delivered, and making high-quality health care accessible and affordable. ABOUT THE URBAN INSTITUTE The Urban Institute is a nonprofit, nonpartisan policy research and educational organization established in Washington, D.C., in 1968. Its staff members investigate the social, economic, and governance problems confronting the nation and evaluate the public and private means to alleviate them. The Institute disseminates its research findings through publications, its Web site, the media, seminars, and forums. v ACKNOWLEDGMENTS The authors gratefully acknowledge the financial support of The Commonwealth Fund. They also thank the following individuals for their thoughtful review of prior drafts of this report: Dawn Horner and Beth Morrow of The Children’s Partnership; Lynn Etheredge of the Health Insurance Reform Project, George Washington University; Zoë Neuberger of the Center on Budget and Policy Priorities, who analyzed the portions of the report that concern the National School Lunch Program; and Dr. Jack A Meyer, president of the Economic and Social Research Institute (ESRI). The authors also wish to thank Joshua McFeeters and Justin Yee of the Urban Institute for research assistance, useful discussions, and suggestions related to the analyses described in Appendix A. Neither The Commonwealth Fund, the experts who reviewed this work, the Urban Institute, nor ESRI are responsible in any way for the views expressed in this report, which are solely those of the authors. vi EXECUTIVE SUMMARY Sixty-two percent of uninsured children qualify for but are not enrolled in Medicaid or the State Children’s Health Insurance Program (SCHIP). Similarly, two- thirds of uninsured, poor parents qualify for Medicaid but are not enrolled. Auto enrollment could reach many of these uninsured families. Under this strategy, eligible parents and children receive health coverage based on information already in the hands of state officials, rather than through full, formal applications for Medicaid or SCHIP. Similar auto-enrollment strategies have achieved remarkable success with other public and private benefit programs. For example: • With retirement savings programs, 10 percent of eligible workers enroll if they must establish Individual Retirement Accounts on their own, and about one-third participate when their employers offer the option to enroll into 401(k) retirement accounts. By comparison, 90 percent of eligible workers participate in 401(k) accounts when they are automatically enrolled by their employers unless the workers actively decline participation. • Medicare Part B, into which seniors are enrolled automatically unless they decline participation, covers 96 percent of eligible seniors. By contrast, the Medicare Savings Programs, for which low-income seniors must apply to receive assistance with cost-sharing and premiums, reach no more than 33 percent of the eligible population. • The drug discount card that was the first benefit to be implemented under the Medicare Modernization Act of 2003 (MMA) included a $600 subsidy for low- income seniors. In states where seniors automatically received this benefit if they had participated in previous subsidy programs, 80 to 90 percent of eligible seniors enrolled. In states where seniors were required to apply for the new benefit, only 4 to 20 percent of eligible seniors enrolled. • Since 1991, school districts participating in the National School Lunch Program (NSLP) have had the option to provide free meals based on direct certification, through which children receive NSLP based solely on the findings of Food Stamp and cash assistance programs. In the 61 percent of districts that used this option in 2001–02, 43 percent of all children approved for free meals were enrolled either through direct certification, without any application for NSLP by their parents, or through other mechanisms that grant free meals based on income determinations of other programs. Direct certification lowered administrative costs, reduced the vii proportion of ineligible children receiving free meals, and increased the number of eligible children receiving free meals. Based on this success, Congress passed legislation in 2004 requiring all NSLP-participating districts to use direct certification by 2008–09. Comparable auto-enrollment strategies could help to enroll many uninsured children and parents into Medicaid and SCHIP. According to data from the 2002 National Survey of America’s Families, more than two-thirds (71%) of uninsured children with family incomes at or below 200 percent of the federal poverty level (FPL) live in families who participate in NSLP, the Special Supplemental Program for Women, Infants, and Children (WIC), or the Food Stamp Program. Since most states extend Medicaid and SCHIP to children with family incomes at or below 200 percent of the FPL, providing health coverage to uninsured children based on their families’ participation in these nutrition programs could reach most low-income children who qualify for Medicaid or SCHIP but are not yet enrolled. Similar strategies may also be effective in enrolling poor parents—that is, parents with incomes at or below the FPL. Among poor parents who are uninsured, 83 percent either live in a family participating in a means-tested nutrition program or have a child who receives Medicaid. Strikingly, 53 percent of poor, uninsured parents fall into the latter category, with children already enrolled in Medicaid. This suggests that one straightforward and promising way for states to identify and enroll potentially eligible but uninsured parents is through their children’s Medicaid records. A number of states have pursued auto-enrollment strategies to provide children with Medicaid or SCHIP based on the findings of other means-tested programs, such as NSLP or the Food Stamp Program. Such strategies have generally failed for two main reasons, each of which could be addressed through modest changes in federal law. First, the computer systems that states use to administer health and non-health programs often cannot communicate with one another. Accordingly, information frequently must be obtained, conveyed, or evaluated by hand in order for a child receiving non-health benefits to be enrolled into Medicaid or SCHIP. This makes ongoing program administration cumbersome, costly, and ultimately unsustainable. This problem could be solved with federal funding for states to develop the information technology (IT) needed to implement auto-enrollment through electronic exchange and analysis of eligibility information. viii Second, federal law forbids child health programs from relying on the final income determinations of other means-tested programs. This is because Medicaid, SCHIP, and non-health programs have slightly different methodologies for determining household income. For example, these programs may have varying definitions of the household members whose resources and needs are taken into account in determining eligibility, or different “disregards” that are subtracted from a household’s gross income to arrive at a net income figure. As a result, even after non-health programs have found children to have family incomes low enough to qualify for Medicaid or SCHIP, parents are nevertheless required to complete a second and generally redundant application before their children can receive health coverage. In several states pursuing auto-enrollment, such applications have been completed for less than a third (25%–31%) of potentially eligible children, leaving the remaining children uninsured. To solve this problem, policymakers could change federal law to give states the flexibility to disregard technical differences between program methodologies and grant health coverage when other means-tested programs have found that families have incomes low enough to qualify for Medicaid or SCHIP. A similar approach already applies to low- income subsidies under the MMA. Through the Medicare Savings Program (MSP), state Medicaid programs have, for many years, covered Medicare cost-sharing for low-income seniors. Under the MMA, subsidies go to certain low-income Medicare beneficiaries, with automatic coverage for seniors who participate in MSP. Although eligibility rules for MMA subsidies and MSP differ in some states, auto-enrollment still takes place, because the programs’ eligibility requirements are substantially the same, though not identical. These two changes to federal law could be accomplished through legislation. They also could be tested within a single state through a federal waiver under Section 1115 of the Social Security Act. Such a waiver could provide enhanced federal funding to develop IT infrastructure and permit the state to rely on the determinations of other means-tested programs in establishing that families meet eligibility requirements for Medicaid or SCHIP. Of course, consent to enrollment in health coverage would be needed, but states could have the flexibility to provide eligible children with health coverage unless their parents object. In this time of partisan division, one health policy goal that unites leaders in both parties is providing health coverage to the millions of uninsured children who qualify for Medicaid and SCHIP but are not enrolled. In pursuing this goal, policymakers may wish to consider giving states the resources and flexibility needed for effective implementation of auto-enrollment, which has proven successful when used with many other public and private programs. ix AUTOMATICALLY ENROLLING ELIGIBLE CHILDREN AND FAMILIES INTO MEDICAID AND SCHIP: OPPORTUNITIES, OBSTACLES, AND OPTIONS FOR FEDERAL POLICYMAKERS INTRODUCTION During most of the 1990s, the percentage of U.S. children without insurance rose steadily, from 12.7 percent in 1991 to 15.0 percent in 1997.1 Medicaid programs in most of the country covered poor children; they provided no help, however, to millions of children in near-poor, working families that earned too much to qualify for Medicaid but too little to afford health coverage on their own. To address this growing problem, federal lawmakers enacted the State Children’s Health Insurance Program (SCHIP) in 1997. States extended subsidies to millions of near-poor, uninsured children; application procedures for child health coverage were streamlined considerably; and the percentage of uninsured children began a steady decline, reaching 11.2 percent in 2004. Despite this substantial progress, almost a decade later three of five uninsured children qualify for Medicaid or SCHIP but are not enrolled. The situation has remained the same for several years. Auto-enrollment holds great promise for breaking this log jam and greatly reducing the number of uninsured children. Under this strategy, children are enrolled in Medicaid and SCHIP when information already in the hands of state agencies shows they meet eligibility standards for health coverage. While parents need to provide their consent in some appropriate fashion, no formal Medicaid or SCHIP application is required. Building on the remarkable success of auto-enrollment strategies in other programs, many children’s advocates and state policymakers have made efforts to place children into Medicaid and SCHIP based on their families’ receipt of other means-tested benefits, such as assistance provided under the National School Lunch Program (NSLP), the Special Supplemental Program for Women, Infants, and Children (WIC), or the Food Stamp Program. Unfortunately, such approaches have reached relatively few children to date.2 To help chart a way forward, this report addresses three questions: • Does auto-enrollment have the potential to help Medicaid and SCHIP reach a large proportion of eligible but previously uninsured children and families? • If so, what obstacles have prevented this strategy from working effectively in the past? • What options are available to policymakers who wish to make auto-enrollment significantly more effective in the future? 1 In addressing those questions, we first provide some background information about various groups of uninsured, including those who qualify for Medicaid and SCHIP but are not enrolled. Second, we describe past results from the use of auto-enrollment with other benefit programs. Third, we outline new findings from the 2002 National Survey of America’s Families, including the fact that a remarkably high proportion of uninsured, low-income children and uninsured, poor parents benefit from means-tested nutrition assistance. Fourth, we summarize prior research describing the barriers that have limited the effectiveness of automatically enrolling children into Medicaid and SCHIP based on their families’ participation in other means-tested programs. Fifth and finally, we explore whether changes to state and federal health policy could help overcome these barriers. BACKGROUND INFORMATION ABOUT THE UNINSURED Most uninsured Americans live in low-income households that have incomes at or below 200 percent of the federal poverty level (FPL) (Figure 1). (In 2006, the FPL is $16,600 a year for a family of three.) However, children, parents, and adults without dependent children each face distinct obstacles to coverage. Figure 1. Distribution of Uninsured by Relationship to Children and Income, 2004 <100% FPL 100%–199% FPL 200%+ FPL Children 4.1 2.4 2.5 Parents 3.7 3.8 3.5 Noncustodial 8.9 6.3 10.3 adults 0 10 20 30 Millions of uninsured Source: C. Hoffman, A. Carbaugh, H. Yang Moore et al., Health Insurance Coverage in America: 2004 Data Update (Washington, D.C.: Kaiser Commission on Medicaid and the Uninsured and the Urban Institute, Nov. 2005), http://www.kff.org/uninsured/upload/Health-Coverage-in-America-2004-Data-Update-Report.pdf. Calculations by ESRI, Feb. 2006. Among uninsured children, 62 percent qualify for Medicaid or SCHIP but are not enrolled (Figure 2).3 In large part, that reflects the comparatively generous income eligibility standards that apply to children. Ninety-three percent of American children live 2 in the 40 states (plus the District of Columbia) that provide Medicaid or SCHIP to all children with family incomes up to or above 200 percent of FPL.4 Figure 2. Distribution of Uninsured Children, by Eligibility for Medicaid and SCHIP, 2002 Eligible for Eligible for Medicaid SCHIP but not enrolled but not enrolled 34% 28% Ineligible for Medicaid and SCHIP 38% Source: T. M. Selden, J. L. Hudson, and J. S. Banthin, “Tracking Changes in Eligibility and Coverage Among Children, 1996–2002,” Health Affairs, Sept./Oct. 2004 24(5):39–50. See also State Health Access Data Assistance Center (SHADAC) and the Urban Institute, Going Without: America’s Uninsured Children, prepared for The Robert Wood Johnson Foundation (Minneapolis: SHADAC, Aug. 2005), http://www.rwjf.org/files/newsroom/ckfresearchreportfinal.pdf, finding that, in 2004, more than 70 percent of uninsured children may have qualified for Medicaid or SCHIP. Parents face a different situation. In the median state, working parents lose Medicaid if their incomes exceed 67 percent of the FPL.5 Conversely, 47 percent of nonelderly adults live in the 21 states that cover parents up to 100 percent of the FPL or higher.6 As a result, the proportion of uninsured parents who qualify for public coverage varies dramatically, depending on whether such parents have incomes above or below the FPL: • Two-thirds (67%) of uninsured parents with incomes below the FPL qualify for Medicaid; but • In households with incomes between 100 and 200 percent of FPL, most uninsured parents (78%) are ineligible for Medicaid and SCHIP (Figure 3). For adults who are not parents, even less help is available. Federal law flatly forbids state Medicaid programs from covering adults—no matter how poor—who are not caring for dependent children, unless such adults are pregnant, severely disabled, or elderly. States may receive a federal waiver to this limitation, but no additional federal funds are provided to cover otherwise ineligible adults.7 As a result, among noncustodial adults with incomes below the FPL, nearly four-fifths (73%) of the uninsured are ineligible for Medicaid (Figure 3). 3 Figure 3. Distribution of Uninsured, Low-Income Adults, by Income, Relationship to Children, and Eligibility for Medicaid, 2002 Percent Eligible for Medicaid Ineligible for Medicaid Parents, 67 33 0–99% FPL Noncustodial adults, 27 73 0–99% FPL Parents, 22 78 100–199% FPL Noncustodial adults, 6 94 100–199% FPL 0 25 50 75 100 Source: A. Davidoff, A. Yemane, and E. Adams, Health Coverage for Low-Income Adults: Eligibility and Enrollment in Medicaid and State Programs, 2002 (Washington, D.C.: Urban Institute & University of Maryland for Kaiser Commission on Medicaid and the Uninsured, Feb. 2005), http://www.kff.org/uninsured/upload/Health-Coverage-for-Low-Income-Adults-Eligibility-and-Enrollment-in-Medicaid-and-State-Programs-2002- Policy-Brief.pdf. Calculations by ESRI, Feb. 2006. While acknowledging that millions of uninsured children and adults are ineligible for any assistance, we focus in this report on the groups of uninsured who qualify for public programs but are not enrolled—namely, children with incomes at or below 200 percent of the FPL and parents with incomes at or below the FPL. In recent years, outreach efforts and streamlined application procedures have greatly increased the proportion of eligible children enrolled in public coverage. For example, one study found that, from 1998 to 2002, take-up rates for Medicaid rose from 72.2 percent to 79.1 percent of eligible children, and such rates for SCHIP increased from 43.5 to 60.4 percent. The study’s authors interpreted these results “as evidence of the effects of improved outreach, reduced stigma, enrollment simplification, continuous coverage, and the myriad other improvements in Medicaid and SCHIP implemented since the mid-1990s.”8 Further incremental improvements along similar lines remain possible.9 However, qualitatively different steps may be needed to enroll the bulk of the remaining uninsured but eligible children. One such step would be to expand parents’ coverage; a significant body of research suggests that, when parents receive coverage, they are more likely to enroll their children.10 Another promising step that departs from past policy—and has received less sustained attention from the research community—would be to automatically enroll eligible families in Medicaid and SCHIP, without requiring formal applications for 4 coverage. This could be done by using information already accessible to state health officials. According to the authors of a meta-analysis of take-up research, “Looking broadly across many programs, it seems clear that automatic enrollment is the best way to increase take-up” of Medicaid and SCHIP by eligible children.11 This report examines whether and how automatic enrollment could extend health coverage to eligible but uninsured children and families. We begin with one important caveat, however: enrollment into health coverage does not necessarily improve access to health care. Prior auto-enrollment efforts have achieved varying amounts of increased health care utilization.12 Accordingly, even if policymakers succeed in automatically enrolling eligible but uninsured families into Medicaid and SCHIP, it may be necessary to supplement these efforts with additional measures to monitor and facilitate families’ subsequent receipt of care. AUTO-ENROLLMENT IN OTHER CONTEXTS Many private and public benefit programs use auto-enrollment, often very effectively. For example, such programs have automatically enrolled new employees into employer-based retirement savings accounts; eligible seniors into Medicare coverage of outpatient care and prescription drug discount cards; and eligible families into the National School Lunch Program, based on their receipt of food stamps or cash assistance. While the economics literature contains many additional examples, these suffice to convey the potential of auto- enrollment to markedly increase utilization of available benefits.13 Retirement Savings Accounts Retirement savings plans with identical tax benefits have very different take-up rates, depending on how much effort is required to enroll: • When such retirement plans are available as Individual Retirement Accounts into which households must enroll on their own, roughly 10 percent of eligible workers participate.14 • When employers offer their workers a retirement savings vehicle in the form of a 401(k) retirement account and workers must act affirmatively to participate, approximately a third of new employees enroll. Even after a year on the job, only half of workers are enrolled. • At other firms, unless workers actively opt out, they are automatically enrolled into 401(k) accounts, funded by regular payroll deductions. In these companies, roughly 90 percent of new employees enroll.15 5 Several studies show that auto-enrollment into retirement plans has the greatest impact on low-income workers.16 Medicare Part B Auto-enrollment’s potential for success is further illustrated by Medicare Part B, which covers most Medicare outpatient care. Unless seniors affirmatively opt out of coverage within a specified period after turning 65, they are enrolled in this program, with premium payments automatically deducted from their monthly Social Security checks. Fully 95.5 percent of eligible seniors enroll in Medicare Part B.17 By contrast, two Medicaid programs that, for more than 15 years, have been available to pay Medicare cost-sharing and/or premiums for low-income seniors are used by only 33 percent of those eligible for the Qualified Medicare Beneficiary (QMB) program and 13 percent of those eligible for the Supplemental Low-Income Medicare Beneficiary (SLMB) program. To obtain the assistance, seniors must submit applications to their state’s Medicaid program.18 Medicare Drug Discount Cards The first prescription drug benefit to be implemented under the Medicare Modernization Act of 2003 (MMA) was a drug discount card, coupled with federal payment of $600 in annual drug costs for certain low-income seniors. This short-term program operated from mid-2004 until the 2006 commencement of full MMA benefits. Some seniors were auto- enrolled into the short-term discount card program based on their prior participation in Medicaid, state-run prescription drug programs, or Medicare managed care plans. Auto- enrollees accounted for 63 percent of all recipients of drug discount cards, including 44 percent of means-tested subsidy beneficiaries.19 According to the Medicare Payment Advisory Commission: Auto-enrollment was far more effective than voluntary enrollment. . . . States that used auto-enrollment achieved high participation rates in a short period of time—these rates ranged from 80 to 90 percent of eligible members. Conversely, the five states that encouraged members to voluntarily enroll in the discount card program experienced much lower enrollment rates, ranging from 2 to 40 percent.… Interviewees repeatedly stressed the success of auto-enrollment in reaching low-income populations.20 National School Lunch Program (NSLP) Generally, children qualify for free lunches under NSLP if their family income is at or below 130 percent of FPL. Reduced-price meals are provided if family income is between 131 and 185 percent of FPL. However, families can qualify for free meals based on their receipt of food stamps or cash assistance, without including any information about their 6 income on the application for NSLP. This basis for assistance is known as “categorical eligibility.”21 Since 1991, school districts have had the option to use a form of auto-enrollment called “direct certification.” Under direct certification, schoolchildren qualify for free meals based on information provided by state agencies administering Food Stamp and cash welfare programs, without any need for families to apply for NSLP. As of 2001–02, 61 percent of districts (which included 68 percent of all public school children) had implemented the direct certification option. In these districts, fully 25 percent of all children approved for free meals were enrolled through direct certification, without any NSLP application by their parents. An additional 18 percent were found to be categorically eligible based on their receipt of food stamps or cash assistance. Put differently, more than 40 percent of children receiving free meals through NSLP were found eligible based, in whole or in part, on the findings of other means-tested programs.22 More than half of school districts participating in direct certification use passive consent procedures. With passive consent, parents are notified that their children have been directly certified as eligible for NSLP and that, unless the parents object, their children will receive free meals. Other districts use an active consent process through which parents are required to respond affirmatively before their children can receive free meals through direct certification. The direct certification option has lowered school district administrative costs (particularly when implemented through computer matching), reduced the proportion of inaccurate eligibility determinations, and increased NSLP participation by eligible children.23 As a result, Congress enacted legislation in 2004 requiring all school districts to implement direct certification (at a minimum, for children in households receiving food stamps) by the 2008–09 school year. The legislation also provided federal grants to help pay school districts’ implementation costs, including for development of information technology infrastructure.24 The history of this auto-enrollment strategy is striking. It began as a pilot project in a handful of places, became an option open to every school district, was implemented with great success in most districts, and in several years will become a standard part of the NSLP nationwide. The remainder of this report analyzes whether and, if so, how a similar strategy could be applied in another context—namely, enrollment into Medicaid and SCHIP of eligible but uninsured children and families whom other programs have already been found to have low enough incomes to qualify for health coverage. 7 FINDINGS FROM THE 2002 NATIONAL SURVEY OF AMERICA’S FAMILIES Using data from the 2002 National Survey of America’s Families (NSAF), this section explores the relationship between health coverage and the receipt of means-tested nutrition assistance by low-income children and poor parents. Earlier research analyzing the 1997 and 1999 NSAF described the proportion of low-income, uninsured children whose families participated in means-tested nutrition programs.25 This study goes beyond this earlier research in several ways. Not only does it use more recent data, it also examines two topics that have not previously been investigated: namely, the proportion of low- income, uninsured adults whose families participate in non-health programs; and, among low-income participants in non-health programs, the proportion who already receive health coverage through employers, Medicaid, or SCHIP. The current analysis examines children and parents with different income levels. As noted above, state Medicaid and SCHIP programs typically cover children with family incomes up to 200 percent of FPL; but the median state’s Medicaid program covers parents up to only 67 percent of FPL, even as an important minority of states extends parental coverage to the full FPL. Accordingly, we focus on children with family incomes up to 200 percent of FPL and parents with incomes up to 100 percent of FPL. Appendix B contains additional information for other groups of children and adults. A preliminary caution here is important. While the numbers in this analysis convey the magnitude of potential coverage gains, they do not show the precise proportion of eligible but unenrolled low-income children and poor parents who would be covered if they were auto-enrolled into health coverage based on their families’ participation in means-tested nutrition programs. Measurement error is always possible with survey data, as respondents can inadvertently mischaracterize their health coverage, income, or other factors. Moreover, the income levels portrayed below are not identical to eligibility standards for health coverage in every state, for a number of reasons: a few state Medicaid programs cover parents with incomes above the FPL; many states set their income eligibility limits for parents well below the FPL; a small number of uninsured children in households receiving nutrition assistance may have incomes above 200 percent of FPL and still qualify for SCHIP because they live in states with unusually high income eligibility limits for the latter program; a small percentage of children live in states that do not extend Medicaid and SCHIP up to 200 percent of FPL; and some uninsured recipients of nutrition assistance may be ineligible for health coverage for reasons other than income, such as immigration status or excess assets. More broadly, because state eligibility rules vary greatly, no single set of nationwide income levels and other selection criteria for analysis 8 could replicate with precision each state’s eligibility criteria for health coverage. Despite these considerations, the groups with estimates shown below are reasonable proxies for typical Medicaid and SCHIP eligibility, particular in the case of children. Uninsured, Low-Income Children and Nutrition Assistance As explained in Appendix A, we analyzed data from the 2002 NSAF concerning the relationship between income, health coverage, and participation in three means-tested nutrition programs: the National School Lunch Program (NSLP), the Special Supplemental Program for Women, Infants, and Children (WIC), and the Food Stamp Program. We found that, in 2002, more than two-thirds (71%) of low-income, uninsured children lived in families who received one or more of these nutrition benefits (Figure 4). Figure 4. Percentage of Low-Income, Uninsured Children Whose Families Participated in Means-Tested Nutrition Programs, 2002 Percent 75 71 59 50 25 22 8 0 NSLP WIC Food stamps Any of these three programs Notes: Low-income children have family incomes at or below 200 percent of FPL; NSLP is National School Lunch Program; WIC is Special Supplemental Program for Women, Infants, and Children. Source: Authors’ tabulations based on 2002 National Survey of America’s Families (NSAF). This suggests that most low-income, uninsured children who qualify for Medicaid or SCHIP could be enrolled in these health programs if they obtained coverage based on their families’ receipt of means-tested nutrition assistance. The maximum income eligibility for NSLP and WIC is 185 percent of FPL. For the Food Stamp Program, gross income levels may not exceed 130 percent of FPL, and the maximum net income level is 100 percent of FPL. Accordingly, children whose families receive these nutrition benefits are nearly always income-eligible for SCHIP, which extends to 200 percent of FPL or higher in most of the country, as noted above. 9 Although families receiving nutrition assistance include the majority of uninsured, low-income children, they also include children who already have insurance from Medicaid, SCHIP, or employers. In fact, the latter significantly outnumber the former (Figure 5). Figure 5. Health Insurance Distribution Among Low-Income Children Whose Families Participated in Means-Tested Nutrition Programs, 2002 Percent 100 6 2 17 12 16 Uninsured 2 2 2 75 Other coverage 50 56 66 84 57 Medicaid/S CHIP 25 25 20 24 ES I 0 8 NSLP WIC Food Any of stamps these three programs Notes: Totals may not equal 100% because of rounding; low-income children have family incomes at or below 200 percent of FPL; ESI is employer-sponsored insurance; NSLP is National School Lunch Program; WIC is Special Supplemental Program for Women, Infants, and Children. Source: Authors’ tabulations based on 2002 National Survey of America’s Families (NSAF). The data shown in Figure 5 serve as a reminder that employer-based health insurance fails to reach many low-income children, thus highlighting the need for publicly funded assistance. In terms of auto-enrollment, these results further suggest that state programs to enroll children into Medicaid and SCHIP based on their families’ receipt of non-health benefits would likely require efficient screening mechanisms to identify children who already have health coverage. As explained in more detail below, such screens would be required to focus scarce administrative resources on enrolling uninsured children (rather than reenrolling those who are already covered) and to comply with federal laws that limit the publicly funded health coverage that states are allowed to provide when children receive employer-sponsored insurance. Uninsured, Poor Parents, Nutrition Assistance, and Medicaid Among uninsured parents who are poor—that is, those with incomes at or below 100 percent of FPL—more than four of five (83%) either (a) live in families who participate in means-tested nutrition programs or (b) have one or more children who receive Medicaid. Over half (55%) of all uninsured parents who are poor have at least one child who participates in the National School Lunch Program. Similarly, 53 percent of uninsured, poor parents have one or more children who already receive Medicaid (Figure 6). 10 Figure 6. Percentage of Uninsured, Poor Parents Whose Families Participated in Means-Tested Nutrition Programs, 2002 Percent 100 83 75 55 53 50 29 22 25 0 NSLP WIC Food stamps One or more Any of these children nutrition or receive child health Medicaid programs Notes: Poor parents have the following characteristics: their income is at or below FPL; they are ages 18–64; and they live with a stepchild, biological child, or adopted child under age 18. Since Medicaid rather than SCHIP covers children with family incomes below FPL, this chart describes children in terms of their receipt of Medicaid rather than SCHIP. NSLP is National School Lunch Program; WIC is Special Supplemental Program for Women, Infants, and Children. Source: Authors’ tabulations based on 2002 National Survey of America’s Families (NSAF). Poor parents whose families receive nutrition assistance have very low rates of employer coverage (16% or less, depending on the type of nutrition assistance received). Among poor families participating in NSLP or WIC, nearly half of the parents (46% and 48%, respectively) are uninsured, and about a third already receive Medicaid (Figure 7). Figure 7. Health Coverage Among Poor Parents Whose Families Participated in Means-Tested Nutrition Programs or Whose Children Received Medicaid, 2002 Percent 100 Uninsured 32 75 46 48 41 46 4 Other coverage 3 50 4 2 3 34 57 49 36 Medicaid/S CHIP 37 25 16 13 8 7 14 ES I 0 NSLP WIC Food Child in Any of stamps Medicaid these programs Notes: Totals may not equal 100% because of rounding. Poor parents have the following characteristics: their income is at or below FPL; they are ages 18–64; and they live with a stepchild, biological child, or adopted child under age 18. ESI is employer-sponsored insurance; NSLP is National School Lunch Program; WIC is Special Supplemental Program for Women, Infants, and Children. Since Medicaid rather than SCHIP covers children with family incomes below FPL, this chart describes the children in terms of their receipt of Medicaid rather than SCHIP. Source: Authors’ tabulations based on 2002 National Survey of America’s Families (NSAF). 11 The previous section’s conclusions about low-income children apply to poor parents as well. That is, auto-enrollment strategies are capable of reaching many uninsured, poor parents who qualify for Medicaid, and the very low percentage of poor parents receiving employer coverage reinforces their need for publicly funded coverage. As with children, auto-enrollment systems for poor parents will need an efficient mechanism to identify parents who already receive health insurance. On the other hand, uninsured, poor parents face a situation that is distinct in important ways. Automatically enrolling them into health coverage based on their families’ receipt of nutrition assistance may have an even more significant “payoff” than such measures could achieve with children, since nearly half of poor parents in households participating in nutrition programs are uninsured—a much higher percentage than the proportion of uninsured among low-income children in such families. In addition, more than half (53%) of poor, uninsured parents have children who receive Medicaid, and very few of this group (7%) have employer coverage. These findings suggest that states wishing to reach poor parents who qualify for Medicaid but are not enrolled could cover many of them quite efficiently by identifying and enrolling them based on their children’s Medicaid case files. As discussed below, such a strategy could be pursued under current law, without a federal policy change, since it does not require any interface between Medicaid and other means-tested programs. By contrast, auto-enrollment strategies that seek to link Medicaid and non-health programs—a category that includes all auto-enrollment strategies aimed at children—cannot be effective without federal policy changes, as shown in the following section. OBSTACLES TO AUTO-ENROLLMENT This section summarizes existing research about auto-enrollment efforts directed at children. It describes significant obstacles that have made it impossible, as a practical matter, for state officials to pursue this strategy effectively. It then identifies other challenges that, while not insurmountable, have complicated auto-enrollment efforts and thus need to be considered in the development of future auto-enrollment programs. Unless otherwise noted, the examples in this part of the report are taken from The Children’s Partnership’s Web site on Express Lane Eligibility.26 12 Limited Information Technology Because of the shortcomings of state agencies’ computer systems, auto-enrollment can be a labor-intensive process. The computer systems used by different public benefit programs are often incompatible. As a result, information about particular children often must be obtained, conveyed, and evaluated by hand in order for a child receiving non-health benefits to be enrolled into Medicaid or SCHIP. In many cases, state agencies administering means-tested programs lack the administrative resources to invest in the new information technology (IT) needed for health programs to interface digitally with non- health programs. Without a significant, upfront investment in IT, ongoing program administration can thus be cumbersome, costly, and ultimately unsustainable. For example: • In recent years, California has mounted a major effort to enroll free school lunch participants into Medicaid, triggered by parental requests for health coverage on the school lunch application form. Unfortunately, information about children’s prior Medicaid enrollment is divided among multiple offices with incompatible computer systems, including social service agencies in 58 counties. As a result, school personnel have been unable to launch automated queries to determine whether particular children were already enrolled in Medicaid or SCHIP. Frequently, parents have requested that their children be enrolled in Medicaid through the school lunch application even though their children already received Medicaid or SCHIP. As a result, tremendous amounts of staff time have been devoted to processing the health coverage applications of children who were already insured, causing frustration over this waste of scarce administrative resources in the schools. • In Maryland, applications for children’s health coverage receive expedited processing if the children or members of their households collect food stamps or cash assistance. Children who appear likely to qualify for health coverage based on information already provided for purposes of food stamps or cash assistance are immediately enrolled in three months of interim health coverage while the state completes the full eligibility determination for Medicaid or SCHIP. However, because of limitations of the state’s computers, the calculation of children’s potential eligibility for even interim health coverage must be done by hand, using paper worksheets. This is one important reason why fewer than 400 applications a month go through this otherwise promising system. • When children apply for health coverage in Nebraska, state workers examine other case files to see if information already provided for purposes of the Food Stamp Program, prior Medicaid applications, or other programs can expedite the eligibility determination process for health coverage. Some case records are 13 computerized, but others are stored in paper form and must be accessed by hand. In addition, even after the requisite files are located, case workers must manually determine whether the information in those files establishes potential eligibility for health coverage. This effort requires a significant time commitment from both staff and supervisors. Accordingly, this potentially effective program does not operate in many parts of the state. Conversely, the states and localities that have achieved some success pursuing auto-enrollment strategies have often benefited from the serendipitous availability of IT. For example: • In New York City, both Medicaid and the Food Stamp Program share a common Management Information System operated by the Human Resources Administration (HRA). HRA staff accordingly can match computer files to identify children who receive food stamps but not Medicaid. As a one-time initiative, HRA sent the parents of such children a notice stating that, unless the parents returned a form declining child health coverage, the children would have their eligibility for Medicaid determined and would be enrolled, if eligible. Only 2 percent of the families opted out, and the remaining children, when found eligible, were enrolled into Medicaid. Altogether, more than 15,000 children received Medicaid through this effort.27 • In Vermont, the state WIC agency conducts Medicaid outreach for children, which has led to the development of effective computer interfaces between the WIC and Medicaid programs. Since 1989, a single application process has been used for both WIC and Medicaid, so that an application arriving at either program is processed for both purposes. As a result, 97 percent of children receiving WIC in Vermont have health coverage. • With a “Y2K” grant, Louisiana made IT improvements connecting the state’s Medicaid program to income databases maintained by the state’s workforce agency and the state agency that administers the Food Stamp Program and cash assistance. Before Hurricane Katrina, the state routinely tapped automated data streams showing family income and other key facts whenever children’s health coverage was slated for re-determination. With 60 percent of re-determinations, this automatic process renewed children’s coverage, without any need to contact the families for more information. The percentage of children whose coverage was terminated for procedural reasons, such as failure to provide requested documentation, dropped from more than 25 percent to less than 4 percent. 14 Inflexible Federal Rules for Medicaid Eligibility A second barrier to auto-enrollment in Medicaid and SCHIP is a federal prohibition against state reliance on the prior determinations of other means-tested programs that have already found families to have low enough incomes that their children qualify for health coverage. For example, a casual observer might think that all children receiving food stamps must necessarily qualify for Medicaid, since both programs extend eligibility to children with net incomes up to 100 percent of FPL. However, despite these identical income-eligibility standards, a small proportion of children who qualify for food stamps may, in theory, be ineligible for Medicaid, since the two programs use slightly different methodologies to determine income. For example, Medicaid and Food Stamps have different definitions of the household members whose resources and needs are taken into account in determining eligibility, different “disregards” that are subtracted from a household’s gross income to arrive at a net income figure, etc.28 If such methodological differences mean that, in theory, a non-health program would classify as income-eligible even one child whom Medicaid would find ineligible, federal law prohibits Medicaid and SCHIP from relying on the final income determinations of that other, means-tested program.29 In a number of states, this federal doctrine has meant that, although non-health programs have already found children to have family incomes low enough to qualify for Medicaid or SCHIP, parents have been required to complete a second and generally redundant application for health coverage. This consumes state administrative dollars and families’ time, without substantial justification. More important, such “repeat applications” can prevent eligible children and parents from enrolling. Even when application forms are simple and require little or no documentation, many families either do not complete them or require costly assistance to complete the forms properly, For example: • In a two-phase initiative, Washington State treated the application form for NSLP as, in effect, commencement of a Medicaid application. In the first phase, the NSLP application form asked families to affirmatively authorize sharing information with the Medicaid program. After state officials did a manual check of case files to identify and exclude children who were already enrolled in Medicaid, the remaining families were sent abbreviated, supplemental forms requesting additional information needed to determine Medicaid eligibility. In the initiative’s second phase, information from the NSLP form was automatically forwarded to the Medicaid program unless parents objected. The Medicaid office used a sophisticated computer algorithm, rather than checking files by hand, to identify families most likely to qualify for health coverage. Only such families were sent 15 supplemental applications. In both phases, only 31 percent of the targeted families completed and returned the supplemental forms needed for children to receive health coverage. • In California, children received presumptive Medicaid based on information provided by the school lunch program. Families were asked to complete a supplemental information form to continue their children’s health coverage beyond the presumptive stage. Coverage ended for most children, since supplemental forms needed for children to continue receiving health coverage were completed and returned for approximately 25 percent of the children who received presumptive eligibility. • Chicago’s application form for the school lunch program asks parents to authorize sharing information with the Medicaid agency as well as community-based agencies, for purposes of outreach and enrollment into child health coverage. Each fall, the Chicago public schools obtain computerized data from the state’s Medicaid program identifying children who are enrolled in the school lunch program but not Medicaid or SCHIP. The schools then send these families written information about health coverage, including invitations to events at which application assistance is provided. School outreach workers screen incoming health applications to make sure they are complete, following up with families to obtain missing information. On average, three to four contacts per family are required to obtain complete applications. Although at initial outreach events 80 percent of applications have been incomplete, the intensive and costly work performed by school staff has led to an 80 percent final approval rate for child health applications that come from the schools. Several final comments on this issue are warranted. First, after the conclusion of presumptive eligibility, more of the above-described families in California and Washington may have completed the Medicaid application process if it had been further simplified through reducing or eliminating requirements for families to collect and present documents showing eligibility. In other words, while auto-enrollment may yield greater take-up than even the simplest application procedure, simpler application procedures do increase enrollment. Second, limited follow-through is not confined to low-income families offered health coverage. For example, as noted above, the vast majority of eligible employees fail to enroll in retirement security plans if they must sign up on their own. One particularly dramatic example concerns retirement savings programs where the employer matches worker contributions dollar-for-dollar, up to a limit. For most workers, participation in 16 such programs has a cost, because workers must reduce their take-home pay to fund the contributions that “draw down” the employer match. However, workers over age 59½ can, in effect, obtain their full employer match without making any contributions. Federal tax law permits these workers, without penalty, to withdraw their contributions at any time, including 10 seconds after such contributions are made. Nevertheless, at jobs where employees must affirmatively enroll, roughly half of these older workers fail to draw down the employer’s full match, even after they receive consumer education and financial incentives to participate, according to one study. On average, these non-participants were found to forgo the equivalent of 1.3 percent of their annual pay. In the words of the study authors, the employees’ failure to complete the forms needed to take full advantage of this employer benefit was like leaving “$100 bills on the sidewalk.”30 Third, when technical differences in eligibility methodology rules are combined with limited IT, the administrative obstacles can become daunting. As described above, New York City officials did a one-time exercise to enroll into Medicaid children who received food stamps. However, while the Food Stamp Program computers and Medicaid agency computers could “talk to each other,” the two programs applied different eligibility methodologies for determining whether children’s net income was below the FPL. As a result, without funding to develop new software, the agency’s computer system could not determine a child’s Medicaid eligibility based on Food Stamp Program information. The determination had to be done manually, by program staff. That is one important reason why this promising effort was never repeated. Other Challenges The above sections describe problems that have prevented states from moving forward effectively. However, other factors can complicate but not necessarily doom a state’s attempt to pursue auto-enrollment strategies. Child Health Programs Divided Between Medicaid and SCHIP Many states operate bifurcated child health programs, with Medicaid serving the poorest uninsured children and separate SCHIP programs serving somewhat higher-income children. In many states, a single family can have its younger children in Medicaid and its older children in SCHIP. This administrative complexity creates multiple challenges, including some that interfere with effective and efficient auto-enrollment. For example, within a single state, SCHIP and Medicaid programs can have different or even incompatible computer systems, making it costly and complex to determine whether a particular child who is being “auto-enrolled” already has coverage. 17 At a more basic level, in states operating two child health programs, it is not enough simply to say that, based on findings of a non-health program, a child has family income low enough to qualify for health coverage. Auto-enrollment systems must determine the correct health program into which the child is placed. In making that determination, the SCHIP statute requires that, before enrolling a child in a separate SCHIP program, a state must screen the child for potential Medicaid eligibility and enroll the child in Medicaid if he or she qualifies. Relying on the final income determinations of non-health programs, as discussed above, could be problematic when such determinations would place the children into SCHIP. The current “screen and enroll” requirement, without modification in such cases, would make auto-enrollment largely ineffective with these near-poor children, since their families would be forced to go through the standard Medicaid application procedure, with the associated administrative burdens for the state and family and the resulting loss of coverage if families fail to complete the process. Immigration Status Immigration status must be considered carefully in the development of auto-enrollment procedures that rely on programs such as NSLP and WIC to establish eligibility for Medicaid and SCHIP. That is because NSLP and WIC serve income-eligible children, regardless of their immigration status. Medicaid and SCHIP, by contrast, are limited to U.S. citizens and certain permanent legal residents (mostly those who have lived in the U.S. for five years or longer). Accordingly, even if the income findings of NSLP and WIC were used to determine income-eligibility for child health coverage, families would still need to provide information about citizenship or immigration status before their children could receive Medicaid or SCHIP. Moreover, establishing auto-enrollment mechanisms from NSLP and WIC into health coverage needs to be done carefully, to avoid discouraging immigrant families from seeking nutrition assistance for their children. While this challenge must be faced, it can be overcome. For example, in California, linking the school lunch program to child health coverage actually increased school lunch enrollment in many cases. Some immigrant families do not apply for health coverage because they fear that such an application could lead to their classification as a potential “public charge,” thereby risking their capacity to legalize or even retain satisfactory immigration status. Some families worry that their immigration sponsors (often spouses or other close family members) could be forced to repay the state for publicly funded health care costs. These 18 fears posed a serious obstacle to health coverage when Chicago attempted to move children from the school lunch program into Medicaid and SCHIP. However, through intensive outreach and costly community education, Chicago officials overcame those fears for most immigrant families. Characteristics of Non-Health Programs In terms of income eligibility standards, there is considerable overlap between many non- health programs and SCHIP (Figure 8). Accordingly, such non-health programs appear to be promising sources of auto-enrollment into child health coverage. Figure 8. Typical Income Eligibility Limits for Children, Various Means-Tested Programs Maximum income eligibility as a percentage of FPL 250 226 200 194 200 185 185 180 150 150 100 100 50 0 SCHIP Food NSLP WIC EITC LIHEAP Section 8 Child care stamps subsidies See “Notes/Sources for Figure 8” on page 37. Despite this promise, particular features of non-health programs can complicate the process of auto-enrolling children into Medicaid and SCHIP. For example, with NSLP, many school districts historically have maintained eligibility records on paper, rather than in electronic form. This obviously makes it much more costly to transfer information about particular children to state health programs. Many education officials are understandably reluctant to take resources away from education to fund such information transfer, even for a laudable effort that seeks to expand children’s health coverage. However, schools are increasingly automating these and other student records at the district or state level. Such automation of student records is being driven by many factors, including the requirements of the No Child Left Behind law, the opportunity to 19 enroll children into NSLP via direct certification, the ability of computerized data checks to verify applicants’ eligibility for NSLP, and, perhaps, the above-described IT grants Congress authorized in 2004.31 To take another example, the federal Earned Income Tax Credit (EITC) appears, at one level, to be a particularly promising vehicle for auto-enrollment. Among families with children who do not receive Medicaid, SCHIP, or food stamps, EITC reaches more than 74 percent with incomes below the FPL and over 66 percent of those with incomes below 200 percent of the FPL.32 However, income determinations for EITC purposes are non-contemporaneous. Typically, individuals receive EITC refunds in a particular calendar year based on the previous year’s income. It may be a challenge to develop systems that translate such information about the previous year’s household income into Medicaid or SCHIP eligibility during the current year. STATE AND FEDERAL POLICY OPTIONS Current federal law permits states to use information from other means-tested programs to target outreach to children and families based on their likely eligibility for Medicaid and SCHIP. However, targeted outreach, while worthwhile, does not provide the full increase in enrollment that would result if eligible children and families automatically received health coverage based on information provided by non-health programs, without any need to complete a Medicaid or SCHIP application form. Current law permits information already in Medicaid case files to be used to qualify additional individuals based on the application of standard Medicaid methods for eligibility determination. Accordingly, states could cover many uninsured parents based on information in their children’s Medicaid or SCHIP case files, as discussed below. However, current federal law forbids states from granting Medicaid or SCHIP based on the final income determinations of non-health agencies—the main type of auto- enrollment that could reach eligible but unenrolled children. As a result, policy changes in federal law may be needed for auto-enrollment to work effectively with children (and with parents whose children are not already enrolled in Medicaid or SCHIP). Finally, the development of IT needed for efficient auto-enrollment has implications for both state and federal policymakers. Federal policymakers could make an enormous contribution by making clear that enhanced federal matching funds can cover the development of such IT. Pending such federal steps, state policymakers need to remain 20 alert to opportunities to incorporate into broader IT efforts the infrastructure needed for auto-enrollment. State Policies to Enroll Parents Based on Their Children’s Medicaid Files Current federal law permits one auto-enrollment strategy: namely, placing uninsured, poor parents into Medicaid based on information in their children’s Medicaid case files. Through computer queries directed at existing Medicaid eligibility files, states could identify the families where a) children receive Medicaid but the parents do not and b) household income is low enough for the parents to receive Medicaid.33 Before extending coverage to such parents, however, the state would need to ascertain that the parents meet Medicaid eligibility criteria other than income. Examples of such criteria include citizenship or satisfactory immigration status (which can differ between parents and children within a particular family); and the parent’s possession of a valid Social Security number (which need not be presented for children to receive health coverage, but which often must be presented before the parent can enroll).34 Notwithstanding these factors, in states that eliminate asset requirements for parents, a large proportion of income-eligible parents are likely to have sufficient information in their children’s Medicaid files to determine their eligibility for Medicaid. However, in the 24 states with asset requirements that apply to parents but not children, and whenever children’s case files do not allow a full determination of parental eligibility, the state will need to obtain supplemental information from the family before granting eligibility to the parents.35 If state officials obtain that information through proactive telephone calls or in-person contact, many more eligible parents will be enrolled than if families must complete forms that request the supplemental information. Giving States Flexibility to Auto-Enroll Based on Other Programs’ Findings Despite ingenuity and commitment by state-level policymakers, the obstacles discussed above have prevented auto-enrollment from reaching more than a relative handful of uninsured children. Following is a discussion of options available to federal policymakers who wish to clear away these obstacles and give states new and practical tools with which to cover eligible children, as well as poor parents whose coverage cannot be established based on their children’s Medicaid case files. This section focuses on relatively narrow policy changes that specifically address auto-enrollment, rather than more sweeping reforms that could tackle a broader range of related issues.36 21 For states to grant Medicaid or SCHIP coverage based on the final income determinations of non-health agencies, Congress would need to amend federal Medicaid law to give states the flexibility to disregard technical differences between programs’ methods for determining income. For example, such flexibility could allow a state to qualify a child for Medicaid based on a finding by the Food Stamp program that the child’s family has net income below 100 percent of FPL. Such a child would receive Medicaid coverage without any need for the program to re-determine income based on Medicaid’s eligibility methodologies, which differ slightly from Food Stamp methodologies. In addition to achieving administrative savings and enrolling eligible individuals who otherwise would fail to complete the application process, such an approach could simplify the IT development required for transmission of information between health and non-health programs. For information related to eligibility, non- health agencies would need to communicate only the final determination of household income, not all of the particular facts that went into that income determination. Giving states this flexibility would have trade-offs. A small number of children could receive Medicaid or SCHIP who ordinarily would be denied health coverage. With the eligibility methodology used by the non-health program, these children would be found to have family incomes low enough to qualify for Medicaid or SCHIP. However, using the methodology ordinarily applied by Medicaid and SCHIP, the children’s family income would be calculated as slightly higher than maximum eligibility levels for Medicaid or SCHIP. For example, the Food Stamp program might classify a family as having net income slightly below FPL because that program subtracts from income certain shelter costs that Medicaid does not subtract. In other words, eligibility standards for health coverage would not change under this approach, but because children arriving at Medicaid and SCHIP via other means-tested programs would have had their income evaluated using different eligibility methodologies, a few families slightly above or below the eligibility threshold for health coverage might move to the other side of the line.37 One little-known element of the Bush Administration’s approach to low-income subsidies under the Medicare Modernization Act of 2003 (MMA) provides a useful precedent for analyzing this trade-off.38 For many years, state Medicaid agencies have operated the Medicare Savings Program (MSP), which pays some or all Medicare cost- sharing and/or premiums for certain low-income Medicare beneficiaries who receive no other Medicaid coverage. Under MMA, these MSP beneficiaries are automatically enrolled in a Medicare prescription drug plan and receive full subsidies if they do not affirmatively choose a plan within a certain period of time.39 The Bush Administration (acting pursuant to Congressional authorization) extends this automatic enrollment and 22 low-income subsidy to all MSP beneficiaries, including the small proportion who would not otherwise qualify for such subsidies.40 These individuals either live in the five states that do not take into account assets in determining eligibility for the MSP—unlike MMA’s low-income subsidies, which are limited to households with assets below $10,000 per beneficiary—or live in a state that uses less restrictive methodologies for evaluating income or assets than those employed by the MMA subsidy program.41 Although the Administration’s approach provides low-income subsidies to a small number of otherwise ineligible people, the approach extends subsidies to a vastly larger number of fully eligible people who otherwise would not receive assistance. Technical differences in program rules that qualify a handful of otherwise ineligible people can be disregarded if, in the words of the MMA statute, “the eligibility requirements under [the Medicare Savings Program and low-income subsidies] are substantially the same.”42 It is hard to see a compelling justification for federal policymakers to provide this automatic eligibility for low-income subsidies to Medicare beneficiaries without providing similar eligibility to low-income children and families seeking Medicaid or SCHIP. In each case, a small number of otherwise ineligible people could receive coverage because of slightly different technical rules. In both cases, the far larger effect is to enroll numerous individuals who otherwise would go without benefits for which they fully qualify, under ordinarily applicable rules. In fact, the approach discussed here for children and families is less expansive than the automatic eligibility that already applies to the MSP and MMA’s low-income subsidies, since the eligibility standards for health coverage would not broaden under the Medicaid and SCHIP approach. For example, consider a state that limits Medicaid eligibility for children ages 6 to 18 to those with family incomes at or below 100 percent of the FPL. If such a state wanted to automate Medicaid enrollment for children receiving WIC, the state would provide Medicaid to children whose family income, as determined by WIC, was at or below 100 percent of FPL; such a state could not grant Medicaid to all children in WIC-participating households, since WIC eligibility in most states extends to 185 percent of the FPL.43 By contrast, even though the MMA statute limits low-income subsidies to individuals with assets valued below specified amounts, the subsidies automatically go to seniors enrolled in the MSP—even in states where low-income seniors are eligible for the MSP, regardless of how many assets they have.44 In other words, the Medicaid/SCHIP auto-enrollment option explored here involves only differences in eligibility methodologies, not the potentially more significant differences in eligibility 23 standards already accepted by Congress and the Department of Health and Human Services for auto-enrollment of low-income Medicare beneficiaries. Policymakers interested in pursuing similar flexibility for children and families who qualify for Medicaid and SCHIP need to consider several relatively technical matters in crafting the details of such a new federal option. First, as is implicit in the above discussion of Medicaid and WIC, the fact of a family’s receipt of another benefit will sometimes be insufficient to establish eligibility for Medicaid. If income-eligibility for the non-health program extends above Medicaid eligibility standards, auto-enrollment would take place based on the income level found by the non-health program, not the non-health program’s granting of a benefit. Second, assets as well as income can affect eligibility, particularly with parents. While 45 states plus the District of Columbia base children’s eligibility for Medicaid and SCHIP entirely on income, without considering assets, 29 state Medicaid programs require parents to meet requirements related both to income and assets.45 Some non- health programs also take assets into account in determining eligibility, as well as other factors beyond income that play a role in Medicaid or SCHIP eligibility, such as state residence, citizenship, and immigration status. In such cases, federal policymakers could give states the flexibility to rely on determinations of other means-tested programs where such determinations show that a particular parent or child meets Medicaid or SCHIP eligibility requirements other than income. Third, in structuring this new option, it would be important to streamline the application of “screen and enroll” requirements when children qualify for SCHIP based on the findings of non-health programs. As noted above, such requirements, without modification, could force Medicaid application forms to be completed before these children enroll into SCHIP, entirely defeating the purpose of auto-enrollment. To prevent that result, these children could receive an expedited screening for Medicaid eligibility. Rather than requiring families to complete even simple Medicaid application forms, state officials could gather necessary information through telephone calls, cross- checking existing databases, and relying on families’ attestation of key facts, subject to random, post-eligibility audits. In addition, if family income, as found by the non-health program, is sufficiently high, Medicaid eligibility may be impossible, as a practical matter, potentially obviating the need for any further screening. Fourth, states could be allowed to use a valid Social Security number (SSN) as an option for documenting immigration status for purposes of Medicaid and SCHIP auto- 24 enrollment from programs such as WIC and NSLP that impose no immigration status requirements.46 While the family may need to supplement the SSN with proof of residence in the U.S. for five years or more, this option would still spare families burdensome and largely pointless documentation requirements that otherwise might prevent many of them from completing their children’s enrollment in Medicaid or SCHIP. Current federal law requires non-citizens seeking Medicaid or SCHIP to produce certain specific documents to show their immigration status, even though such documents serve little purpose for people known to have valid SSNs who can show residence in the U.S. for the requisite period.47 Fifth, federal statutory authorization may be required for non-health programs to convey to state health agencies information about the people who receive non-health benefits. In providing that authorization, however, federal policymakers would need to establish appropriate safeguards to protect privacy and limit to auto-enrollment the use of information provided by non-health agencies. A useful model for such safeguards is provided by NSLP legislation authorizing school districts to share information with Medicaid and SCHIP programs. That legislation permits disclosure only if the following requirements are met: the sole purpose of such disclosure must be identifying and enrolling into health coverage children who qualify for such coverage; the district and the state health agency must have a written agreement to that effect; and families must be notified about and have an opportunity to prevent such information disclosure.48 Sixth, the failure of a non-health agency to find that a given household has low enough income for children to receive Medicaid or SCHIP would need to leave room for such children to qualify for health coverage based on the standard application of the health programs’ eligibility methodologies. Under current federal law, state officials may not deny a Medicaid application until the application has been reviewed for potential eligibility under all potentially applicable categories.49 However, if streamlined auto- enrollment did not yield a finding of eligibility for health coverage, and information in state officials’ hands was not sufficient to determine eligibility based on the health agency’s normal methodology, officials could ask the applicants for additional information needed to determine eligibility under that standard methodology. Finally, federal policymakers pursuing this approach would need to decide whether states should have the same flexibility that school districts enjoy under NSLP to choose between opt-in and opt-out approaches to default enrollment. In either case, consent is required for enrollment into health coverage, but consent is manifested in different ways under these two approaches. With opt-in rules, children are not enrolled unless the 25 parents affirmatively request enrollment. Under opt-out procedures, children are enrolled unless their parents object. Each of these policy designs involves trade-offs. Opt-out procedures gain the full benefit of auto-enrollment strategies, producing a much larger increase in coverage. This is true even if the opt-in alternative would simply ask the parent to request child health coverage by checking a box on the application form for non-health benefits. On the other hand, opt-in rules reduce the odds that children will receive subsidized health coverage against their parents’ will. The latter concern may be particularly important with immigrant families who fear that applications for subsidized health coverage could harm their later ability to legalize. While that concern may or may not be well-founded, it is heartfelt in many quarters and needs to be taken into account in the development of outreach and enrollment procedures. In addition, families who enroll based on the absence of a parental opt-out may require additional monitoring and education to ensure that they use available services. Ultimately, federal policymakers may decide to give states the flexibility to make careful, case-by-case judgments on the application of opt-in and opt- out procedures to particular beneficiary groups. Federal Funding for Information Technology That Enables Efficient Data Exchange When a state develops a Medicaid Management Information System (MMIS), the federal government pays 90 percent of the cost. However, according to longstanding federal regulations, this enhanced matching rate is unavailable for “eligibility determination systems.”50 Statutory or regulatory change may thus be needed to extend this MMIS match to IT development related to auto-enrollment. Alternatively, federal policymakers could provide targeted grants to help states finance such IT improvements. New resources to develop IT infrastructure would make it possible for auto- enrollment to happen electronically, rather than by hand. This would substantially reduce the ongoing cost and administrative inconvenience of auto-enrollment, helping make this strategy sustainable. Stronger linkages between the data systems of different programs could yield the following benefits: • Enhanced IT infrastructure could automatically identify the recipients of non- health assistance who fit within categories of individuals slated for enrollment into Medicaid or SCHIP; without such infrastructure, such individuals often have to be identified through inspection of paper files and manual calculations. 26 • Information about the identity of the individuals slated for enrollment into Medicaid or SCHIP could be transmitted electronically from the non-health to the health coverage program, rather than relayed by hard copy and entered manually into the health program’s eligibility files. • Such identifying information could be compared, through automated query rather than manually, against lists of beneficiaries already receiving health coverage, thereby preventing duplicative enrollment and wasteful administrative expenses for redundant health coverage applications. • Near the end of an individual’s Medicaid or SCHIP coverage, improved IT interfaces between health and non-health programs could facilitate the automatic re-determination of continuing Medicaid or SCHIP eligibility, without requiring beneficiaries to complete forms or provide information already in the files of non- health programs. Beefing up this IT infrastructure could meet other goals as well, such as rapidly identifying and correcting erroneous eligibility determinations and fraud. It also could help state Medicaid programs comply with Section 6035 of the Deficit Reduction Act of 2005 (DRA), which requires states to obtain information about all employer-sponsored insurance provided to Medicaid-eligible individuals. This provision’s purpose is to improve states’ ability to collect reimbursement from employer plans when beneficiaries receive Medicaid services that are also covered by employer-based insurance. Federal funding that establishes a robust, electronic connection between state health programs and employer-sponsored plans could lower the public and private operational costs of complying with this new law.51 At the same time, an automated connection between employer health plans and state health agencies would be important to facilitating efficient auto-enrollment. As noted above, employer-sponsored insurance already covers many participants in means-tested nutrition programs. Federal law states that if children receive employer coverage, SCHIP may not enroll them. Similarly, Medicaid coverage is limited to services and costs not covered by employer plans. Accordingly, when children receiving non-health benefits are presented for possible auto-enrollment into Medicaid or SCHIP, states may need to identify which of those children receive employer coverage, since such identification would be needed to exclude SCHIP coverage and properly tailor Medicaid to “wrap around” the employer policy. Giving states the capacity to make such identifications efficiently would be an important role played by improved IT infrastructure. 27 Pending changes in federal policy for reimbursing investments in eligibility systems related to auto-enrollment, state policymakers interested in pursuing auto-enrollment can ensure that broader IT initiatives, many of which offer time-limited opportunities for involvement, include the IT infrastructure needed for auto-enrollment. Examples of such initiatives include the following: • Many states are developing new MMIS systems.52 The Centers for Medicare and Medicaid Services (CMS) is taking a broad and systems-wide approach to developing Medicaid IT, known as the Medicaid Information Technology Architecture (MITA). Aligned with the wider National Health Infrastructure Initiative, MITA is “intended to foster integrated business and information technology transformation across the Medicaid enterprise to improve the administration of the Medicaid program.”53 As states proceed with MMIS and MITA initiatives, policymakers could include functionalities needed for auto- enrollment. One approach worth considering in this context would integrate eligibility information into electronic health records (EHRs) that also include clinical data; EHR development is likely to be a central feature of many MITA projects. • DRA Section 6081 appropriates $75 million a year for a new program of Medicaid Transformation Grants in Fiscal Years 2007 and 2008. Such grants support “the adoption of innovative methods to improve the effectiveness and efficiency in providing” Medicaid. State matching funds are not required. Although the purposes for which such grants can be used are open-ended, pertinent examples include “methods for reducing waste, fraud and abuse.” States must report results in the areas of quality improvement, clinical outcomes, and cost savings. In awarding grants, preference will be given to “States that design programs that target health providers that treat significant numbers of Medicaid beneficiaries.” Many of the purposes for which such funding is provided involve improved clinical management of health care and tighter administration of payment for services. However, IT infrastructure development related to auto-enrollment may also qualify, based on the terms of the statute. IT that connects state Medicaid agencies with private health plans, as discussed above, would help achieve the purposes underlying this new grant program because it would allow state Medicaid programs to lower costs and prevent waste by pursuing third-party liability claims against insurers and employers. Furthermore, IT infrastructure development that connects Medicaid systems to eligibility data maintained by other public agencies could, as noted above, help detect erroneous and fraudulent Medicaid applications, in addition to enrolling eligible but uninsured children.54 28 • A number of states are developing Integrated Eligibility Systems (IES) that determine eligibility simultaneously for Medicaid and certain other benefits, such as food stamps and cash assistance.55 Such systems could be built to facilitate grants of health coverage based on information in the hands of other programs that are included within IES modules as well as to make electronic connections with programs, such as NSLP, that may be outside such modules. State officials developing proposals for IES funding should be strongly encouraged to collaborate from the outset with their colleagues at other agencies. Such early collaboration could help develop integrated funding requests for IT that allow effective and efficient inter-program data exchange to serve many important goals, including enrolling into health coverage eligible but previously uninsured children. Vehicles for Federal Policy Change Federal initiatives along these general lines could be enacted through legislation. For example, in the current Congress, S. 1049, sponsored by Senate Majority Leader Frist (R- Tenn.) and Senators Bingaman (D-N.M.), Lugar (R-In.), Cantwell (D-Wash.), Santorum (R-Pa.), Collins (R-Maine), Cochran (R-Miss.), Murray (D-Wash.), Feinstein (D-Calif.), Bond (R-Mo.), Nelson (D-Fla.), Talent (R-Mo.), and Jeffords (I-Vt.), would authorize $50 million in annual grants to states for innovative outreach and enrollment efforts (which could potentially include IT infrastructure development) and give state Medicaid and SCHIP programs the option to rely on the income determinations of non-health programs in establishing children’s health coverage. Policy innovations like those described here could also be tested within a state through a federal waiver under Section 1115 of the Social Security Act. That statute authorizes “any experimental, pilot, or demonstration project which, in the judgment of the Secretary [of Health and Human Services], is likely to assist in promoting the objectives of” the Medicaid and SCHIP statutes.56 Such waiver approval could authorize: a) enhanced MMIS federal matching rates for the state’s development of IT needed for efficient auto-enrollment; b) the state’s reliance on the determinations of other means- tested programs in establishing automatic Medicaid or SCHIP eligibility; and c) other policies outlined above. Such a waiver could result in the collection of useful information to guide future policy development, such as the number of otherwise ineligible children receiving coverage through auto-enrollment, the administrative costs of infrastructure establishment and program operations, or other data. In some ways more important, a waiver could cover uninsured but eligible children while federal lawmakers take the time required to enact national legislation. 29 CONCLUSION In this time of great partisan division, one of the few health policy goals that unites leaders in both parties is to enroll the millions of uninsured children who qualify for public health coverage. A promising strategy to reach this goal targets children whom other programs have already found to have family income low enough to qualify for Medicaid or SCHIP. Under this approach, states would have the flexibility to automatically enroll those children into health coverage, without requiring their families to submit a second and generally redundant application. Similar auto-enrollment strategies have achieved tremendous success increasing enrollment into multiple public and private programs, ranging from Medicare Part B to 401(k) retirement accounts. However, for auto-enrollment to succeed with children’s health coverage, federal action is needed. Currently, states generally lack the IT systems required to make automatic enrollment efficient; federal resources could help fill this gap. In addition, the federal Medicaid statute forbids states from relying on the determinations of other programs that assess family income using technical methods that differ even slightly from those in Medicaid. Giving states the flexibility, under federal law, to disregard such technical differences could significantly enhance states’ ability to pursue auto-enrollment initiatives. These changes could be made by statute. Alternatively, they could be tested as demonstration projects in individual states. Either way, in the short term, policymakers have an opportunity to take important steps forward that would empower states to do the hard work required to reach the remaining uninsured children and families who qualify for Medicaid and SCHIP. 30 NOTES 1 U.S. Census Bureau, Table HI-3. Health Insurance Coverage Status and Type of Coverage—Children Under 18 by Age: 1987 to 2004, Current Population Survey, 1988 to 2005 Annual Social and Economic Supplements, Health Insurance Historical Tables, Aug. 30, 2005. http://www.census.gov/hhes/www/hlthins/historic/hihistt3.html. 2 See the examples collected at http://www.expresslaneinfo.org/AM/Template.cfm? Section=Program_Examples1. 3 T. M. Selden, J. L. Hudson, and J. S. Banthin, “Tracking Changes in Eligibility and Coverage Among Children, 1996–2002,” Health Affairs, Sept./Oct. 2004 23(5):39–50. http://content.healthaffairs.org/cgi/content/full/23/5/39. Another study, examining Current Population Survey data for 2004, found that more than seven of 10 uninsured children may have qualified for Medicaid or SCHIP. State Health Access Data Assistance Center (SHADAC) and the Urban Institute, Going Without: America’s Uninsured Children, Prepared for The Robert Wood Johnson Foundation, Aug. 2005, http://www.rwjf.org/files/newsroom/ckfresearchreportfinal.pdf. 4 D. Cohen Ross and L. Cox, In a Time of Growing Need: State Choices Influence Health Coverage Access for Children and Families, Center on Budget and Policy Priorities, prepared for Kaiser Commission on Medicaid and the Uninsured (KCMU), Oct. 2005, http://www.kff.org/ medicaid/upload/In-a-Time-of-Growing-Need-State-Choices-Influence-Health-Coverage- Access-for-Children-and-Families-Report.pdf. U.S. Census Bureau, Table 1-RES: Estimates of the Resident Population by Selected Age Groups for the United States and States and for Puerto Rico: July 1, 2004 (SC-EST2004-01-RES). Release Date: Feb. 25, 2005, http://www.census.gov/ popest/states/asrh/SC-est2004-01.html. Calculations by ESRI, Feb. 2006. 5 Cohen Ross and L. Cox, 2005. In the median state, this is the maximum income eligibility level for a working parent with two children. 6 Cohen Ross and L. Cox, 2005. U.S. Census Bureau, Feb. 25, 2005. Calculations by ESRI, May 2006. 7 As of January 2004, 15 states and Washington, D.C., used waivers or state-only dollars to cover childless adults. S. Dorn, S. Silow-Carroll, T. Alteras et al., Medicaid and Other Public Programs for Low-Income Childless Adults: An Overview of Coverage in Eight States, Economic and Social Research Institute, Aug. 2004, http://www.kff.org/medicaid/loader.cfm?url=/commonspot/security/ getfile.cfm&PageID=46175. In addition to the programs noted in this study, Connecticut also operates a state-funded program for indigent childless adults. 8 T. M. Selden, J. L. Hudson, and J. S. Banthin, 2004. 9 Particularly significant opportunities in this area include: a) increasing the length of eligibility periods; b) extending to adults the enrollment mechanisms that typically apply to children; c) restoring procedural improvements that a number of states have undone in recent years; d) streamlining determinations of continued eligibility based on information already in the hands of government agencies, without asking families to meet new paperwork requirements; and e) reducing families’ requirements for producing documentation. See generally Cohen Ross and L. Cox, 2005. 10 A. Aizer and J. Grogger, Parental Medicaid Expansions and Health Insurance Coverage, Economic Research Initiative on the Uninsured Working Paper 20, June 2003, http://www.umich.edu/~eriu/pdf/wp20.pdf; R. E. Curtis and E. Neuschler, “Premium Assistance,” Health Insurance for Children: Creative Solutions, The Future of Children, Spring 2003, http://www.futureofchildren.org/usr_doc/tfoc13-1_syn11.pdf; A. Davidoff, L. Dubay, G. Kenney et al., “The Effects of Parents’ Insurance Coverage on Access to Care for Low-Income 31 Children,” Inquiry 40 (2003): 254–68, http://www.inquiryjournalonline.org/inqronline/?request= get-abstract&issn=0046-9580&volume=40&issue=3&page=254; and L. Dubay and G. Kenney, “Expanding Health Insurance Coverage to Parents: Effects on Children’s Coverage Under Medicaid,” Health Services Research 38 (2003):1283–1302, http://www.blackwell-synergy.com/doi/abs/ 10.1111/1475-6773.00177;jsessionid=eThFl8oOY1K5W9sZmR?cookieSet=1&journalCode=hesr. 11 D. K. Remler and S. A. Glied, “What Other Programs Can Teach Us: Increasing Participation in Health Insurance Programs,” American Journal of Public Health, Jan. 2003 93(1), http://www.cmwf.org/publications/publications_show.htm?doc_id=221496. 12 See, e.g., T. M. Westmoreland, Senate Democratic Policy Committee Hearing, “An Oversight Hearing on Implementation of the Medicare Prescription Drug Benefit,” Feb. 27, 2006, http://democrats.senate.gov/dpc/hearings/hearing29/westmoreland.pdf; D. C. Horner with B. Morrow and W. Lazarus, Building an On-Ramp to Children’s Health Coverage: A Report on California’s Express Lane Eligibility Program, The Children’s Partnership prepared for KCMU and The California Endowment, Sept. 2004, http://www.expresslaneinfo.org/AM/Template.cfm? Section=Home2&Template=/CM/ContentDisplay.cfm&ContentFileID=1012; The California Endowment, “California’s School Lunch/ Medi-Cal Express Enrollment Program,” Oct. 2004, http://www.expresslaneinfo.org/AM/Template.cfm?Section=Program_Examples1&template=/ customSource/programExample.cfm&programID=27&detailID=All; D. Horner, The Children’s Partnership, personal communication, 2005; Ayres, McHenry & Associates, Tracking Survey of Seniors Who Are Enrolled in the Medicare Prescription Drug Benefit, prepared for America’s Health Insurance Plans, Mar. 2006, http://www.ahip.org/content/fileviewer.aspx?docid=15332&linkid=132138. 13 See, e.g., C. R. Sunstein and R. H. Thaler. “Libertarian Paternalism Is Not an Oxymoron,” University of Chicago Law Review, forthcoming (draft paper available at http://www.bos.frb.org/economic/conf/conf48/papers/thaler.pdf). 14 L. Etheredge, Health Insurance Coverage At Transitions: What Works, What Doesn’t Work, Maryland Department of Health and Mental Health, State Planning Grant. Apr. 11, 2003, http://www.dhmh.state.md.us/hrsa/pdf/LynnEtheredge.pdf. See also C. Copeland., “401(k)-Type Plan and IRA Ownership,” Employee Benefit Research Institute, Jan. 2005, http://www.ebri.org/pdf/notespdf/0105notes.pdf. 15 D. Laibson, Impatience and Savings, National Bureau of Economic Research, Fall 2005, http://www.nber.org/reporter/fall05/laibson.html. See also J. J. Choi, D. Laibson, B. C. Madrian et al., For Better or For Worse: Default Effects and 401(k) Savings Behavior, National Bureau of Economic Research Working Paper 8651, Dec. 2001, http://papers.nber.org/papers/w8651; J. J. Choi, D. Laibson, and B. C. Madrian, Defined Contribution Plans for Passive Investors, July 5, 2004, http://post.economics.harvard.edu/faculty/laibson/papers/passive.pdf; J. J. Choi, D. Laibson, B. C. Madrian et al., Saving for Retirement on the Path of Least Resistance, Updated Draft: July 19, 2004, originally prepared for Tax Policy and the Economy 2001 under the title “Defined Contribution Pensions: Plan Rules, Participant Choices, and the Path of Least Resistance,” http://post.economics.harvard.edu/faculty/laibson/papers/savingretirement.pdf. 16 J. J. Choi, D. Laibson, B. C. Madrian et al., 2001; S. Holden and J. VanDerhei, “The Influence of Automatic Enrollment, Catch-Up, and IRA Contributions on 401(k) Accumulations at Retirement,” EBRI Issue Brief No. 283, Employee Benefit Research Institute and the Investment Company Institute, July 2005, http://www.ebri.org/pdf/briefspdf/EBRI_IB_07-20054.pdf. 17 D. K. Remler and S. A. Glied, 2003. 32 18 Access to Benefits Coalition, Pathways to Success: Meeting the Challenge of Enrolling Medicare Beneficiaries with Limited Incomes, The National Council on the Aging, June 2005, http://www.accesstobenefits.org/library/pdf/ABC%20ReportFNL62305.pdf. To the same effect is A. D. Federman, B. C. Vladeck, A. L. Siu, “Avoidance of health care services because of cost: impact of the Medicare Savings Program,” Health Affairs, Jan./Feb. 2005 24(1), http://content.healthaffairs.org/cgi/reprint/24/1/263. For a higher estimate of MSP take-up, see M. Nadel, L. Alecxih, R. Parent, and J. Sears, “Medicare Premium Buy-in Programs: Results of SSA Demonstration Projects,” Social Security Bulletin, 2000 63(3), cited in L. Summer and R, Friedland, How Asset Tests Block Low-Income Medicare Beneficiaries from Needed Benefits, Center on an Aging Society, Georgetown University, prepared for The Commonwealth Fund, May 2004, http://www.cmwf.org/usr_doc/summer_assettests_ib_727.pdf. 19 Government Accountability Office (GAO), Drug Card Education and Outreach, Nov. 18, 2005. GAO-06-139R, http://www.gao.gov/new.items/d06139r.pdf. Calculations by ESRI, Nov. 2005. 20 Medicare Payment Advisory Commission (MedPAC), Report to the Congress: Issues in a Modernized Medicare Program, June 2005, http://www.medpac.gov/publications/congressional_reports/ June05_Entire_report.pdf. 21 Eligibility based on receipt of cash assistance is limited to states that did not expand income eligibility for cash assistance after enactment of national welfare reform legislation in 1996. Categorical eligibility includes linkages to several other programs as well. For a more detailed explanation, see Z. Neuberger, Reducing Paperwork and Connecting Low-Income Children with School Meals: Opportunities Under the New Child Nutrition Reauthorization Law, Center on Budget and Policy Priorities, revised Nov. 23, 2004, http://www.cbpp.org/11-16-04fa.pdf. 22 P. Gleason, T. Tasse, K. Jackson et al., Direct Certification in the National School Lunch Program—Impacts on Program Access and Integrity—Final Report, Mathematica Policy Research, Inc., and Decision Information Resources, Inc., prepared for the Food Assistance & Nutrition Research Program of the U.S. Department of Agriculture, Oct. 2003, E-FAN-03-009, http://www.ers.usda.gov/Publications/EFAN03009/. 23 P. Gleason et al., 2003. See also N. Cole and C. Logan, Preliminary Report on the Feasibility of Computer Matching in the National School Lunch Program, Abt Associates, Inc., prepared for Office of Analysis, Nutrition, and Evaluation, USDA, Food and Nutrition Service, Jan. 2005, http://www.fns.usda.gov/oane/MENU/Published/CNP/FILES/NSLPDataMatch.pdf. 24 Z. Neuberger, 2004. 25 G. M. Kenney, J. M. Haley, and F. Ullman, Most Uninsured Children Are in Families Served by Government Programs, The Urban Institute, Series B, No. B-4, Dec. 1999, http://www.urban.org/ UploadedPDF/anf_b4.pdf; Children’s Partnership, Uninsured Children and Express Lane, United States (undated), citing Urban Institute tabulations, http://www.expresslaneinfo.org/AM/Template.cfm? Section=See_Facts_and_Statistics&Template=/CM/ContentDisplay.cfm&ContentID=5791. 26 See http://www.expresslaneinfo.org/AM/Template.cfm?Section=Program_Examples1. 27 In addition to information available through The Children’s Partnership Web site, see also S. Parrott, D. Cohen Ross, and L. Schott, Streamlining and Coordinating Benefit Programs’ Application Procedure, Center on Budget and Policy Priorities, June 22, 2005, http://www.cbpp.org/6-22-05prosim.pdf. 28 General Accounting Office, Means-Tested Programs: Determining Financial Eligibility Is Cumbersome and Can Be Simplified, Nov. 2001, GAO-02-58, http://www.gao.gov/new.items/d0258.pdf; S. Parrott and S. Dean, Aligning Policies and Procedures in Benefit Programs: An Overview of the Opportunities and Challenges Under Current Federal Laws and Regulations, Center on Budget and Policy Priorities. Jan. 6, 2004, http://www.cbpp.org/1-6-04wel.pdf. 33 29 The federal SCHIP statute gives states significantly more latitude in determining eligibility than does the Medicaid statute. Accordingly, an SCHIP program could grant coverage based on the findings of non-health programs. However, as a precondition of granting SCHIP eligibility, the children involved must be screened for full Medicaid eligibility; during the latter process, Medicaid would bring to bear its ordinarily applicable methodologies, regardless of any prior findings of non-health programs. As explained by the Centers for Medicare and Medicaid Services (CMS), only if “the income standard and rules for that [non-health] program are the same as or more restrictive than the rules . . . under Medicaid, the Medicaid agency can rely on the [other] program’s income determination.” CMS, Continuing the Progress: Enrolling and Retaining Low-Income Families and Children in Health Care Coverage, Jan. 14, 2002. This means that, if any individuals ineligible for Medicaid would be found income-eligible by the other program, Medicaid may not rely on the determination of that other program, since that other program would no longer be “the same as or more restrictive.” According to the State Medicaid Manual, “When a particular … policy can result in a more liberal treatment of a person than under equivalent … [Medicaid] policy, the [former] policy is more liberal than [Medicaid] even though it may not be in all circumstances.” State Medicaid Manual, Section 3420.2. Note: the latter section construed the meaning of “more restrictive” standards and methods used to determine eligibility for disability benefits under certain state programs, compared with standards used for SSI. In theory, states have the option to change their health eligibility methodologies to incorporate methodologies of non-health programs as additional eligibility pathways. For example, a state could qualify a child for Medicaid if the child’s income was below the applicable eligibility threshold based on either the income methodology of NSLP or Medicaid’s normal rules. However, that would require Medicaid eligibility workers to learn the income methodologies of both the standard Medicaid program and NSLP and to apply such dual methodologies to every application for a child’s Medicaid coverage, not just to potential auto-enrollees identified by NSLP. No state has ever chosen to implement this option, which is complex and potentially costly to administer. 30 J. J. Choi, D. Laibson, B. C. Madrian, $100 Bills on the Sidewalk: Suboptimal Saving in 401(k) Plans, Prepared for the National Bureau of Economic Research, revised Draft: July 16, 2005, http://post.economics.harvard.edu/faculty/laibson/papers/undersavers.pdf. 31 See, e.g., N. Cole and C. Logan, 2005. Another challenge with NSLP is that some children receive assistance without any individualized eligibility determination. Rather, all children within a given school or school district automatically receive free meals for several years after children in that region achieve certain participation levels in NSLP. In such cases, districts have no current information about the family income of most participating children. However, only 6 percent of students in schools that have NSLP receive benefits under this program feature, known as “Provision 2” and “Provision 3.” P. Gleason, T. Tasse, K. Jackson et al., 2003. Policymakers could either exclude such “geographically eligible” children from auto-enrollment into health coverage or decide that, if a particularly area is so depressed economically that all school children automatically receive federally subsidized, free meals, it may not be unreasonable to find such children eligible for health coverage as well based on likely family income at or below 200 percent of FPL, provided that children who received private health coverage could be identified and excluded from SCHIP. Among the children in these areas with income above 200 percent of FPL (the SCHIP eligibility threshold in most states), most are likely to have private coverage, since 80 percent of children with family incomes at or above 200 percent of FPL receive employer-sponsored coverage, and 5 percent have other private insurance. C. Hoffman, A. Carbaugh, H. Yang Moore et al., Health Insurance Coverage in America: 2004 Data Update (Washington, D.C.: Kaiser Commission on Medicaid and the Uninsured and the Urban Institute, Nov. 2005), http://www.kff.org/uninsured/upload/Health-Coverage-in-America- 2004-Data-Update-Report.pdf. Calculations by ESRI, Mar. 2006. Put differently, policymakers could decide that the additional cost of individualized income determinations for children living in such areas is not worth the minor weeding out of ineligible children that would result. 34 32 S. Zedlewski, G. Adams, L. Dubay et al., Is There a System Supporting Low-Income Working Families? (Washington, D.C.: Urban Institute, Feb. 2006), http://www.urban.org/ UploadedPDF/311282_lowincome_families.pdf. Calculations by ESRI, Nov. 2005. 33 In a few cases, the children’s case files may not contain the information needed to determine parental income-eligibility. For example, if a parent remarries, the new spouse’s income counts in determining Medicaid eligibility for the parent but not the child; such spousal income may or may not be in the child’s case files. In addition, with states that use different methodologies to determine income for parents than for children, the children’s case files may or may not have enough information to fully apply the methodology used for parents. 34 In some poor families, parents but not children may be enrolled in employer coverage. That would not affect parental eligibility, but it would limit Medicaid coverage to wraparound services and cost-sharing. 35 D. Cohen Ross and L. Cox, 2005. Calculations by ESRI, Mar. 2006. 36 Such larger reforms that will not be discussed in this paper but that deserve consideration include integrating Medicaid and SCHIP into a single national program with a federal matching rate between current Medicaid and SCHIP levels; increasing state flexibility to provide health coverage to low-income children, regardless of immigration status; changing the rules governing the interaction between employer coverage and public programs; and changing the MMIS regulations to provide enhanced federal match for eligibility-related IT improvements in appropriate cases. 37 This creates horizontal equity problems as well. Under the approach described in the text, some children who would be denied coverage if they were to apply directly for Medicaid or SCHIP could instead receive coverage if they were auto-enrolled from a different, means-tested program. In other words, within this small group of children, receipt of coverage would depend not on the children’s need for help, but on how their application was submitted. 38 Other precedents are provided by adjunctive eligibility for WIC, pursuant to which Medicaid beneficiaries are automatically deemed eligible for WIC; and receipt of certain forms of cash assistance, which automatically establish Medicaid eligibility. In fact, federal regulations forbid state Medicaid agencies, under certain circumstances, from requiring any applications from individuals who receive Supplemental Security Income (SSI) or state supplements of SSI. 42 CFR 435.909. More recently, the Child Nutrition and WIC Reauthorization Act of 2004 authorized school districts to disregard, for purposes of verifying NSLP eligibility, differences between eligibility methodologies for Medicaid, NSLP, food stamps, and cash assistance. Z. Neuberger, 2004. 39 CMS, Frequently Asked Questions, Medicare Modernization Act, “How will Medicare Savings Program beneficiaries be treated by the Medicare Prescription Drug Benefit?” Answer 4037, June 9, 2004, http://questions.cms.hhs.gov; “How will full-benefit dual eligible beneficiaries be affected by the Medicare Prescription Drug Benefit?” Answer 4034, Feb. 14, 2005, http://questions.cms.hhs.gov; CMS, Introduction to the Monthly Deemed Notice, CMS Pub. No. 11166, Oct. 2005. CMS refers to enrollment as “facilitated” rather than “automated” when Medicaid coverage is limited to payment of Medicare cost-sharing. CMS, Auto-Enrollment and Facilitated Enrollment of Low-Income Populations, Apr. 5, 2005, http://www.cms.hhs.gov/States/Downloads/AutoversusFacilitatedEnrollment.pdf. 40 42 U.S. Code Section 1395w-114(a)(3)(B)(v)(II); CMS, Guidance to States on the Low-Income Subsidy, May 25, 2005, http://www.cms.hhs.gov/States/Downloads/GuidancetoStatesonLimited- IncomeSubsidy.pdf. 41 CMS, Medicare Savings Programs (MSP) Eligibility Criteria, June 9, 2005, http://www.cms.hhs.gov/States/Downloads/MSPEligibilityCriteriaChart.pdf; K. Glaun, K. Davenport, and A. Cohen, The Medicare Low Income Drug Subsidy: Strategies to Maximize 35 Participation, Medicare Rights Center, prepared for The Commonwealth Fund, Jan. 2005, http://www.medicarerights.org/lowincomeissuebriefframeset.html. 42 42 U.S. Code Section 1395w-114(a)(3)(B)(v)(II). For state officials to implement the kind of option described here, they may need the ability to rely conclusively on the findings of non-health programs, just as low-income subsidies are granted based on the MSP determinations of state Medicaid agencies, without any “piercing the surface” of such determinations to evaluate their basis. Under this approach, if a Medicaid agency correctly evaluated and processed a final income determination made by the Food Stamp Program, the Medicaid agency would not be held liable for errors made by the Food Stamp Program. That way, Medicaid agencies would be liable only for errors within their capacity to prevent and correct. 43 U.S. Department of Agriculture, Food and Nutrition Service, WIC Income Eligibility Guidelines 2005–2006, modified Mar. 16, 2005, http://www.fns.usda.gov/wic/howtoapply/ incomeguidelines05-06.htm. 44 42 U.S. Code Section 1395w-114(a)(3)(D) and (E). 45 D. Cohen Ross and L. Cox, 2005. Calculations by ESRI, Mar. 2006. Put differently, since only five states consider assets in determining children’s eligibility and 29 states consider such assets in determining eligibility for parents, 24 states consider assets for parents but not for children. 46 This policy option is taken from D. Horner and B. Morrow, Opening Doorways to Health Care for Children: 10 Steps to Ensure Eligible but Uninsured Children Get Health Insurance, prepared by The Children’s Partnership for the Kaiser Commission on Medicaid and the Uninsured, Apr. 2006. http://www.kff.org/medicaid/upload/7506.pdf. Of course, proffering a valid Social Security number would be merely one of several options for applicants to document satisfactory immigration status. Individuals who currently are not required to provide a Social Security number as a condition of applying for Medicaid would not have to do so under this approach. 47 State Medicaid Manual, Sections 3212.2 and 3212.4. Social security numbers’ validity can be verified digitally by the Social Security Administration through the State Verification and Exchange System. N. Cole and C. Logan, 2005. 48 42 U.S. Code Sections 1758(b)(2)(C)(iii)(V) and 1758(b)(2)(C)(vi). In addition, it would be helpful to clarify that the Health Insurance Portability And Accountability Act does not bar necessary disclosure of public or private health coverage information to state Medicaid and SCHIP agencies. On an analogous question, the HHS Office of Civil Rights (OCR) ruled that state Medicaid programs and Medicare Advantage plans may share information about enrollees, thereby allowing the identification of dually enrolled individuals, because this information relates to such individual’s potential eligibility for coverage, which is a subject about which disclosure is permitted, so long as the minimum necessary information is disclosed. OCR, Health Information Privacy and Civil Rights Questions & Answers, “May a Medicaid State agency and a Medicare Advantage plan share PHI to identify dually eligible enrollees?” Answer No. 1040. Nov. 7, 2005, http://healthprivacy.answers.hhs.gov. A similar analysis would presumably apply in the current context. 49 J. Perkins and S. Somers, An Advocate’s Guide to the Medicaid Program, National Health Law Program, June 2001, http://www.healthlaw.org/library.cfm?fa=detail&id=76626&appView=folder. 50 Under 24 U.S. Code Section 1396b(a)(3), the federal government pays 90 percent of the cost of establishing such a system and 75 percent of operating costs. However, 42 CFR 433.112(c) and 42 CFR 433.111(b)(3) exclude “eligibility determination systems” from such enhanced match. 51 For example, data from private plans could be warehoused in a secure setting that states would access by inputting the Social Security numbers of applicants and enrollees, thereby limiting state access to legitimate purposes needed for administration of health coverage. It would be fair 36 for Medicaid to pay the costs of developing such a system, without asking health plans to contribute, since the system would benefit Medicaid financially but harm private insurers. No such data warehouse exists currently, in any state. 52 For a state-by-state listing of the status and time frames for development of new MMIS systems, see CMS, MMIS Fiscal Agent Contract Status Report, Feb. 15, 2006, http://www.cms.hhs.gov/MMIS/Downloads/mmisfaqr.pdf. 53 W. Branch and K. Connor, CMS and State Medicaid Programs Modernization: Medicaid Information Technology Architecture (MITA), Oct. 26, 2005, http://www.omg.org/news/meetings/ workshops/ADM_2005_Proceedings_FINAL/2-2_Branch-Connors.pdf. For more information about MITA, including white papers and presentations, see http://www.cms.hhs.gov/ MedicaidInfoTechArch/01_Overview.asp#TopOfPage. 54 The pertinent provisions are codified at 42 U.S. Code Section 1396b(z). 55 See, e.g., S. Peterson, “Social Services Synchronicity,” Government Technology, Mar. 28, 2005, http://www.govtech.net/magazine/channel_story.php?channel=17&id=93498. 56 42 U.S. Code Section 1315(b). NOTES/SOURCES FOR FIGURE 8 (page 19) Notes: NSLP is National School Lunch Program; WIC is Special Supplemental Program for Women, Infants, and Children; EITC is Earned Income Tax Credit; LIHEAP is Low Income Home Energy Assistance Program; and Section 8 is a program to subsidize rental housing. For many of these programs, income eligibility standards vary by state, family size, and/or household type. In such cases, this chart presents a typical eligibility level found in many states along with a common household configuration, such as a parent with two children. These income eligibility levels reflect “net” household income, after subtracting income disregards. For example, the Food Stamp Program is limited to households with “gross” income up to 130% FPL but “net” income at or below 100% FPL. Sources: N. Cole and C. Logan, Preliminary Report on the Feasibility of Computer Matching in the National School Lunch Program, prepared for Office of Analysis, Nutrition, and Evaluation, USDA Food and Nutrition Service (Cambridge, Mass.: Abt Associates, Inc., Jan. 2005), http://www.fns.usda.gov/ oane/MENU/Published/CNP/FILES/NSLPDataMatch.pdf; Division of Energy Assistance, Office of Community Services, Administration for Children and Families, U.S. Department of Health and Human Services, User Notes on 2000 Decennial Census Tabulations of Households Estimated to be Income Eligible for LIHEAP (Washington, D.C.: DHHS, June 2005), http://www.acf.dhhs.gov/ programs/liheap/data/census02tech.doc; S. Zedlewski, G. Adams, L. Dubay et al., Is There a System Supporting Low-Income Working Families? (Washington, D.C.: Urban Institute, Feb. 2006), http://www.urban.org/UploadedPDF/311282_lowincome_families.pdf; U.S. Department of Housing and Urban Development, NOTICE PDR-2005-01, Regarding Estimated Median Family Incomes for FY 2005 (Washington, D.C.: HUD, Feb. 2005), http://www.huduser.org/datasets/ il/il05/HUD-Medians-2005Notice.pdf. Calculations by ESRI, Mar. 2006. 37 APPENDIX A. METHODOLOGY The estimates presented in Figures 4 through 7 and the Appendix Tables below are based on the 2002 National Survey of America’s Families (NSAF). The NSAF is a national household survey providing information on more than 100,000 children and adults and is representative of non-institutionalized civilian residents of the United States younger than 65 years old. The survey oversamples the low-income population, defined as households with incomes less than 200 percent of the FPL, in 13 states and the nation as a whole. Detailed information on children was collected from the primary parent, i.e., the adult in the household with the most knowledge regarding the health care and education of up to two focal children in the household (one five years old or younger and/or one ages 6 to 17). The sample includes observations on around 13,000 low-income children (defined as children ages 17 and under) and 16,000 low-income adults (defined as adults ages 18 through 64) in 2002. Insurance coverage for both children and adults is defined at the time of the survey using a hierarchy. The health insurance status variables are classified into four mutually exclusive groups in the following hierarchy: 1) the individual is covered under an employer-sponsored health insurance plan, which includes military coverage; 2) the individual is enrolled in Medicaid, SCHIP, or another state plan; 3) the individual is enrolled in a health insurance plan not included in the previous two categories; or 4) the individual is uninsured. Estimates are presented for three publicly funded nutrition programs—the National School Lunch Program (NSLP), the Special Supplemental Program for Women, Infants, and Children (WIC), and the Food Stamp Program. Families whose incomes appeared to be below 300 percent of the FPL as measured at the time of the survey and who had school-age children were asked about participation in the NSLP in the year prior to the survey. Questions about participation in WIC at some point in the past year were asked of families whose incomes appeared to be below 300 percent of the FPL and who had preschool-age children. Note that NSLP and WIC participation are measured over a different time frame than insurance coverage. Furthermore, only adults living with children ages 17 and under can be classified as being in a family that participates in the NSLP or WIC or having a child covered by Medicaid or SCHIP. Separate estimates were made for low-income adults (those having incomes less than 200 percent of the FPL), poor adults (those having incomes at or below 100 percent of the FPL), for all nonelderly adults, and for parents and other adults separately. Parents, 38 defined as those living with at least one stepchild, biological child, or adopted child under the age of 18, constitute the bulk of the adults who can be reached through these programs (e.g., 82.6 percent of the low-income adults who live in families that participate in the NSLP are parents.) This is true even for the Food Stamp Program, in which 61.3 percent of the adults who live in families receiving food stamps are parents. Insurance information is available for up to two focal children per family. It is possible that some parents who have children covered by Medicaid or SCHIP will be missed in cases where the focal children do not have Medicaid or SCHIP coverage even though another child or other children in the family have such coverage. Information from other surveys tells us that this circumstance is rare at this point in time (unpublished tabulations of the National Health Interview Survey). All survey respondents were asked about participation in the Food Stamp Program at the time of the survey. It should be noted that underreporting of participation is likely in all of these government programs. 39 APPENDIX B. ADDITIONAL TABLES Appendix Table 1. Percentage of Uninsured Children Whose Families Participated in Means-Tested Nutrition Programs, by Income, 2002 <200% FPL <300% FPL Food stamps 8% 7% NSLP 59% 55% WIC 22% 19% Any nutrition program 71% 64% None of the above 29% 36% Notes and source follow Appendix Table 6. Appendix Table 2. Percentage of Uninsured Children with Incomes Below 200 Percent of FPL Whose Families Participated in Means-Tested Nutrition Programs, by Citizenship Status, 2002 Citizens Non-citizens All children Food stamps 9% 7% 8% NSLP 56% 71% 59% WIC 21% 29% 22% Any nutrition program 69% 78% 71% None of the above 31% 22% 29% Notes and source follow Appendix Table 6. 40 Appendix Table 3. Percentage of Uninsured, Nonelderly Parents Whose Families Participated in Means-Tested Nutrition Programs, by Income, 2002 <100% FPL <200% FPL Food stamps 22% 15% NSLP 55% 53% WIC 39% 34% Child Medicaid 53% 52% Any of the above 83% 81% None of the above 17% 19% Notes and source follow Appendix Table 6. Appendix Table 4. Health Insurance Coverage Distribution of Children, by Income and Family Participation in Nutrition Programs, 2002 NSLP WIC Food stamps Any program <200% FPL <300% FPL <200% FPL <300% FPL <200% FPL <300% FPL <200% FPL <300% FPL ESI 25% 30% 20% 25% 8% 9% 24% 29% Medicaid/SCHIP 56% 51% 66% 62% 84% 84% 59% 54% Other coverage 2% 2% 2% 3% 2% 2% 2% 2% Uninsured 17% 16% 12% 11% 6% 5% 16% 15% Notes and source follow Appendix Table 6. 41 Appendix Table 5. Health Insurance Coverage Distribution of Children with Incomes Below 200 Percent of FPL, by Citizenship Status and Family Participation in Nutrition Programs, 2002 NSLP WIC Food stamps Any program Citizens Non-citizens Citizens Non-citizens Citizens Non-citizens Citizens Non-citizens ESI 26% 13% 21% 12% 9% 4% 25% 13% Medicaid/SCHIP 58% 34% 68% 37% 85% 62% 60% 36% Other coverage 2% 3% 2% 2% 2% 3% 2% 3% Uninsured 14% 50% 10% 49% 5% 32% 13% 48% Notes and source follow Appendix Table 6. Appendix Table 6. Health Insurance Coverage Distribution of Nonelderly Parents, by Income and Family Participation in Nutrition Programs, 2002 Child in NSLP WIC Food stamps Medicaid/SCHIP Any program <100% <200% <100% <200% <100% <200% <100% <200% <100% <200% FPL FPL FPL FPL FPL FPL FPL FPL FPL FPL ESI 16% 32% 13% 27% 8% 11% 7% 13% 14% 29% Medicaid/SCHIP 34% 25% 37% 28% 57% 56% 49% 42% 36% 27% Other coverage 4% 3% 2% 2% 4% 3% 3% 3% 3% 3% Uninsured 46% 40% 48% 42% 32% 30% 41% 42% 46% 41% Notes to all tables: FPL is federal poverty level; NSLP is National School Lunch Program; WIC is Special Supplemental Program for Women, Infants, and Children. For an individual to be listed in connection with a particular nutrition program, someone in the individual’s household needs to participate in that program. That person may or may not be a family member. For an adult to be listed in connection with children’s coverage through Medicaid or SCHIP, one of those health programs must cover a child who lives in the same household as that adult. That child and adult may or may not be related. ESI is employer-sponsored insurance. A parent is an adult ages 18 to 64 who lives in the same household as the adult’s stepchild, biological child, or adopted child under age 18. A child is a person under 18 years of age. Totals may not equal 100 percent because of rounding. Source for all tables: Authors’ tabulations based on 2002 NSAF. 42 RELATED PUBLICATIONS Publications listed below can be found on The Commonwealth Fund’s Web site at www.cmwf.org. Instability of Public Health Insurance Coverage for Children and Their Families: Causes, Consequences, and Remedies (June 2006). Laura Summer and Cindy Mann, Georgetown University Health Policy Institute. According to the authors of this report, Medicaid and SCHIP coverage instability can largely be averted by adopting key policies and procedures, like limiting the frequency of required renewals; developing easy, seamless transitions among public coverage programs; and setting affordable limits on premium costs. Generosity and Adjusted Premiums in Job-Based Insurance: Hawaii Is Up, Wyoming Is Down (May/June 2006). Jon Gabel, Roland McDevitt, Laura Gandolfo et al. Health Affairs, vol. 25, no. 3 (In the Literature summary). The authors of this article found that employees in states with large urban populations, such as California, Massachusetts, New York, and Pennsylvania, tend to get more value for their premium dollar than those in rural states. Gaps in Health Insurance: An All-American Problem—Findings from the Commonwealth Fund Biennial Health Insurance Survey (April 2006). Sara R. Collins, Karen Davis, Michelle M. Doty, Jennifer L. Kriss, and Alyssa L. Holmgren, The Commonwealth Fund. Among many findings noted in this survey report—prepared for the Fund’s Commission on a High Performance Health System—two of five working-age Americans with annual incomes between $20,000 and $40,000 were uninsured for at least part of the past year, which represents a dramatic and rapid rise from 2001, when just over one-quarter of this group was uninsured. Recent Growth in Health Expenditures (March 2006). Stephen Zuckerman and Joshua McFeeters, The Urban Institute. Prepared for the Commonwealth Fund/Alliance for Health Reform 2006 Bipartisan Congressional Health Policy Conference, this report reviews trends in health expenditures in the United States over the past decade, examines differences between public and private spending, and considers explanations for the growth in spending and strategies intended to contain it. Workers’ Health Insurance: Trends, Issues, and Options to Expand Coverage (March 2006). Paul Fronstin, Employee Benefit Research Institute. Prepared for the Commonwealth Fund/Alliance for Health Reform 2006 Bipartisan Congressional Health Policy Conference, this report highlights recent trends in employment-based health benefits and compares an array of policy approaches that seek to expand coverage. Rising Out-of-Pocket Spending for Medical Care: A Growing Strain on Family Budgets (February 2006). Mark Merlis, Douglas Gould, and Bisundev Mahato. In this report the authors examine the components of out-of-pocket spending and characteristics of families with high out-of-pocket costs, including income level and insurance coverage. Entrances and Exits: Health Insurance Churning, 1998–2000 (September 2005). Kathryn Klein, Sherry Glied, and Danielle Ferry. The authors of this issue brief analyze Medical Expenditure Panel Survey data for the years 1998–99 and 1999–2000 and report that 22 percent of the U.S. population experienced at least one spell without any health coverage over the two-year period, in addition to the 9 percent who were uninsured for the full two years. Churn, Churn, Churn: How Instability of Health Insurance Shapes America’s Uninsured Problem (November 2003). Pamela Farley Short, Deborah R. Graefe, and Cathy Schoen. This issue brief’s analysis of health insurance coverage in America reveals a complex and troubling picture of insurance instability and gaps in coverage over time. Eighty-five million people, or 38 percent of the population under age 65, were uninsured at some point from 1996 through 1999, based on findings from a survey that followed people’s health coverage for four years. 43