Enrolling Eligible Children In Medicaid And CHIP: A Research Update

Covering Kids doi: 10.1377/hlthaff.2009.0142 HEALTH AFFAIRS 29, NO. 7 (2010): 1350 1355 2010 Project HOPE The People-to-People Health Foundation, Inc. By Benjamin D. Sommers Enrolling Eligible Children In Medicaid And CHIP: A Research Update Benjamin D. Sommers (bsommers@post.harvard.edu) is an instructor in the Department of Health Policy and Management, Harvard School of Public Health, in Boston, Massachusetts. ABSTRACT Keeping children who are eligible for Medicaid and the Children s Health Insurance Program (CHIP) enrolled in these programs remains an important policy challenge. An earlier study showed that one-third of all uninsured children in 2006 had been enrolled in Medicaid or CHIP the previous year. Updated results show that in 2008, children enrolled in Medicaid were somewhat more likely to remain in the program than in 2006. However, more than a quarter of all uninsured children in 2008 had been enrolled in Medicaid or CHIP the year before. In other words, roughly two million children became uninsured in 2008, despite their ongoing eligibility for these programs. It is possible that fewer children may also be enrolling in public programs since 2006 because of requirements that their U.S. citizenship status be documented. Nearly six million uninsured children in the United States are eligible for public insurance, based on 2007 estimates. 1,2 Many policy makers have assumed that this is attributable to low enrollment among eligible children. However, recent research suggests that poor retention keeping eligible children enrolled in public health insurance plays a nearly equal role. 3,4 One analysis, using 2006 data, estimated that 33 percent of all uninsured children had been enrolled in Medicaid or the Children s Health Insurance Program (CHIP) during the previous year, and 42 percent of all children who were uninsured but eligible for one of the programs had been enrolled the year before. 5 This paper analyzes more recent data to explore trends since 2006, in the context of a worsening economy, evolving state enrollment procedures, and a new requirement that Medicaid applicants provide documentation of U.S. citizenship. Disenrollment from public insurance occurs for several reasons. For example, recipients can lose their eligibility through an increase in income; they can acquire other health insurance; or they can drop out, losing coverage (often unintentionally) that they are still eligible for. 4 Medicaid and CHIP eligibility must be renewed once or twice annually typically via paper application or telephone. This process can require additional documentation of income, residence, and other criteria, depending on the state. If Medicaid or CHIP enrollees do not complete the renewal process, they lose coverage. Since 2006, several policy and economic changes have occurred that may have affected the rate at which individuals remain eligible for and participate in Medicaid or CHIP. States have revised aspects of their application and renewal processes. The national economy went into a deep recession during 2008 and 2009. A new citizenship documentation requirement for individuals applying for Medicaid took effect in July 2006. New data on trends in the rate at which individuals enroll in public insurance known as take-up of coverage and the rates at which they stay enrolled called retention can provide important information to guide Medicaid and CHIP policy. This is especially true in light 1350 HEALTH AFFAIRS JULY 2010 29:7

of the CHIP Reauthorization Act of 2009 and the Patient Protection and Affordable Care Act of 2010, which both feature significant changes to public insurance programs for children. Study Data And Methods Data The general methodological approach follows that of a similar analysis of retention versus take-up for 2000 2006. 5 The primary data source is the Current Population Survey (CPS) March Supplements, 2002 2008. The CPS is a nationally representative survey conducted by the U.S. Census Bureau. It contains detailed information on insurance, demographics, income, and state of residence. In a given March Supplement, half of the households minus attrition are surveyed the following March. This allows for the creation of two-year linked observations, using household and personal identification numbers as well as sex and age to ensure correct linking. The sample (N ¼ 11; 116) contains all children ages 1 17 in Year 2 of the survey who did not have health insurance during Year 2. Infants under twelve months in Year 2 are not included, because there are no data for them in Year 1. Methods Eligibility for public health insurance is imputed based on household income and the state- and year-specific threshold percentages of the federal poverty level for CHIP, whose eligibility rules are more generous than those of Medicaid. 6 11 Noncitizens who were not enrolled in Medicaid or CHIP at any point in the study are considered ineligible for public coverage. This is a conservative approach, because permanent residents who have been in the United States for more than five years can be eligible for the programs. This exclusion removed 8.1 percent of the uninsured sample from the study. Analysis The primary outcome variables are the survey-weighted means stratified by year for the following measures of disenrollment. Outcome Measure 1 is the percentage of all uninsured children in Year 2 ( all uninsured ) who had been enrolled in Medicaid or CHIP during Year 1. This measure indicates how much public insurance disenrollment contributes to the overall number of uninsured children in the United States. Outcome Measure 2 identifies what proportion of children who were uninsured during Year 2 despite being eligible for public insurance ( eligible but uninsured ) had been enrolled in Medicaid or CHIP during Year 1. This measure reflects the roles of poor take-up and dropout in the case of children in this group. A Wilcoxon-based test, a type of statistical test for trends, was used to examine possible time trends for Medicaid and CHIP together (any public insurance) and for Medicaid alone and CHIP alone. 12 Data were tested for trends in 2002 6 and in 2006 8. The choice of 2006 as the cutoff point for analysis was based on prior robust findings of a significant trend of worsening retention from 2000 to 2006, as well as the fact that the new Medicaid citizenship documentation requirement took effect in July 2006. 5 All analyses used March Supplement survey weights and standard errors clustered at the household level to adjust for the nonindependence of insurance for children in the same home. Data analysis was performed using the statistical software Stata 7.0. Study Results Outcome Measure 1 Exhibit 1 presents the proportion of all uninsured children who had been enrolled in Medicaid or CHIP the prior year, for the years 2002 2008. In 2008, 26.8 percent of uninsured children had been enrolled in Medicaid or CHIP the previous year: 21.7 percent in Medicaid, and 5.1 percent in CHIP. Exhibit 1 reveals two distinct patterns: an increasing trend from 2002 to 2006, followed by a sharp decline since 2006. The increasing trend was statistically significant (Wilcoxon trend p<0:01, or not likely to be due to chance alone) and peaked in 2006 at 33 percent. The decline since 2006 represented a significant negative trend (p <0:01, also not likely to be due to chance). The post-2006 decline was limited to Medicaid: The share of uninsured children previously in CHIP did not experience a statistically significant decline at that point, although the CHIP curve demonstrated an isolated increase in 2005. Outcome Measure 2 Exhibit 2 presents the proportion of children who were eligible but uninsured those who had been enrolled in Medicaid or CHIP the previous year for the years 2002 2008. In 2008, 35.5 percent of these children had been enrolled in Medicaid or CHIP during the prior year: 28.6 percent in Medicaid and 6.9 percent in CHIP. The same pattern evident in Exhibit 1 persisted, as the dropout rate among eligible but uninsured children increased from 2002 to 2006. The rate peaked in 2006 at nearly 43 percent but then declined significantly (p <0:01, not likely to be due to chance) from 2006 to 2008. This effect was limited to Medicaid only. Again, the CHIP curve demonstrated an increase in 2005 but did not produce a statistically significant trend from 2006 to 2008. In terms of which program these uninsured children were eligible for, the percentage varied JULY 2010 29:7 HEALTH AFFAIRS 1351

Covering Kids EXHIBIT 1 Percentage Of All Uninsured Children Who Were In Medicaid Or The Children s Health Insurance Program (CHIP) In The Previous Year, 2002 2008 Medicaid or CHIP a Medicaid a CHIP Percent SOURCE Author s analysisof2002 2008 linked samples from the Current Population Survey March Supplements. NOTES Aversionof this exhibit showing standard error bars is available in the online Appendix, which can be accessed by clicking on the Appendix link in the box to the right of the article online. Sample (N ¼ 11; 116) consists of all children who were uninsured during Year 2 in the survey. Percent referstothefractionofthesechildrenforwhommedicaidorchipcoveragewasreportedduringyear1ofthesurvey. a 2002 2006 test of increasing trend, and 2006 2008 test of decreasing trend: p < 0:01. from 55 percent to 65 percent for Medicaid during the study period. The other 35 45 percent of uninsured children were eligible for CHIP, without a statistically significant time trend. Discussion The presence of eligible children who drop out of Medicaid and CHIP continues to pose a major policy problem. In 2008, dropping out from the programs directly contributed to roughly two million children s becoming uninsured despite their ongoing eligibility for public coverage. (This figure is the result of multiplying 35.5 percent [from Exhibit 2] by the estimated six million uninsured but eligible U.S. children.) 1 However, the trend of worsening dropout rates from 2000 to 2006 has reversed. From 2006 to 2008, the dropout rate in public health insurance played a decreasing role as a cause of uninsurance among children, while poor takeup of Medicaid and CHIP was an increasingly EXHIBIT 2 Percentage Of Uninsured But Eligible Children Who Were In Medicaid Or The Children s Health Insurance Program (CHIP) In The Previous Year, 2002 2008 Medicaid or CHIP a Medicaid a CHIP Percent SOURCE Author s analysisof2002 2008 linked samples from the Current Population Survey March Supplements. NOTES Aversionof this exhibit showing standard error bars is available in the online Appendix, which can be accessed by clicking on the Appendix link in the box to the right of the article online. Sample (N ¼ 6; 763) consists of all children who were uninsured during Year 2 in the survey and who were eligible for public coverage. Percent refers to the fraction of these children for whom Medicaid or CHIP coverage was reported during Year 1 of the survey. a 2002 2006 test of increasing trend, and 2006 2008 test of decreasing trend: p < 0:01. 1352 HEALTH AFFAIRS JULY 2010 29:7

important factor. Comparison To Underlying Insurance Trends It is important to compare these statistics to the underlying trends of insurance coverage among children. Exhibit 3 presents the same retention statistic as Exhibit 2 the percentage of uninsured but eligible children who were in public health insurance the previous year compared with three other key statistics: the percentage of all children who were uninsured, were in Medicaid or CHIP, or had private insurance. The latter three statistics come from official Census Bureau analyses of data from the Current Population Survey. 13 Overall, from 2002 to 2008, Medicaid and CHIP enrollment increased gradually, from 22.7 percent to 28.1 percent of all U.S. children, and private health insurance declined, from 68.8 percent to 64.2 percent. Meanwhile, the proportion of children without any insurance was essentially flat, varying between 10.5 percent and 11.7 percent. None of these measures correlate with the dropout trend s peak in 2006 and subsequent decline. This suggests that the changes in retention since 2006 are not due to shifts in the denominators of uninsured children or those enrolled in Medicaid. Rather, the declining dropout statistics in Exhibits 1 and 2 reflect greater relative retention and worsened take-up of public programs since 2006. Reasons For Declining Take-Up Why has take-up (enrollment) in Medicaid and CHIP become relatively more difficult, and retention relatively less so, since 2006? STATE-LEVEL POLICY CHANGES: One explanation is the implementation since 2005 of several state-level policies aimed at improving retention. Three states switched to a twelvemonth renewal period for CHIP from a morestringent six-month renewal; three states made a similar switch for Medicaid. Three states adopted twelve-month continuous eligibility for CHIP enrollees; one state did so for Medicaid, which made dropping out much less likely in the first year of coverage. One state eliminated its requirement of a face-to-face interview for CHIP renewal, and another state began using a joint renewal form for both programs. 9-11 Other, more difficult-to-measure interventions such as community outreach that targeted dropping out may also have played an important role in improving retention. However, these changes on their own are unlikely to explain the dramatic change in trajectory of the retention measures in Exhibits 1 and 2. CITIZENSHIP DOCUMENTATION: Another contributing factor that coincides with the change in trends is the federal requirement for increased citizenship documentation for Medicaid applicants. This policy, implemented as part of the Deficit Reduction Act of 2005, took effect on 1 July 2006. Before that date, most states required only that applicants sign a statement, under penalty of perjury, saying that they were citizens. 14 Since the change, people renewing or EXHIBIT 3 Comparing Retention Trends With Overall Insurance Enrollment And Uninsurance Rates Among Children, 2002 2008 Percent of all children with private insurance Percent Percent of uninsured but eligible children who were in Medicaid or CHIP the previous year a Percent of all children who were in Medicaid/CHIP Percent of all children who were uninsured SOURCES Author s analysisof2002 2008 linked samples from the Current Population Survey March Supplements; and DeNavas-Walt C, Proctor BD, Smith JC. Income, poverty, and health insurance coverage in the United States: 2007. Washington (DC): U.S. Census Bureau; 2008. NOTE A version of this exhibit showing standard error bars is available in the online Appendix, which can be accessed by clicking on the Appendix link in the box to the right of the article online. a Sample (N ¼ 6; 763) consists of all children who were uninsured during Year 2 in the survey and who were eligible for public coverage. JULY 2010 29:7 HEALTH AFFAIRS 1353

Covering Kids CHIPRA February 2009 CHIP reauthorization in 2009 includes features designed to improve eligible children s participation. These include an express lane to integrate data from other means-tested programs. applying for Medicaid coverage have had to produce either a passport or a birth certificate to prove their citizenship unless they have already done so on a prior application. Several studies suggest that the documentation requirement led to decreases in public health insurance coverage, although these analyses have not demonstrated a clear causal effect of the policy. 15,16 The timing of the policy s implementation coincides precisely with the downturn in dropout versus take-up in Exhibits 1 and 2. Furthermore, several features of the citizenship requirement suggest that it may have a larger effect on take-up than on retention. Current Medicaid enrollees are more likely to have already demonstrated their citizenship in the course of applying for other programs. For instance, most states require citizenship documentation for Supplemental Security Income (SSI). Once an individual produces the proper documents, he or she no longer has to do so again. This fact suggests that the greatest impact of the new requirement would be on first-time applicants and first-time renewals, rather than on individuals with longer periods of participation in Medicaid. 17 However, there is no direct evidence as of yet on the effect of the Deficit Reduction Act on retention versus take-up, so this discussion is speculative. The fact that many new enrollees are infants, whose families presumably have ready access to their birth certificates, may in fact place some new enrollees at an advantage when it comes to the documentation requirement. DIFFERENT TRENDS IN MEDICAID AND CHIP: Finally, one additional piece of evidence in support of this hypothesized effect of the Deficit Reduction Act comes from the differential trends in Medicaid and CHIP. The 2006 8 decline in the dropout rate in Exhibits 1 and 2 was limited to Medicaid; CHIP was not affected. The brief increase in 2005 for the CHIP curves in Exhibits 1 and 2 occurred before the Deficit Reduction Act took effect. Similarly, the act applied directly only to Medicaid procedures, although most states use shared processes for the two programs which means that many, but not all, CHIP applicants were affected by the policy. 18 Thus, it may have had a larger negative effect on Medicaid take-up, sparing retention to a greater degree. Limitations One notable limitation of this analysis stems from the fact that the Current Population Survey undercounts Medicaid participation. However, the overall effect of the undercount would be to underestimate the extent of Medicaid and CHIP dropout rates as a factor in the persistence of children who are eligible but uninsured. 5 Another key limitation is that the study design provides only a two-year window to monitor health insurance coverage. Thus, the study cannot identify currently uninsured children who were enrolled in Medicaid or CHIP more than a year before. This limitation means that the statistics in Exhibits 1 and 2 further underestimate the dropout rate among uninsured children. Lastly, the study is limited by the nature of income and asset data in the CPS, which is not as detailed as the information used by states to determine eligibility, and which uses annual rather than monthly income. Fortunately, the eligibility imputation approach used here has been shown to provide estimates similar to those from other analyses using more-detailed eligibility imputation. For instance, comparable estimates on the number of children who are eligible but uninsured have been obtained using differing imputation approaches. 1,5,19 Conclusions And Future Prospects Poor retention in Medicaid and CHIP remains a significant reason why millions of children who are eligible for public coverage are nonetheless uninsured. However, the situation appears to have improved over the past three years, although problems with initial enrollment of eligible children remain. Some of this improvement is probably attributable to streamlined renewal processes at the state level, and further steps in this direction are warranted. Much of the change in the relative impacts of poor retention and poor take-up since 2006 may be due to the new requirement that applicants document their citizenship, which may selectively deter new Medicaid applicants to a greater extent than returning applicants. The coming years offer the promise of improved Medicaid and CHIP take-up and retention as a result of the recent CHIP Reauthorization Act (CHIPRA), which was signed into law in February 2009. CHIPRA includes several features designed to improve participation rates among children who are eligible but uninsured, such as the use of bonus payments to states for streamlining the application and renewal processes. States also have a new option to use an express lane procedure that incorporates data from related means-tested programs such as school lunches and food stamps to identify and enroll eligible children in Medicaid and CHIP. Finally, CHIPRA contains a relaxed version of the citizenship documentation requirement, through which applicants can simply provide their names and Social Security numbers, which will be cross-referenced with the Social Security Administration, instead of having to present documents. 20 1354 HEALTH AFFAIRS JULY 2010 29:7

It remains to be seen whether these changes improve public insurance take-up, retention, or both. Similarly, it is unclear how the implementation of the Patient Protection and Affordable Care Act of 2010 will effect these issues. The effects of CHIPRA and health reform on these various aspects of Medicaid and CHIP will, no doubt, be an important subject for future research, as the nation continues efforts to expand coverage to millions of uninsured children who are already eligible for public health insurance. The author is grateful to Sarah Dine for suggesting this project, and to Melissa Wachterman and two anonymous reviewers for helpful revisions. NOTES 1 Holahan J, Cook A, Dubay L. Characteristics of the uninsured: who is eligible for public coverage and who needs help affording coverage? [Internet]. Issue Paper. Washington (DC): Kaiser Commission on Medicaid and the Uninsured; 2007 Feb [cited 2010 May 10]. Available from: http://www.kff.org/ uninsured/upload/7613.pdf 2 Hudson JL, Selden TM. Children s eligibility and coverage: recent trends and a look ahead. Health Aff (Millwood). 2007;26(5):w618 29. 3 Phillips JA, Miller JE, Cantor JC, Gaboda D. Context or composition: what explains variation in CHIP disenrollment? Health Serv Res. 2004;39(4 Pt 1):865 86. 4 Sommers BD. From Medicaid to uninsured: drop-out among children in public insurance programs. Health Serv Res. 2005;40(1):59 78. 5 Sommers BD. Why millions of children eligible for Medicaid and CHIP are uninsured: poor retention versus poor take-up. Health Aff (Millwood). 2007;26(5):w560 7. 6 National Governors Association. Maternal and child health (MCH) update 2005: states make modest expansions to health care coverage [Internet]. Washington (DC): NGA Center for Best Practices; 2005 [cited 2010 May 10]. Available from: http://www.nga.org/files/pdf/ 0609MCHUPDATE.PDF 7 Ross DC, Cox L. Preserving recent progress on health coverage for children and families: new tensions emerge [Internet]. Washington Medicaid and the Uninsured; 2003 Jul [cited 2010 May 10]. Available loader.cfm?url=/commonspot/ security/getfile.cfm&pageid=14355 8 Ross DC, Cox L. Beneath the surface: barriers threaten to slow progress on expanding health coverage to children and families [Internet]. Washington (DC): Kaiser Commission on Medicaid and the Uninsured; 2004 Oct [cited 2010 May 10]. Available from: http:// www.kff.org/medicaid/upload/ Beneath-the-Surface-Barriers- Threaten-to-Slow-Progress-on- Expanding-Health-Coverage-of- Children-and-Families-pdf.pdf 9 Ross DC, Cox L. In a time of growing need: state choices influence health coverage access for children and families [Internet]. Washington Medicaid and the Uninsured; 2005 Oct [cited 2010 May 10]. Available upload/in-a-time-of-growing-need- State-Choices-Influence-Health- Coverage-Access-for-Children-and- Families-Report.pdf 10 Ross DC, Cox L, Marks C. Resuming the path to health coverage for children and parents: a 50 state update on eligibility rules, enrollment and renewal procedures, and costsharing practices in Medicaid and SCHIP in 2006 [Internet]. Washington Medicaid and the Uninsured, 2007 Jan [cited 2010 May 10]. Available upload/7608.pdf 11 Ross DC, Horn A, Marks C. Health coverage for children and families in Medicaid and SCHIP: state efforts face new hurdles; a 50-state update on eligibility rules, enrollment and renewal procedures, and cost-sharing practices in Medicaid and SCHIP in 2008 [Internet]. Washington Medicaid and the Uninsured; 2008 Jan [cited 2010 May 10]. Available upload/7740_es.pdf 12 Cuzick J. A Wilcoxon-type test for trend. Stat Med. 1985;4(1):87 90. 13 DeNavas-Walt C, Proctor BD, Smith JC. Income, poverty, and health insurance coverage in the United States: 2007. Washington (DC): U.S. Census Bureau; 2008. 14 Kaiser Commission on Medicaid and the Uninsured. Citizenship documentation requirements in Medicaid [Internet]. Key Facts. Washington (DC): The Commission; 2007 Mar [cited 2010 May 10]. Available from: http://www.kff.org/medicaid/ upload/7533-02.pdf 15 Ross DC. New Medicaid citizenship documentation requirement is taking a toll [Internet]. Washington (DC): Center on Budget and Policy Priorities; 2007 Mar 13 [cited 2010 May 10]. Available from: http://www.cbpp.org/files/2-2-07 health.pdf 16 U.S. Government Accountability Office. Medicaid: states reported that citizenship documentation requirement resulted in enrollment declines for eligible citizens and posted administrative burdens [Internet]. Washington (DC): GAO; 2007 Jun [cited 2009 Feb 10]. Available from: http://www.gao.gov/new.items/d07889.pdf 17 Centers for Medicare and Medicaid Services. Documentation of citizenship for Medicaid recipients and applicants that declare they are U.S. citizens: frequently asked questions [Internet]. Baltimore (MD): CMS; 2006 [cited 2009 Feb 4]. Available from: http://www.cms.hhs.gov/ MedicaidEligibility/downloads/ AdditionalInformationon DocumentationRequirements.pdf 18 National Health Policy Forum. The basics: the Medicaid citizenship documentation requirement [Internet]. Washington (DC): NHPF; 2007 [cited 2010 Jun 7]. Available from: http://www.nhpf.org/library/ details.cfm/2740 19 Selden TM, Hudson JL, Banthin JS. Tracking changes in eligibility and coverage among children, 1996 2002. Health Aff (Millwood). 2004;23(5):39 50. 20 Horner D, Guyer J, Mann C, Alker J. The Children s Health Insurance Program Reauthorization Act of 2009 [Internet]. Washington (DC): Georgetown University, Health Policy Institute; 2009 Feb [cited 2009 Feb 17]. Available from: http:// ccf.georgetown.edu/index/thechildren-s-health-insuranceprogram-reauthorization-actof-2009 JULY 2010 29:7 HEALTH AFFAIRS 1355