PEDIATRICS Vol. 106 No. 1 July 2000, pp. 14-21

A Comparison of the Socioeconomic and Health Status Characteristics of Uninsured, State Children's Health Insurance Program-Eligible Children in the United States With Those of Other Groups of Insured Children: Implications for Policy

Gayle R. Byck, PhD

From the School of Public Health, University of Illinois at Chicago, Chicago, Illinois.

	ABSTRACT

Top Abstract Methods Results Discussion Conclusion References

Objectives.  To describe the sociodemographic and health status characteristics of the national uninsured, State Children's Health Insurance Program (SCHIP)-eligible population, and to compare this population with Medicaid-enrolled children, privately insured children, and privately insured children who have family income in the SCHIP eligibility range.

Procedures.  Data were analyzed for 50 950 children 0 to 18 years of age included in the 1993 and 1994 National Health Interview Surveys. The survey obtained information on insurance coverage and sociodemographic and health status measures. Bivariate analyses were conducted to identify the relationships between SCHIP eligibility and sociodemographic and health status characteristics. Multivariate analyses were conducted to assess the independent association of the sociodemographic and health status variables with the likelihood of being uninsured, SCHIP-eligible.

Primary Findings.  Results indicate that SCHIP children exhibit markedly different socioeconomic and health status characteristics than do both Medicaid- enrolled and privately insured children, although these differences are less significant in privately insured children. SCHIP children more often live with college- educated (39.4%) and employed adults (91.2%) than do Medicaid-enrolled children (23.0% and 53.9%, respectively). However, SCHIP children live with college-educated and employed adults less than do all privately insured children (66.7% and 96.9%, respectively) and privately insured/same-income children (57.8% and 97.0%, respectively). Parents of SCHIP-eligible children are also disproportionately self-employed or employed in industries (eg, retail trade) and occupations in which health insurance coverage is less available or affordable. SCHIP-eligible children are also 2 times more likely to be adolescents and 11/2 times more likely to be in excellent health than Medicaid-eligible children. Compared with privately insured children, SCHIP-eligible children are nearly 3 times more likely to be Hispanic and nearly 2 times more likely to be rated in fair or poor health.

Conclusions.  The results demonstrate that uninsured, SCHIP-eligible children are substantially different from children in these groups, particularly compared with Medicaid-enrolled children. These differences need to be taken into account when setting policies and implementing programs intended to increase health insurance coverage and access to health care.  Key words:  children, health insurance, State Children's Health Insurance Program, Medicaid.

There are ~11.6 million uninsured children in the United States, or 15.3% of all children in 1997.¹ Lack of health insurance affects children in all facets of their life, not just health. Children who do not receive primary and preventive care generally use inappropriate, more expensive services and have more serious medical problems. Uninsured children are more likely to miss school and fall behind in school,² which may affect their future educational and employment opportunities and prevent them from achieving their full potential. Uninsured children have decreased access to care and lower utilization of health care services than either privately insured or publicly insured children.^3-11

The State Children's Health Insurance Program (SCHIP) made federal funds available to states as of October 1997 to expand health insurance coverage for children <19 years of age with family incomes up to 200% of the federal poverty level, or up to 50 percentage points above the state's current Medicaid level. As of December 31, 1998, 43 of 52 approved state SCHIP plans were in operation. States have been implementing SCHIP through a Medicaid expansion (26 states), creation of a separate program (14 states), or a combination of these 2 options (12 states).¹² In addition, some states, such as Illinois, are instituting an insurance rebate program to subsidize parents whose family incomes meet SCHIP income criteria for at least some of their costs of employer-based health insurance for their children.¹³

Implementation of SCHIP programs through either a Medicaid expansion or a separate state program, while at least partially based on financial and political considerations, may presume similarity of SCHIP-eligible children to Medicaid-enrolled or privately insured children in terms of sociodemographic and health status characteristics. SCHIP focuses on uninsured children in a very specific family income level. This specific group of uninsured children has not been well-studied, raising questions about how programs should be implemented. In addition, there is limited knowledge about how the specific uninsured population eligible for SCHIP differs from Medicaid-enrolled children and privately insured children, including privately insured low-income children.

This article compares the sociodemographic and health status characteristics of uninsured children who are eligible for SCHIP with those of other groups of children, using 1993 and 1994 National Health Interview Survey (NHIS) data. The following research questions are addressed: 1) What are the characteristics of the national uninsured, SCHIP- eligible population? For this analysis, SCHIP-eligible is defined as children <19 years of age living in households with incomes <200% of the federal poverty level; and 2) How does the uninsured, SCHIP-eligible population differ from Medicaid-enrolled children, all privately insured children, and privately insured children with family income that matches SCHIP eligibility? The results are then discussed in terms of state choices for SCHIP policy and implementation. This research is of considerable current policy significance, because states are in the process of implementing, refining, and evaluating their SCHIP programs.

	METHODS

Top Abstract Methods Results Discussion Conclusion References

Data Source

The 1993 and 1994 NHIS were used for this research.^14,15 Only the second half of 1993 was used because the health insurance questions were not asked of respondents during the first 2 quarters of 1993. The 2 years were pooled to increase the sample size, particularly for subgroup analyses. Combining the 1993 and 1994 data provided 50 950 children 0 to 18 years of age who were never married and are not living alone, for this study for whom family level data (ie, employment status of parents/guardians) could be obtained.

The NHIS collects information about the health and other characteristics of each living member of the sample household. Variables of importance to this research include sociodemographic characteristics, insurance status, health status, chronic conditions, doctor visits, usual source of care, and number of bed days.

Description of Variables Used

Insurance Coverage The health insurance supplement to the NHIS included a series of questions on insurance status. Children in the sample were classified as uninsured if they were not reported to be covered by private health insurance, Medicaid, Medicare, CHAMPUS, the Indian Health Service, or other public assistance programs during the month before the interview date. Children with unknown insurance status (15%) were not included in the analysis.

Socioeconomic and Health Status Characteristics The core survey instrument includes questions on sociodemographic and health status characteristics, including age, race, ethnicity, region, urban/rural status, health status rating, number of chronic conditions, and activity limitations. Children were also matched to the reference person to obtain family education and employment information. Item nonresponse rates were negligible for the sociodemographic variables, and <10% for the health status variables.

Statistical Analysis

Data analyses were conducted using STATA (Stata Corporation, College Station, TX), a statistical program that accounts for the complex sample design and uses the weights to make national estimates.¹⁶ Results of bivariate comparisons between the SCHIP group and each of the 3 other groups are presented here. Multivariate logistic regression was also used to assess the independent association of the socioeconomic (age, race and ethnicity, family education and employment, family structure, region, and urban/rural status) and health status variables (health rating, chronic conditions, and activity limitations) with the likelihood of insurance coverage. The school activity limitation variable was included in the bivariate comparisons but was excluded from the multinomial logistic regression because it only applied to children 5 through 17 years of age.

	RESULTS

Top Abstract Methods Results Discussion Conclusion References

Characteristics of Uninsured, SCHIP-Eligible Children and Other Children

Children who were uninsured, SCHIP-eligible were different from the other groups of children in terms of sociodemographic, family level, and health status characteristics. Relative to the Medicaid-enrolled population, the SCHIP population is proportionately older, less minority, more suburban and rural, and live in better educated and more 2-parent families; they are also in better health and have fewer chronic health conditions and activity limitations. Compared with the all privately insured group, the SCHIP group is more Hispanic and urban and also live in households with parents/guardians who are less educated and less likely to both be employed, as well as in fewer 2-parent families. When the SCHIP group is compared with privately insured/same- income children, the findings are the same as when it is compared with all privately insured children.

Sociodemographic Characteristics The distribution of sociodemographic characteristics for the 4 groups of children and the odds ratios (ORs) and 95% confidence intervals (CIs) for bivariate comparisons between the SCHIP group and each of the other 3 groups are presented in Table 1. Looking at age and race/ethnicity, the majority of uninsured, SCHIP-eligible children are between 5 and 18 years of age (74.6%), white (83.5%), and non-Hispanic (78.2%). More of these children reside in the South (40.9%) or West (26.6%), and reside in metropolitan statistical areas outside the central city (43.5%).

Family Level Characteristics Table 2 presents the distribution and ORs, respectively, for the family level variables. A strong majority of children who are SCHIP-eligible live with both parents (79.9%) and live in families in which the head of the household has at least a high school education (81.7%) and at least 1 parent is employed (91.2%). The privately insured groups have higher education levels for heads of household than do the uninsured/SCHIP-eligible and Medicaid-enrolled groups. Among all the groups, the key seems to be in the proportion of responsible adults with at least some college education; the SCHIP group is only .32 (95% CI: .28-.37) and .47 (95% CI: .41-.54) times as likely as the all privately insured and privately insured/same-income groups, respectively, to have a parent with a college education.

Health Status SCHIP-eligible children are overwhelmingly in good health and most have no chronic health conditions (76.5%) or school (94.9%) or major (94.0%) activity limitations (Table 3). Although proportionately more SCHIP-eligible children are in fair or poor health than privately insured children (OR: 1.84; 95% CI: 1.39-2.45), they are more likely to not have any chronic health conditions (OR: 1.18; 95% CI: 1.06-1.31; Table 3).

Factors Associated With SCHIP Eligibility

Because it is important to understand what influences SCHIP eligibility, and because many socioeconomic variables such as education and employment are interrelated, multinomial logistic regression was performed. (To obtain mutually exclusive groups, the all privately insured group was redefined to consist only of those children who were privately insured but not in the SCHIP income eligibility range.) The multivariable results reinforce the results presented above (not shown). It is important to note that of the 3 comparison groups, the uninsured/SCHIP group is most similar to the privately insured/same-income group; however, key differences still remain in terms of head-of-household education and employment, as well as Hispanic ethnicity and geographic region. Also, there are numerous statistically and practically significant differences between the SCHIP and Medicaid groups that should caution policymakers about lumping these 2 populations together.

	DISCUSSION

Top Abstract Methods Results Discussion Conclusion References

The results demonstrate that the uninsured, SCHIP-eligible group is different in terms of both statistical and policy relevant significance than the Medicaid-enrolled, privately insured, and privately insured/same-income groups. This should affect the ongoing design and implementation of SCHIP policies and programs. For example, outreach and enrollment have been a prominent concern of states, the federal government, and children's advocates. Take-up rates among the Medicaid-eligible seem to be lower at the higher end of the income eligibility spectrum, as well as among 2-parent, higher educated, working families.^17,18 Because the SCHIP group is even stronger on these characteristics, enrollment has been, and will continue to be, a challenge. In states using a Medicaid expansion, even with a different program name, the stigma associated with Medicaid will likely be a factor in enrolling children of educated and employed parents. Cost-sharing provisions also affect enrollment rates. Evidence from voluntary enrollment in state-subsidized health insurance programs for the working poor show that even low premium levels discourage participation^19-22; premiums may be especially discouraging for parents of healthy children.

Policy and Program Considerations for SCHIP as a Medicaid Expansion

Implementing SCHIP through a Medicaid expansion may be based on assumptions that the SCHIP and Medicaid groups are similar, when in fact the results indicate that this is not so. Medicaid has been traditionally focused on caring for infants and young children, whereas SCHIP is widely expected to benefit adolescents.^23,24 The SCHIP population consists of disproportionately more adolescents than does the Medicaid group and will require a different mix of health care services (eg, substance abuse and sexually transmitted disease prevention and treatment rather than immunizations and treatment for otitis media) to be available from appropriately trained providers. Working with the SCHIP group, a group with different needs and characteristics than the Medicaid population, will be a challenge to program administrators and providers used to dealing with the poorest of the poor. These parents generally do not frequent locations where outreach for low- income programs has traditionally occurred, such as welfare offices and inner-city locations. Although Medicaid recipients have been traditionally concentrated in inner-city areas, there will need to be an increase in participation rates of providers in suburban areas to meet increased demand from the SCHIP population.

Children who are uninsured/SCHIP-eligible are also rated in better health, have fewer chronic health conditions, and are less likely to have school and activity limitations than are Medicaid-enrolled children. This may also discourage families from enrolling in state SCHIP plans that require premium contributions. These differences in health may translate into lower service needs and, thus, change the dynamics of the services and population needs. In turn, this may encourage providers in states with managed care Medicaid programs who receive capitation to increase their participation in the SCHIP portion of the program at the expense of the traditional Medicaid population.

Policy and Program Considerations for SCHIP as a Separate Program

Differences between the SCHIP population and the privately insured groups are perhaps most relevant for states choosing to establish a separate program. Implementing SCHIP through a separate program through a health plan not used to caring for low-income individuals will, as in 1 example, require oversight and training to ensure that nonfinancial access barriers, such as language and transportation, are addressed. The much greater proportion of Hispanics in the SCHIP population than in the privately insured populations necessitates having materials printed in Spanish, as well as hiring and training bilingual health care providers and administrative personnel. The large proportion of Hispanics in the SCHIP group is especially notable and has implications for SCHIP outreach and illustrates potential inequities in income, education, and employment that may have led to this distribution.

Children eligible for SCHIP are less likely to live in 2-parent households than are privately insured children, a characteristic associated with having increased access to private health coverage.²⁵ In addition, these SCHIP-eligible children also have parents or guardians less educated than do those who are privately insured; plans and providers may need to spend more time and effort educating their parents on how to work within the system, especially if it is a managed care plan, and on health promotion and prevention practices. The need to focus on prevention and health promotion is also indicated by the data showing that the SCHIP group has slightly lower health status ratings than the privately insured groups, although they have fewer chronic health conditions. These children are also more likely to live in urban areas, so accessibility (proximity and transportation access) to providers and services must be addressed.

The explanation for the higher proportion of privately insured children with chronic conditions is not clear. One possible explanation is that parents experience job lock or are forced into certain types of employment (eg, working for a large corporation vs working for a small company or being self-employed) because of the existence of a chronic health condition in their child and the corresponding need to assure private health coverage. Another explanation may be that better educated and employed parents are more aware of chronic health problems. In contrast, low-income parents with healthy children may have less incentive to purchase health insurance coverage.

Programs for the Privately Insured Working Poor

The comparison between the uninsured/SCHIP-eligible and the privately insured/same-income as SCHIP group is especially relevant in states like Illinois, which is implementing an insurance rebate program as part of its overall SCHIP programs. The insurance rebate program in Illinois allows for reimbursement to families meeting eligibility requirements that are already enrolled in employer-based or private insurance; however, there are no requirements that the private insurance plans provide coverage for preventive or well-child care. This approach addresses the desire to keep children and families in private, as opposed to public, insurance plans. The most notable differences between these groups are that the SCHIP group has less educated heads of households, more families with no employed parents or 1 of 2 parents employed, more urban residents, more Hispanics, and more single- or no-parent households; the health status indicators are mixed. In many respects, the group eligible for the insurance rebate is much like the entire population of privately insured. However, because they may be underinsured, participation in the program, at least in Illinois, may not bring their health care coverage up to par with the preventive services offered to both SCHIP enrollees and Medicaid recipients. Future research should examine differences in utilization and health outcomes to determine the effect of underinsurance among low-income populations.

Policy Considerations for Children in Fair or Poor Health

There are also likely concerns about poor health status of uninsured children and associated demand for, and cost of, health care services. Although the proportion of SCHIP-eligible children with fair or poor health may be larger than the proportion of privately insured children, the actual numbers tell a different story. The 2.9% of uninsured, SCHIP-eligible children in fair or poor health represent ~170 000 children nationwide, which is only 17% of the number of all privately insured children in fair or poor health (984 000) and 15% of Medicaid-enrolled children with similar health status (1 165 000). Insuring these children should not overwhelm the health care system in terms of health care utilization or expenditures.

SCHIP and Parental Education Levels

The differences related to education are perhaps the most striking in light of previous research and theory on how educational status affects health care access, utilization, and health status, as well as vice versa.^26-29 Education may be seen as a predictor of occupational and labor market position; because our health insurance system is primarily employment-based, this has profound implications for health care coverage. Individuals who are better educated may be more open to health education, less inclined to partake in risky behaviors, and may have more skills that enable them to navigate complex health care bureaucracies. Studies have shown that higher levels of educational attainment are associated with lower adult mortality. In addition, low maternal education is associated with lower utilization of health care services for children.³⁰

Differences in head-of-household education levels among these groups are important. Among the working poor, a college education may be the difference between having a job with health insurance benefits and having one without this benefit. When comparing the SCHIP and Medicaid-enrolled groups, higher education may mean that the adult was able to find a job, albeit one that does not offer health insurance benefits. The higher education and employment levels of the SCHIP population, compared with the Medicaid-enrolled group may require markedly different outreach strategies than have occurred with Medicaid, even if the state is implementing SCHIP through a Medicaid expansion. For example, outreach will have to take place in workplaces. However, the differences in employment patterns between the SCHIP and private insurance groups point to a need to target specific industries (eg, retail trade) and types of workers (eg, self-employed) to maximize outreach efforts. For example, perhaps the self-employed can be reached through professional associations and chambers of commerce. It may be possible to approach businesses that do not offer family health insurance benefits to help spread the word about SCHIP to their employees; of course, the political feasibility of this may be tempered by concerns over crowd-out. Regarding the option of implementing SCHIP through a Medicaid expansion or through the creation of a separate program, the high employment rate of SCHIP parents may indicate that this is a group that is more comfortable enrolling in a private sector program than in a government program.

Limitations of the Data

Income is the key criterion for defining SCHIP eligibility; unfortunately, the income variable on the NHIS is categorical rather than continuous, making the definition of the target population imprecise, as described in the example in the "Methods" section. A noted limitation of the NHIS is that it only measures insurance status at the time of the survey and does not have information on an individual's insurance status over a whole year. This is likely to underestimate the number of persons who are uninsured at some time over the course of a year. Even for insured persons, it is not possible to distinguish between those with adequate coverage and those who are underinsured.³

Other limitations focus on the appropriateness of the data and variables studied. It may be that variables that were not part of the NHIS would have shown important differences between the SCHIP group and the other groups. Because the NHIS is a cross-sectional dataset, the variables studied do not indicate causality; for example, the results do not inform us whether children are uninsured because they are in poor health or vice versa.

	CONCLUSION

Top Abstract Methods Results Discussion Conclusion References

This study illustrates the important differences that exist among uninsured, SCHIP-eligible children and Medicaid-enrolled children, all privately insured children, and privately insured/same-income children. These differences must be addressed and accounted for when building on existing, or creating new, health care financing and delivery systems for this population. Outreach and enrollment, provider network design, provider training, and provision of services are all examples of areas in which strategies must be tailored to the specific characteristics of SCHIP-eligible children and their families.

	ACKNOWLEDGMENTS

This research was conducted as part of my doctoral research and was supported by Project MCJ-009041 from the Maternal and Child Health Bureau (Title V, Social Security Act), Health Resources and Services Administration, Department of Health and Human Services.

I acknowledge the assistance of Carol Simon, Sylvia Furner, Arden Handler, Judy Cooksey, and Naomi Morris.

	FOOTNOTES

Currently, Dr Byck is a visiting senior research specialist at the Illinois Center for Health Workforce Studies, University of Illinois at Chicago, Chicago, Illinois.

Received for publication Sep 21, 1999; accepted Nov 8, 1999.

Reprint requests to (G.R.B.) University of Illinois at Chicago, Illinois Center for Health Workforce Studies (MC 275), 850 West Jackson Blvd, Suite 400, Chicago, IL 60607-3025. E-mail: gbyck1{at}uic.edu

	ABBREVIATIONS

SCHIP, State Children's Health Insurance Program; NHIS, National Health Interview Survey; OR, odds ratio; CI, confidence interval.

	REFERENCES

Top Abstract Methods Results Discussion Conclusion References

1. American Academy of Pediatrics. Health Insurance Status of US Children Under Age 19, 1993-1997. Elk Grove Village, IL: American Academy of Pediatrics; 1998
2. Children's Defense Fund. Key Facts About Uninsured Children. Washington, DC: Children's Defense Fund; 1998
3. Holl JL, Szilagyi PG, Rodewald LE, Byrd RS, Weitzman ML Profile of uninsured children in the United States. Arch Pediatr Adolesc Med 1995; 149:398-406
4. Kogan MD, Alexander GR, Teitelbaum MA, The effect of gaps in health insurance on continuity of a regular source of care among preschool-aged children in the US. JAMA 1995; 274:1429-1435
5. Newacheck PW, Hughes DC, Stoddard JJ Children's access to primary care: differences by race, income, and insurance status. Pediatrics 1996; 97:26-32
6. Newacheck PW, Stoddard JJ, Hughes DC, Pearl M Health insurance and access to primary care for children. N Engl J Med 1998; 338:513-519
7. Weigers ME, Weinick RE, Cohen JW. Children's Health, 1996. Rockville, MD: Agency for Health Care Policy and Research; 1998. Medical Expenditure Panel Survey chartbook 1, Agency for Health Care Policy and Research publication 98-0008
8. Weinick RM, Weigers ME, Cohen JW Children's health insurance, access to care, and health status: new findings. Health Aff 1998; 17:127-136
9. Stoddard JJ, St Peter R, Newacheck P Health insurance status and ambulatory care for children. N Engl J Med 1994; 330:1421-1425
10. Spillman BC The impact of being uninsured on utilization of basic health care services. Inquiry 1992; 29:457-466
11. Wood DL, Hayward RA, Corey CR, Freeman HE, Shapiro MF Access to medical care for children and adolescents in the United States. Pediatrics 1990; 86:666-673
12. Department of Health and Human Services. A Preliminary Estimate of the Children's Health Insurance Program Aggregate Enrollment Numbers Through December 31, 1998: Background Only. Washington, DC: Department of Health and Human Services; 1999
13. National Governor's Association Center for Best Practices. NGA Supplement, State Children's Health Insurance Program, 1998: Annual Report, As of March 22, 1999. Washington, DC: National Governor's Association Center for Best Practices; 1999
14. National Center for Health Statistics. 1993 National Health Interview Survey [database on CD-ROM]. CD-ROM series 10, no. 9, SETS Version 1.21a, accompanied by 1 manual. Washington, DC: US Government Printing Office; 1993
15. National Center for Health Statistics. 1994 National Health Interview Survey [database on CD-ROM]. CD-ROM series 10, no. 9, SETS Version 1.21a, accompanied by 1 manual. Washington, DC: US Government Printing Office; 1994
16. Stata Corporation. Stata Statistical Software, Release 5.0. College Station, TX: Stata Corporation; 1997
17. Selden TM, Banthin JS, Cohen JW Waiting in the wings: eligibility and enrollment in the State Children's Health Insurance Program. Health Aff 1999; 18:126-133
18. Dubay LC, Kenney GM The effects of Medicaid expansions on insurance coverage of children. Future Child 1996; 6:152-161
19. Call KT, Lurie N, Jonk Y, Feldman R, Finch MD Who is still uninsured in Minnesota? Lessons from state reform efforts. JAMA 1997; 278:1191-1195
20. Diehr P, Madden CW, Cheadle A, Martin DP, Patrick DL, Skillman S Will uninsured people volunteer for voluntary health insurance? Experience from Washington State. Am J Public Health 1996; 86:529-532
21. Ku L, Coughlin TA. The Use of Sliding Scale Premiums in Subsidized Programs. Washington, DC: Urban Institute; 1997
22. Madden CW, Cheadle A, Diehr P, Martin DP, Patrick DL, Skillman S Voluntary public health insurance for low-income families: the decision to enroll. J Health Polit Policy Law 1995; 20:955-971
23. Newacheck PW, Brindis CD, Cart CU, Marchi K, Irwin CE Adolescent health insurance coverage: recent changes and access to care. Pediatrics 1999; 104:195-202
24. Newacheck PW, Hughes DC, Cisternas M Children and health insurance: an overview of recent trends. Health Aff 1995; 14:244-254
25. Weinick RM, Monheit AC Children's health insurance coverage and family structure. Med Care Res Rev 1999; 56:55-73
26. Blane D Social determinants of healthsocioeconomic status, social class, and ethnicity. Am J Public Health 1995; 85:903-905
27. Adler NE, Boyce WT, Chesney MA, Folkman S, Syme SL Socioeconomic inequalities in health: no easy solution. JAMA 1993; 269:3140-3145
28. Lantz PM, House JS, Lepkowski JM, Williams DR, Mero RP, Chen J Socioeconomic factors, health behaviors, and mortality: results from a nationally representative prospective study of US adults. JAMA 1998; 279:1703-1708
29. Sorlie PD, Backlund E, Keller JB US mortality by economic, demographic, and social characteristics: the National Longitudinal Mortality Study. Am J Public Health 1995; 85:949-956
30. Bates AS, Wolinsky FD Personal, financial, and structural barriers to immunization in socioeconomically disadvantaged urban children. Pediatrics 1998; 101:591-596