Concerns about access to care for low-income children. Health Care Access And Use Among Low-Income Children: Who Fares Best?

Transcription

1 C o v e r i n g K i d s Health Care Access And Use Among Low-Income Children: Who Fares Best? Simply providing access to public or private coverage for kids is not enough to ensure that they are getting appropriate care. by Lisa Dubay and Genevieve M. Kenney 112 LOW-INCOME CHILDREN ABSTRACT: In this paper we assess how access to care and use of services among low-income children vary by insurance status. Although 40 percent of low-income children rely on private health insurance, little is known about how this coverage compares with Medicaid coverage in meeting their health care needs. We find that Medicaid and privately insured low-income children appear to have fairly comparable access but that Medicaid-covered children are more likely to receive services and to have more visits when they receive care. Expanding public coverage may not be sufficient to ensure that all low-income children have access to comprehensive and high-quality care. It may require improvements in preventive and dental care for children with private coverage, an area in which states have limited influence. Concerns about access to care for low-income children under Medicaid are long-standing, even though the Medicaid benefit package is comprehensive, including services such as well-child and dental visits under the Early and Periodic Screening, Diagnosis, and Treatment (EPSDT) program, and requires little, if any, cost sharing. Low reimbursement rates, administrative hassles, and residential segregation between providers and patients have been cited historically as factors that contribute to access problems for children covered by Medicaid. 1 Little attention, however, has been focused on barriers that may affect use of care among low-income children with private coverage. Deductibles, copayments, and limited coverage of preventive care and other benefits may decrease the likelihood that children from low-income families with private coverage receive recommended care. 2 In 1997 fully 40 percent of all children in families with incomes below 200 percent of the federal poverty level ($32,100 for a family of four) were covered by private insurance, as many as were covered by Medicaid. 3 Understanding the relative effectiveness of Medicaid and private Lisa Dubay is a senior research associate and Genevieve Kenney is a principal research associate at the Urban Institute s Health Policy Center. H E A L T H A F F A I R S ~ V o l u m e 2 0, N u m b e r Project HOPE ThePeople-to-PeopleHealthFoundation, Downloaded Inc. from HealthAffairs.org on October 30, 2018.

2 L O W - I N C O M E C H I L D R E N coverage for low-income children is important given that many states are establishing separate non-medicaid programs under the State Children s Health Insurance Program (SCHIP). Enacted in 1997, SCHIP provides states with $40 billion in federal funding over ten years to expand coverage for low-income children. Unlike previous expansions, which built upon existing Medicaid programs, states can set up separate programs to serve SCHIP enrollees. States that so choose have greater flexibility in designing benefit packages and may impose some cost sharing. Consequently, these SCHIP programs may look more like private insurance than Medicaid. 4 Surprisingly, there is little recent evidence regarding whether low-income children are better served by private insurance or by public programs such as Medicaid. The literature on the topic uses data from the 1980s and examines a limited set of use measures. 5 One recent study compared access and use for poor Medicaid-covered children to those of nonpoor children with private coverage and found children with Medicaid coverage to be worse off along a number of dimensions. 6 However, it is not clear from that analysis whether Medicaid-covered children fare worse than privately insured children because of problems within Medicaid or because they have lower incomes and socioeconomic status. 7 In this study we assess how access to care and use of services among low-income children vary by insurance coverage. We use data from the 1997 National Survey of America s Families (NSAF), fielded just prior to implementation of SCHIP, and examine a broader set of access and use measures than previous studies have done. Study Methods NSAF is a household survey that provides information on more than 100,000 children and adults representing the noninstitutionalized civilian population under age sixty-five. 8 The survey includes measures of insurance coverage, access to care, health care use, and health status. 9 With the exception of the questions about mental health and dental visits, which are asked about children age three and older, the use questions are asked about all children. Questions are asked of the adult who knew the most about a child s education and health. Although efforts were made to ascertain the actual amount of care each child received, these data, like all survey data, are subject to potential bias. 10 We focus on children in families with incomes below 200 percent of poverty. We categorize children by whether they had private coverage, were uninsured, or had Medicaid coverage for the entire year. 11 We use the past year s coverage because our access and use measures reflect patterns of care over the past year. COVERING 113 KIDS H E A L T H A F F A I R S ~ J a n u a r y / F e b r u a r y

3 C o v e r i n g K i d s 114 LOW-INCOME CHILDREN We analyzed data on 12,680 low-income children and computed variance estimates using a replication method that adjusts for the sample design. 12 We conducted multivariate analyses to control for differences across the various insurance categories in demographic, socioeconomic, geographic, and health status indicators known to be associated with access and use. 13 Ordinary least squares regression models were estimated for the visits equations, and logit models were used for the other measures. We transformed the logit coefficients into their implied marginal effects, which enables us to interpret the results as the percentage-point difference in the outcome between children in the different coverage categories. Findings On Access And Use Among Poor Children Exhibit 1 presents the descriptive and multivariate findings for the access measures, while Exhibit 2 presents the findings for the use measures. n Usual source of care. Having a usual source of care is considered a necessary component of continuous primary care and therefore a key indicator of access. 14 Comparable proportions of Medicaid-covered and privately insured low-income children in our study lacked a usual source of care or relied on the emergency room (Exhibit 1). Other things equal, uninsured children were 8.8 percentage points more likely than Medicaid-covered children were to rely on the emergency room or to have no usual source of care. n Ability to obtain needed care. Medicaid-covered and privately insured low-income children had similar levels of unmet need for medical or surgical care and for dental care. Medicaid-covered children reported higher levels of unmet need for prescription drugs than privately insured children did, and this difference persists when other factors are controlled for, but the incidence of problems appears to be low for both groups. Uninsured low-income children were much more likely than Medicaid-covered and privately insured children were to have unmet needs for medical or surgical care and dental care. Other things equal, uninsured children were 2.8 percentage points more likely than Medicaid-covered children were to have an unmet need for medical or surgical care and 7.4 percentage points more likely to have an unmet need for dental care. n Confidence and satisfaction. The reported levels of dissatisfaction with the quality of the medical care received by their family during the previous twelve months were similar for Medicaid and privately insured children. Both the descriptive and the multivariate analyses indicate that low-income Medicaid-covered children were more likely than privately insured children were to have parents who lacked confidence that their family members could obtain H E A L T H A F F A I R S ~ V o l u m e 2 0, N u m b e r 1

4 L O W - I N C O M E C H I L D R E N EXHIBIT 1 Access To Care Among Low-Income Children, By Insurance Status, 1997 Usual source of care Children with ER as usual source or no usual source of care 6.1% 5.4% 24.1% d d Unmet or delayed health needs Children not getting or postponing medical or surgical care Children not getting or postponing dental care Children not getting or postponing prescription drugs Confidence and satisfaction Children in families not confident about obtaining needed care Children in families not satisfied with care they received Number of observations 4,841 5,503 2,336 8,446 8, d 7.8 d 10.1 SOURCE: 1997 National Survey of America s Families. NOTE: ER is emergency room. a The regression models control for health status (excellent/very good, good, fair/poor), activity limitations, age (0 2, 3 5, 6 12, 13 14, 15 18), race and ethnicity (Hispanic, black/non-hispanic, other/non-hispanic), family size, education of most knowledgeable adult (less than high school, high school, college), work status of most knowledgeable adult (full time, part time, or not working), birthplace of most knowledgeable adult (U.S. or foreign-born), family income as a percentage of poverty, geographic location (central county of metropolitan area, other county in metropolitan area, nonmetropolitan county), and state. b These numbers reflect regression-adjusted percentage-point difference in the outcome between privately insured and Medicaid-covered children. c These numbers reflect the regression-adjusted percentage-point difference in the outcome between uninsured and Medicaid-covered children. d Significantly different from Medicaid at the.01 level. 8.3 d 17.7 d d 15.3 d d 5.4 d d 7.4 d d 4.4 d needed care, but this difference disappears when we focus on Medicaid-covered children whose parents are also likely to be covered by Medicaid. Among Medicaid-covered children whose families were receiving Aid to Families with Dependent Children (AFDC), 7.8 percent had parents who lacked confidence that their families could obtain needed care, compared with 14.4 percent of the Medicaid-covered children whose families were not receiving AFDC. 15 Children who received AFDC were far more likely than other Medicaid-covered children were to live in families in which the whole family was covered by Medicaid. Thus, when all family members tend to be insured, levels of confidence are similar for Medicaid-covered and privately insured children. Greater levels of dissatisfaction and lower levels of confidence were found among uninsured children compared with both privately and publicly insured children. Uninsured children were 9.2 percentage points more likely to be in families that were not confident that they could get needed care and were 4.4 percentage points more likely to be in families that were not satisfied with the care H E A L T H A F F A I R S ~ J a n u a r y / F e b r u a r y

5 C o v e r i n g K i d s their families had received, other factors equal. n Use of physician and other health services. Almost 90 percent of Medicaid-covered children had at least one visit to a physician or other health professional in the year preceding the survey, compared with 82.5 percent of privately covered children and only 61.2 percent of uninsured children (Exhibit 2). Similar patterns are found when physician and other professional visits are considered separately, and when we control for other factors. Medicaid-covered children were more likely than privately insured and uninsured children were to receive well-child care. Other EXHIBIT 2 Use Of Services Among Low-Income Children, By Insurance Status, 1997 Children to physician or other health professional Total visits among children with at least one visit Children with at least one physician visit in past year Children to other health professional in past year Children receiving at least one visit for well-child care in past year Children to dentist or dental hygienist Children receiving at least one mental health services visit in past year Children to hospital emergency room in past year 89.6% % % % % % % % d 4.0 d 72.4% e 3.1 d 29.9% % d 1.7 d 76.5% % 5.7 d 24.4% d 1.6 d 61.2% d 3.2 d 44.9% d 2.7 d 26.5% d 2.2 d 40.5% d 1.7 d 48.7% d % % d 5.5 d 1.21 e 5.6 e 1.13 d d 0.69 d 9.7 d d 17.4 d 1.67 d 25.6 d 1.37 d d 0.54 d 29.4 d Children hospitalized at least once over the past year 5.4% 2.5% d 2.0% d 1.1 e 2.0 d 1.6 d 8.2 d d 0.51 d SOURCE: 1997 National Survey of America s Families. a See Exhibit 1 for regression details. b With the exception of the visits equations, these numbers reflect regression-adjusted percentage-point difference in the outcome between privately insured and Medicaid-covered children. For the visits equations, these numbers can be interpreted as the regression-adjusted difference in the number of visits received between privately insured and Medicaid-covered children. c With the exception of the visits equations, these numbers reflect the regression-adjusted percentage-point difference in the outcome between uninsured and Medicaid-covered children. For the visits equations, these numbers can be interpreted as the regression-adjusted difference in the number of visits received between uninsured and Medicaid-covered children. d Significantly different from Medicaid at the.01 level. e Significantly different from Medicaid at the.05 level. H E A L T H A F F A I R S ~ V o l u m e 2 0, N u m b e r 1

6 L O W - I N C O M E C H I L D R E N things equal, Medicaid-covered children were 8.8 percentage points more likely than privately insured children and 25.6 percentage points more likely than uninsured children were to have received a well-child visit in the year prior to the survey. Medicaid and privately insured children were about equally likely to have received dental care during the preceding year, while uninsured children were about two-thirds as likely. However, other things equal, we find that Medicaid-covered children were 9.7 percentage points more likely to have had a dental visit than privately insured children were. The average number of well-child and dental visits among children using those services is about the same across the three categories. Regarding mental health services, we find no significant differences in the likelihood of receiving such services across categories. Medicaid-covered children who used mental health services received more visits than did children with private insurance and a similar number as uninsured children. Medicaid-covered children were the most likely to have had emergency room visits and to have received more visits when they used the emergency room. Other things equal, Medicaid-covered children were 8.2 percentage points more likely than privately insured children were to have had an emergency room visit in the year before the survey. Medicaid-covered children also were more likely to have been hospitalized over the past year relative to privately insured and uninsured children. Discussion And Policy Implications Our analysis provides new insight into how Medicaid and private coverage compare for low-income children. A pattern of slightly lower access was found for Medicaid-covered relative to privately insured children when we examined usual source of care, unmet need, and confidence. However, unmet need for prescription drugs, which was higher among Medicaid than among privately insured children, was experienced by only 3 percent of Medicaid-covered children. The other access differential that favored privately insured children was found only among certain Medicaid-covered children. Taken together, our results suggest that access to care is comparable for low-income Medicaid-covered and privately insured children. Moreover, parents of both were equally satisfied with the care their families received. With regard to use of services, Medicaid-covered children appear to have used more health care services than privately insured children did. Our findings are consistent with previous research from an earlier period that compared Medicaid-covered and privately insured children. 16 Interpreting whether lower use of services by pri- COVERING 117 KIDS H E A L T H A F F A I R S ~ J a n u a r y / F e b r u a r y

7 C o v e r i n g K i d s 118 LOW-INCOME CHILDREN vately insured children reflects lower access is not straightforward. Lower use of physician and hospital services by privately insured children could be the result of unmeasured health status differences, underuse by these children, or overuse by Medicaid-covered children. In contrast, greater use of well-child and dental care by Medicaid-covered children is likely to reflect greater access because these services are recommended for all children, regardless of health status. Thus, our findings suggest that there are some important problems with private coverage for low-income children. While our analysis suggests that Medicaid-covered children are faring as well as or better than their privately insured counterparts, it does not conclusively indicate adequate access to care, nor does it speak to the quality of the services provided for either group of children. 17 In fact, we have found evidence of problems within Medicaid. Reported use of dental services by Medicaid-covered children falls short of minimum recommended guidelines. Moreover, it is not clear that the Medicaid-covered children who received dental services were receiving preventive care. Almost four in ten Medicaid-covered children paid at least one visit to the emergency room in the twelve months prior to the survey. It is unclear whether these patterns of care reflect a lack of access to convenient primary care, quality problems, or discontinuities within the primary care system available to Medicaid children, or other factors specific to the circumstances of families living in or near poverty. n Effect of SCHIP. Implementation of SCHIP should reduce the number of low-income children lacking insurance coverage. Our analysis, which finds large gaps in access and use between insured and uninsured low-income children on the eve of SCHIP, suggests that gaining insurance coverage under either SCHIP or existing Medicaid programs should translate into greater access to care and service use. By examining a broader range of services than has been considered previously, we also bring to light additional dimensions of the problems faced by low-income uninsured children. Only 40.5 percent received a well-child visit, and fewer than half received any dental care in the twelve months preceding the survey. Moreover, our finding that more than a quarter of these parents lacked confidence that their family would obtain needed care provides an indication of the anxiety parents feel when their children lack health coverage. Previous research has shown that insuring children can reduce stress on families, which may be another benefit of SCHIP. 18 An innovation of SCHIP is that states can establish separate state programs to serve SCHIP enrollees. The majority of states are taking advantage of this option. 19 As mentioned previously, having separate SCHIP programs means that states can be more flexible in designing H E A L T H A F F A I R S ~ V o l u m e 2 0, N u m b e r 1

8 L O W - I N C O M E C H I L D R E N their benefit packages than under the Medicaid expansion option. While SCHIP plans must offer a benchmark equivalent plan that covers well-child care and has limited cost sharing, SCHIP plans can exclude or restrict coverage of services such as dental and mental health care. Consequently, it will be important to monitor the scope of benefits offered under these programs, the effects of cost sharing, and whether gaps in needed services arise. It is possible that SCHIP plans will afford even better access than under private coverage because SCHIP benefit packages may be broader and financial burdens lower than under private plans. If non-medicaid SCHIP programs raise fees or capitation rates or reduce the administrative burden that providers associate with Medicaid, greater access may also be achieved relative to Medicaid. n Crowding out. Much concern was voiced about whether SCHIP would substitute for or crowd out employer-based coverage. 20 Fear of crowding out led to provisions in the SCHIP legislation that restrict eligibility to children without creditable coverage. Thus, states have built in safeguards such as waiting periods to discourage families from dropping private coverage and enrolling their children in SCHIP. 21 However, our analysis suggests that if low-income families substitute public for private coverage, some of their children s health care needs may be better met. Therefore, it may be prudent for policymakers to reconsider what constitutes creditable coverage for children and whether, under SCHIP, states should be allowed to supplement private benefit packages with wraparound coverage for preventive care. All states responded to the SCHIP legislation by expanding eligibility for publicly subsidized insurance coverage. However, simply expanding coverage may not be sufficient to ensure that all low-income children have timely access to preventive and curative care that is comprehensive and of high quality. To meet this objective, states will need to enroll large numbers of uninsured children in Medicaid and SCHIP, correct problems in the Medicaid program, and develop SCHIP plans with adequate benefit packages and accessible service delivery systems. Unfortunately, meeting this objective also requires improvements in preventive and dental care for low-income children with private coverage, an area in which states have limited influence. COVERING 119 KIDS This report was written under the Urban Institute s State Children s Health Insurance Program Evaluation, which is supported by the Robert Wood Johnson Foundation as part of the Assessing the New Federalism (ANF) project. The ANF project has received funding from the Annie E. Casey, W.K. Kellogg, Robert Wood Johnson, Henry J. Kaiser Family, Ford, John D. and Catherine T. MacArthur, Charles Stewart Mott, David and Lucile Packard, McKnight, Stuart, Weingart, H E A L T H A F F A I R S ~ J a n u a r y / F e b r u a r y

9 C o v e r i n g K i d s Lynde and Harry Bradley, Joyce, and Rockefeller Foundations; the Commonwealth Fund; and the Fund for New Jersey. We are grateful to Niall Brennan, John Holahan, Theodore Joyce, Alan Weil, and Stephen Zuckerman for their helpful advice on earlier drafts and to Grace Ko for her painstaking research assistance. Opinions expressed are those of the authors and do not necessarily reflect the positions of the Urban Institute or its funders. 120 LOW-INCOME CHILDREN NOTES 1. F. Sloan, R. Mitchell, and J. Cromwell, Physician Participation in State Medicaid Programs, Journal of Human Resources (Supplement 1978): ; J. Mitchell, Participation in Medicaid Revisited, Medical Care ( July 1991): ; J.W. Fossett et al., Medicaid and Access to Child Health Care in Chicago, Journal of Health Politics, Policy and Law 17, no. 2 (1992): ; and P. Newacheck et al., The Role of Medicaid in Ensuring Children s Access to Care, Journal of the American Medical Association 280, no. 20 (1998): J.P. Newhouse and the Insurance Experiment Group, Free for All? Lessons from the RAND Health Insurance Experiment (Cambridge: Harvard University Press, 1994); and H.E. Freeman and C.R. Corey, Insurance Status and Access to Health Services among Poor Persons, Health Services Research 28, no. 5 (1993): N. Brennan, J. Holahan, and G. Kenney, Health Insurance Coverage of Children, Snapshots of America s Families, Snapshot B-1 (Washington: Urban Institute, 1999). 4. U.S. General Accounting Office, Children s Health Insurance Program: State Implementation Approaches Are Evolving, Pub. no. GAO/HEHS (Washington: GAO, May 1999). 5. M.L. Rosenbach, The Impact of Medicaid on Physician Use by Low-Income Children, American Journal of Public Health 79, no. 9 (1989): ; P.F. Short and D.C. Lefkowitz, Encouraging Preventive Services for Low-Income Children: The Effect of Expanding Medicaid, Medical Care 30, no. 9 (1992): ; and Freeman and Corey, Insurance Status and Access. 6. Newacheck et al., The Role of Medicaid. 7. R. Kaestner, Medicaid and Children s Access to Care, Journal of the American Medical Associatio n 281, no. 14 (1999): For details on NSAF, see P. Brick et al., National Survey of America s Families: Survey Methods and Data Reliability, NSAF 1997 Methodology Series, no. 1, July 1999, <newfederalism.urban.org/nsaf/methodology1997.html>. 9. We use imputed data on the health insurance, access, and utilization variables with missing values. Imputed values account for 1.3 percent or less of all observations for these measures. 10. In this case, there may be a tendency to report more care than was received so that caregivers do not appear negligent. Because these data are self-reported, they may understate the extent to which children fail to receive physician, well-child, or dental care. While reported levels of care may be overstated, it is unlikely that the regression-adjusted differences in utilization that are reported by insurance status will be biased. 11. Medicaid coverage includes a small proportion (5.4 percent) who indicated that they were covered by state-sponsored insurance programs. Measurement of insurance coverage on NSAF is defined in S. Rajan, S. Zuckerman, and N. Brennan, Confirming Insurance Coverage in a Telephone Survey: Evidence from the National Survey of America s Families, Inquiry (Fall 2000): We omitted children with mixed types of coverage in the past year (21 percent of all low-income children) but found, with few exceptions, that their inclusion did not alter our main results in any meaningful way. We excluded from H E A L T H A F F A I R S ~ V o l u m e 2 0, N u m b e r 1

10 L O W - I N C O M E C H I L D R E N our analysis children who receive Supplemental Security Income (SSI), have Medicare coverage, receive special education, or are in foster care (together, 8.1 percent of low-income children) because previous research indicates that their patterns of care are much different from those of other children. See I. Flores-Cervantes, M.J. Brick, and R. DiGaetano, 1997 NSAF Variance Estimation, Report No. 4, NSAF Methodology Series (Washington: Urban Institute, 1999). 13. R. Anderson, A Behavioral Model of Families Use of Health Services (Chicago: Center for Health Administration Studies, University of Chicago, 1968); and M. Grossman, The Demand for Health: A Theoretical and Empirical Investigation, National Bureau of Economic Research Occasional Paper no. 119 (New York: Columbia University Press, 1972). Control variables include perceived health status; presence of activity limitations; age, sex, race, and ethnicity of the child s family income as a percentage of poverty; urban/rural location; and state. We also control for the education, work status, and birthplace of the most knowledgeable adult. We undertook a number of supplemental analyses because of concern that the access and use differences found across different types of insurance coverage were due to unobserved differences in the children in the three different coverage categories that are not accounted for by the control variables included in the regression analyses. Because we thought that unmeasured differences in health status posed the most serious possible threat to the validity of our findings, we reestimated our multivariate models on subgroups that were likely to be more homogeneous with respect to health status. On the whole, the pattern of our findings (available on request from the authors) did not change when we narrowed the analysis to focus on these smaller subgroups, although the differences tended to have weaker statistical significance. We also found that our results on the effects of insurance coverage were not sensitive to the control variables that were included in the multivariate models. 14. A.S. Johanses, B. Starfield, and J. Harlow, Analysis of the Concept of PrimaryCare for Children and Adolescents (Baltimore: Child and Adolescent Health Policy Center, Johns Hopkins University, 1994); and B. Starfield, Primary Care: Concept, Evaluation, and Policy (New York: Oxford University Press, 1992). 15. AFDC was replaced in 1996 welfare reform legislation by Temporary Assistance for Needy Families (TANF). 16. Rosenbach, The Impact of Medicaid ; Short and Lefkowitz, Encouraging Preventive Services ; and J. Currie and D. Thomas, Medical Care for Children: Public Insurance, Private Insurance, and Racial Differences in Utilization, Journal of Human Resources 30, no. 1 (1995): M.S. Marquis and S.H. Long, Reconsidering the Effect of Medicaid on Health Care Services Use, Health Services Research 30, no. 6 (1996): J.R. Lave et al., Impact of a Children s Health Insurance Program on Newly Enrolled Children, Journal of the American Medical Associatio n 279, no. 22 (1998): F. Ullman, I. Hill, and R. Almeida, CHIP: A Look at Emerging State Programs, Assessing the New Federalism Policy Brief no. A-35 (Washington: Urban Institute, September 1999). 20. L. Dubay and G. Kenney, Lessons from the Medicaid Expansions for Children and Pregnant Women: Implications for Current Policy, Statement before the House Ways and Means Subcommittee on Health, Hearing on Children s Access to Health Coverage, Serial (Washington: U.S. Government Printing Office, 1999); and J. Gruber, Statement before the House Ways and Means Subcommittee on Health, Hearing on Children s Access to Health Coverage, Serial (Washington: U.S. GPO, 1999). 21. GAO, Children s Health Insurance Program. COVERING 121 KIDS H E A L T H A F F A I R S ~ J a n u a r y / F e b r u a r y