Family structure, socioeconomic status, and access to health care for children. (Research Brief).
Research has produced inconsistent findings on the level of health care access and utilization among children of single mothers as compared with children in two-parent families. A recent study reported that children with single mothers were slightly less likely than those living with two parents to have a regular source of medical care or to see a physician and were more likely to have an unmet health care need (Simpson et al. 1997). Another analysis found that children in families headed by single mothers were less likely to receive health care than children in two-parent families (Cunningham and Hahn 1994). Conversely, Newacheck (1992) found that family structure was not associated with the number of physician visits that children receive. An older study found that children in two-parent families had more physician visits, but among healthier children, children of single mothers had more physician visits (Cafferata and Kasper 1985).
Whereas health policy makers cannot influence family structure, they set rules around Medicaid and Child Health Insurance Program (CHIP) eligibility that relate to family structure. Thus, an understanding of its impact on child health insurance rates is critical. For this study, we hypothesized that overall, children in single-mother families would have less adequate health care access and utilization than children in two-parent families and this difference would result from an interaction between family structure and SES. We expected there would be few family structure-related differences in health care access and utilization at high levels of maternal SES, whereas children of single mothers would be worse off than children in two-parent families at low levels of SES. The rationale for the hypothesis of an overall family structure disparity in access to children's health care was the evidence that lower SES is consistently related to lower levels of health care access and utilization, and single mothers ten d to be of lower SES. In addition, single mothers may have less social support, greater time demands, and less employer-sponsored health insurance than married mothers, resulting in greater difficulty obtaining care for their children. However, we expected to observe no family structure disparity in access to care among families of high SES because higher SES single mothers might have resources that would allow them to overcome barriers to care.
This analysis uses a nationally representative data set to examine the relationships between family structure, health insurance, and access to health care for children. This study expands on previous research by linking parents to children and by including several measures of access and utilization to paint a fuller picture of access to care among children of single mothers. Although a previous analysis in this area (Newacheck 1992) found that family structure was not related to physician visits after controlling for socioeconomic factors, that study did not specifically address the impact of the potential interaction, which we examine here, between family structure and SES on access to health care for children.
This analysis used the National Health Interview Surveys (NHISs) for 1993, 1994, and 1995. Observations from the first half of 1993 were excluded because information on health insurance coverage, available from a supplement to the standard NHIS questionnaire, was not ascertained during the first half of 1993. The NHIS, a continuing survey of a stratified national household probability sample of the civilian, noninstitutionalized population, is conducted by the National Center for Health Statistics (Massey et al. 1989). This analysis included data from the core interview as well as supplements on health insurance, family resources, and access to health care (Adams and Marano 1995). This study focused on children under the age of 18. A linkage was performed between data for parents living in the home and child data; 96 percent of children were successfully linked to one or two parents. Most of the unlinked children were living with a relative other than a parent, whereas there was insufficient information for l inkage for the remaining children. Analyses were limited to children living either in two-parent households or with a single mother; children who lived in single-father or other types of households were excluded because there were too few children living in other family structures for meaningful inference (approximately 5 percent of children overall). Assignment to single-mother or two-parent family structure was based on linkage to parents and further information about adults in the household.
This analysis examined health care access and utilization, comparing children of single mothers with children in two-parent families. Weissman and Epstein (1994) have characterized access to health care as "the attainment of timely, sufficient, and appropriate health care of adequate quality such that health outcomes are maximized" (p. 6). In practice, access is a complex concept that may encompass several dimensions, such as having a single provider or group of providers who the child sees regularly at one location to provide continuity of care, or perceiving few barriers to care, such as high cost, long waiting times, or restrictive appointment times. For this study, three measures of inadequate health care access and utilization were used, involving the parent reporting (1) the child had not visited a physician in the past year, (2) the child did not have a usual source (person or place) where they obtained health care, and (3) the child had any unmet health care needs. Children whose parents said emergenc y rooms were their usual source of care were considered to have no usual source of care, as emergency rooms do not generally provide continuity of care. Unmet health care needs were ascertained via several questions: (1) Did the child need but was unable to get medical care? (2) Was medical care delayed due to cost? (3) Did the child need dental care but was unable to get it? (4) Did the child need prescription medicines but was unable to get them? (5) Did the child need eyeglasses but was unable to get them, or (6) did the child need mental health care but was unable to get it? The association between family structure and unmet need for health care was similar across these various measures of unmet need; thus, they were combined into a summary variable of having any unmet need. In this study, the word access is used to signify both health care access and utilization.
All data analyses were performed using SUDAAN (Shah, Barnwell, and Bieler 1995) to account for the complex sample design of the NHIS. To examine our hypothesis that SES would affect the relationship between family structure and access to care, models were computed that included an interaction term between maternal education (as the measure of SES) and family structure. SES may encompass various measures; maternal education was selected because it may be the most appropriate variable when addressing issues of access to care, as more educated mothers may have greater knowledge about children's health care needs and because it had the least amount of missing data. Interaction terms were significant (p < .001 for each model), and thus, the analysis was stratified by maternal education (less than 12 years, 12 years, 13 to 15 years, 16 or more years) to examine the impact of family structure on access at each education level.
Unadjusted logistic regression models were computed first (model A) to determine the overall effects of family structure on access to care. Adjusted models (model B) were used to examine the impact of additional variables considered as potential confounders of the relationships: race-ethnicity, as family structure and access to care vary by race and ethnicity (measured as non-Hispanic White, non-Hispanic Black, or Hispanic); child's health status, as health status is likely to impact odds of accessing care and varies by SES (good to excellent, or fair to poor); child's age, as younger children are more likely to access health care (0 to 4, 5 to 9, or 10 to 17), and whether or not the mother was employed, as employment impacts health insurance as well as time to access care (working part-time or full-time vs. unemployed or not in labor force). Several other variables were also considered as potential confounders but were not included in multivariable models because they did not influence the point estimates for the variable of interest: metropolitan statistical area (MSA) residence, the presence of a grandmother or other female adult in the house hold, and family income.
Health insurance coverage plays a unique role in this analysis. Health insurance, measured here as employer-sponsored private insurance, Medicaid, other type, or uninsured, is an important determinant of access to health care. The child's health coverage lies in the causal pathway between family structure and access to health care for children because public programs such as Medicaid are associated with family structure: During 1993-95, children receiving AFDC were automatically Medicaid eligible, but during this time period, program rules in many states excluded children in two-parent families from receiving AFDC. Health insurance coverage would likely be a primary reason for an impact of family structure on access to care. Thus, cross-tabulations were used to examine the relationships between single-parent status, health insurance coverage, access to care, and maternal education, and adjusted models were computed that excluded health insurance coverage. As a final step, health insurance coverage was added to adjusted models (model C) to examine whether residual disparities in access to care could be explained by family structure differences in health insurance coverage.
This analysis was limited to the 71,428 children ages 0 to 17 in the NHIS between the second half of 1993 and the end of 1995 who were either Hispanic, Black, or White and who lived in either a two-parent or a single-mother household. Children were excluded from models if they were missing data on health insurance (n = 6,375), maternal education (n = 94), physician visits (n = 315), unmet needs (n = 357), or usual source of health care (n = 444) or if they had missing data for any other potential confounders (n = 789). These exclusions left 63,054 children to be included in final analyses.
Almost one out of five children had not visited a doctor in the past year, and this proportion did not vary by family structure (Table 1). The percentage with no usual source of health care varied only slightly by family structure (6 to 7 percent). Children of single mothers were somewhat more likely to have unmet health care needs than children in two-parent families (12 vs. 8 percent). Single mothers were more likely to have fewer than 12 years of education and less likely to have 16 or more years of education than married mothers. Age did not vary between children of single mothers and children in two-parent families. Children of single mothers were more likely to be non-Hispanic Black or Hispanic and less likely to be non-Hispanic White than children in two-parent families. Children of single mothers were more likely to be in only fair or poor health (5 vs. 2 percent) than children in two-parent families.
Children of single mothers were substantially less likely to have employer-sponsored health insurance (35 vs. 71 percent) and were more likely to have Medicaid coverage (47 vs. 10 percent) than children in two-parent families; however, rates of coverage by another type of health insurance were similar between the two groups, and approximately 14 percent of each group was uninsured. Single mothers were less likely than married mothers to be working. Children of single mothers lived predominantly in central cities of MSAs, whereas most children in two-parent families lived in an MSA but outside of the central city (in suburbs). Single-mother families were more likely than two-parent families to have a grandmother or other adult female relative in the household, and their family incomes were lower than those of two-parent families.
In addition to family structure, maternal education also played a role in health insurance (Table 2). Children in single-mother families were less likely than children in two-parent families to have employer-sponsored private insurance at each level of maternal education. Medicaid coverage was 2 to 10 times as common among children of single mothers as among children in two-parent families at each level of maternal education. The proportion of children who were uninsured declined steeply with increasing maternal education in two-parent families but declined only slightly in single-mother families.
Stratified and Adjusted Models
A significant interaction was found between maternal education and family structure (p < .001 for each access measure), and thus, models were stratified on four levels of maternal education. Results for bivariate stratified models (Table 3, model A) showed there was a socioeconomic gradient in access disparities by family structure for each measure of access to care. For all three measures, children of single mothers had fewer access problems than children in two-parent families at the lowest level of maternal education. At the highest level of maternal education, children of single mothers had more access problems (although the difference for usual source of care was not statistically significant).
When child's race, age, health status, and maternal employment were added into stratified models, the relationship between family structure and all three measures of health care access and utilization continued to differ significantly by maternal education (Table 3, model B). For each measure, the socioeconomic gradient in the relationship between family structure and access to care remained, but the strength of the socioeconomic gradient was reduced. Among lower-SES children, children of single mothers were more likely to have a physician visit in the past year and to have a usual source of care; unmet health care needs were similar in the two groups. In contrast, at the highest level of maternal education, there were no significant family structure differences in physician visits or usual source of care; disparities in unmet needs increased as maternal education increased. Among children whose mothers had 16 or more years of education, those in single-mother families were over three times as likely to have an unmet health care need as those in two-parent families.
When health insurance was added to these models, the differences in the relationships by SES disappeared for physician visits and usual source of care (Table 3, model C). Children of single mothers remained more likely to have had physician visits in the past year than children in two-parent families, with the exception of children whose mothers had 16 or more years of education; physician contacts did not vary by family structure in this group. Family structure was not associated with no usual source of health care at any level of maternal education. The socioeconomic gradient in unmet health care needs remained, although the gradient was somewhat less pronounced than for the previous model. When education levels were combined into a single model and an interaction term for family structure and maternal education was included, the interaction was no longer significant for physician visits or usual source of care but remained significant for unmet health care needs.
We hypothesized that health status and utilization would be worse among low SES, but not high SES, children of single mothers and the gap would be explained by access disparities by family structure at lower SES. This hypothesis was only partially sustained. Analysis of two of the access measures, no physician visits in the past year and no usual source of health care, found that family structure had no impact at the highest level of maternal education, as predicted. However, unmet health care needs did not follow the same pattern. Rather than decreasing as maternal education increased, the family structure disparity in unmet needs increased with maternal education, such that at high levels of maternal education, children of single mothers had substantially more unmet needs than children in two-parent families. After adjustment for health insurance coverage, there was no significant interaction between family structure and SES for physician visits or usual source of care, but a positive relationship remained for unmet needs.
The hypothesis that lower SES children of single mothers would have poorer access to care than lower SES children of married mothers was not found to be true. For physician visits and usual source of care, children of single mothers were not relatively worse off at low levels of SES; in fact, among children whose mothers had fewer than 16 years of education, children of single mothers were more likely than children in two-parent families to have seen a physician in the past year, and there was no difference by family structure in the adjusted odds of having a usual source of care. Using the most objective access measure, no physician visit in the past year, lower SES children in two-parent families appeared to have the most problems accessing services.
However, despite the seeming equal or superior access for children of single mothers within SES categories, single mothers reported their children to have more unmet needs for care, even after adjusting for the health status of the child and other confounders. This is similar to a previous finding (Angel and Worobey 1988): Single mothers appear to be more worried about their children's health care than are married mothers, even after controlling for variations in health status. The stresses associated with single parenting may help to explain this observation. Children of single mothers may have more health care needs, which may not have been fully accounted for in the health status measure used. In addition, the questions on unmet need included services such as eye and dental care, which may be less readily accessed than physician care, which could also help to explain why this measure behaves somewhat differently than the other two access measures.
A recent report from the Institute of Medicine states that "insurance coverage is the major determinant of whether children have access to health care" (Institute of Medicine 1998, p. 3). Our analysis found that the higher frequency of Medicaid coverage among children of single mothers was a major reason for children of less educated single mothers having fewer problems accessing care. The greater Medicaid coverage among single-mother families is likely a result of program requirements. Medicaid is an important part of the safety net in caring for children of single mothers, but children in two-parent families whose mothers were less educated did not always have access to that resource. In the next few years, the state CHIP legislation may reduce the family structure disparity in insurance coverage among lower SES children, as eligibility for insurance under the CHIP programs does not depend on the marital status of the child's parents. However, many states have thus far had disappointing results in enrollin g low-income families, and thus, CHIP may not entirely eliminate this problem.
This analysis had several limitations. The NHIS had relatively little information on family relationships and did not distinguish among biological, adoptive, and step parents. Data linkage was incomplete; parents could not be identified for approximately 1.5 percent of children who may have lived with a parent. The analysis used data from 1993-95, and health insurance patterns may have changed since then. Maternal education did not represent all aspects of SES. This analysis also had several strengths. The NHIS includes a nationally representative, large sample of parents and children in a variety of social and economic circumstances. This analysis was able to link parents with children, providing information not available in many other analyses. The measures of access to health care captured a range of dimensions. Information about health insurance coverage was detailed.
Among children of low SES, those with single mothers may actually be advantaged compared with those in two-parent families in obtaining health insurance and health care. Single motherhood can be beneficial for receipt of social services. This disparity is an unintended result of the AFDC regulations, which were originally enacted to assist widows and their families; more recently, the new Temporary Assistance to Needy Families (TANF) programs in some states have tried to counteract this effect. Medicaid rolls have dropped dramatically since TANF was enacted, despite the delinkage of Medicaid from TANF; thus, the greater coverage we observed among low-income children of single mothers may have been reduced over the past few years.
Although many researchers have expressed concerns regarding the well-being of children of single mothers, this analysis unexpectedly found that children in low-SES, two-parent families, rather than children of single mothers, may be most vulnerable to problems of health care access and utilization. The reason for this finding appears to be that health insurance coverage, particularly Medicaid, is more common among children of single mothers than among two-parent families at low levels of maternal education. The employers of low-SES parents provide health insurance to only approximately one third of these children, and thus, many are uninsured. Because private health insurance coverage has declined over time, expansion of public insurance coverage, particularly to children in low-SES two-parent families, is critical in ensuring adequate access to health care for children.
Table 1: Characteristics of Children in Single-Mother and Two-Parent Families (National Health Interview Survey, 1993-95) Single Mother Two Parent n = 14,703 n = 48,351 Percent SE Percent Physician visit in past year 82.7 0.4 82.1 Has a usual source of care 93.0 0.3 94.3 Unmet health care need (*) 12.1 0.4 8.1 Maternal education 0 to 11 years 27.1 0.7 13.9 12 years 40.8 0.7 38.5 13 to 15 years 22.5 0.6 24.5 16 or more years 9.6 0.4 23.1 Child's age 0 to 4 28.5 0.5 29.2 5 to 9 28.0 0.4 28.1 10 to 17 43.5 0.5 42.6 Race ethnicity Hispanic 18.2 0.6 13.6 Black, non-Hispanic 38.0 0.8 8.3 White, non-Hispanic 43.8 0.8 78.1 Child's health status Good to excellent 95.2 0.2 97.9 Fair to poor 4.8 0.2 2.1 Health insurance coverage Employer sponsored 34.8 0.7 71.2 Medicaid 46.7 0.7 9.8 Other coverage 4.2 0.3 5.6 Uninsured 14.3 0.5 13.4 Maternal employment Working 56.5 0.7 63.6 Unemployed 7.0 0.4 3.0 Not in labor force 36.5 0.7 33.4 Family income Less than $20,000 66.8 0.7 17.6 $20,000 or more 31.7 0.7 80.9 Two Parent n = 48,351 p value for SE difference Physician visit in past year 0.3 .242 Has a usual source of care 0.2 .001 Unmet health care need (*) 0.2 .001 Maternal education 0 to 11 years 0.3 .001 12 years 0.4 13 to 15 years 0.3 16 or more years 0.4 Child's age 0 to 4 0.3 .311 5 to 9 0.2 10 to 17 0.3 Race ethnicity Hispanic 0.3 .001 Black, non-Hispanic 0.3 White, non-Hispanic 0.4 Child's health status Good to excellent 0.1 .001 Fair to poor 0.1 Health insurance coverage Employer sponsored 0.4 .001 Medicaid 0.3 Other coverage 0.2 Uninsured 0.3 Maternal employment Working 0.4 .001 Unemployed 0.1 Not in labor force 0.4 Family income Less than $20,000 0.3 .001 $20,000 or more 0.4 (*)Unmet needs were defined as needing medical care, dental care, prescription medicine, eyeglasses, or mental health services but being unable to get it or delaying medical care because of cost. Table 2: Percent Distribution of Health Insurance Coverage by Family Structure and Maternal Education, National Health Interview Survey, 1993--95 Maternal Employer Education Family Sponsored Medicaid Other (Years) Structure Private Coverage Coverage Uninsured 0-11 Single mother 10.2 72.4 1.5 15.9 Two parent 36.2 30.5 2.5 30.7 12 Single mother 34.0 47.1 4.0 15.0 Two parent 69.1 10.4 5.6 14.9 13-15 Single mother 48.5 33.9 5.8 11.9 Two parent 78.3 5.5 6.5 9.7 16+ Single mother 69.5 10.7 8.8 11.0 Two parent 87.3 1.5 6.5 4.8 Maternal Education (Years) p Value 0-11 0.001 12 0.001 13-15 0.001 16+ 0.001 Table 3: Models for Health Care Access and Utilization (National Health Interview Survey, 1993-95) Access Measure No Physician Mother's Family Visit Education Structure OR (*) CI (*) Model A (**) 0-11 years Single mother 0.68 (0.60-0.77) Two parents 1.00 12 years Single mother 0.83 (0.75-0.92) Two parents 1.00 13-15 years Single mother 0.92 (0.80-1.07) Two parents 1.00 16+ years Single mother 1.39 (1.14-1.68) Two parents 1.00 Model B (**) 0-11 years Single mother 0.69 (0.60-0.79) Two parents 1.00 12 years Single mother 0.81 (0.72-0.91) Two parents 1.00 13-15 years Single mother 0.82 (0.70-0.95) Two parents 1.00 16+ years Single mother 1.02 (0.82-1.26) Two parents 1.00 Model C (**) 0-11 years Single mother 0.83 (0.71-0.96) Two parents 1.00 12 years Single mother 0.84 (0.74-0.95) Two parents 1.00 13-15 years Single mother 0.83 (0.71-0.97) Two parents 1.00 16+ years Single mother 0.91 (0.73-1.14) Two parents 1.00 Access Measure No Usual Unmet Health Mother's Source of Care Care Needs Education OR CI OR CI Model A (**) 0-11 years 0.66 (0.54-0.81) 0.84 (0.70-1.00) 1.00 1.00 12 years 1.26 (1.04-1.53) 1.32 (1.15-1.51) 1.00 1.00 13-15 years 1.18 (0.90-1.54) 1.97 (1.65-2.37) 1.00 1.00 16+ years 1.43 (0.97-2.12) 3.50 (2.71-4.52) 1.00 1.00 Model B (**) 0-11 years 0.68 (0.54-0.86) 0.99 (0.82-1.20) 1.00 1.00 12 years 1.18 (0.96-1.45) 1.49 (1.27-1.75) 1.00 1.00 13-15 years 1.10 (0.82-1.48) 1.99 (1.64-2.42) 1.00 1.00 16+ years 1.27 (0.86-1.89) 3.05 (2.33-4.01) 1.00 1.00 Model C (**) 0-11 years 0.85 (0.65-1.11) 1.19 (0.97-1.47) 1.00 1.00 12 years 1.07 (0.87-1.31) 1.44 (1.21-1.71) 1.00 1.00 13-15 years 0.98 (0.70-1.38) 1.87 (1.50-2.31) 1.00 1.00 16+ years 0.77 (0.47-1.26) 2.59 (1.95-3.43) 1.00 1.00 (*)Odds ratio and 95% confidence interval. (**)Model A, unadjusted. Model B, model A + race -ethnicity + maternal employment + child's age + child's health status. Model C, model B + health insurance.
Adams, P. F., and M. A. Marano. 1995. "Current Estimates from the National Health Interview Survey, 1994. National Center for Health Statistics" Vital Health Statistics 10 (193).
Angel, R., and J. L. Worobey. 1988. "Single Motherhood and Children's Health." Journal of Health and Social Behavior 29: 38-52.
Bureau of the Census. 2000. Income Table FINC-03. Presence of Related Children Under 18 Years Old--All Families, by Total Money Income in 1999, Type of Family Work Experience in 1999, Race and Hispanic Origin of Reference Person. Suitland, MD: Bureau of the Census.
Bureau of the Census. 2001. Table CH-1. Living Arrangements of Children Under 18 Years Old: 1960 to Present. http://www.census.gov/population/socdemo/hh-fam/tabCH-1.txt. Suitland, MD: Bureau of the Census.
Cafferata, G. L., and J. D. Kasper. 1985. "Family Structure and Children's Use of Ambulatory Physician Services." Medical Care 23: 350-60.
Cunningham, P. J., and B. A. Hahn. 1994. "The Changing American Family: Implications for Children's Health Insurance Coverage and the Use of Ambulatory Care Services." The Future of Children: Critical Health Issues for Children and Youth 4 (3): 24-42.
Federal Interagency Forum on Child and Family Statistics. 1998. America's Children Key National Indicators of Well-Being, 1998. Federal Interagency Forum on Child and Family Statistics, Washington, DC: Government Printing Office.
Institute of Medicine, Committee on Children, Health Insurance, and Access to Care. America's Children: Health Insurance and Access to Care. Washington, DC: National Academy Press, 1998.
Massey, J. T., T. F. Moore, V. L. Parsons, and W. Tadros. 1989. "Design and Estimation for the National Health Interview Survey, 1985-94." National Center for Health Statistics. Vital Health Statistics 13 (121).
Meyer, D. R. 1997. "Health Insurance and Child Support." Health Affairs 16: 207-17. Newacheck, P. W. 1992. "Characteristics of Children with High and Low Usage of Physician Services." Medical Care 30: 30-42.
Shah, B. V., B. C. Barnwell, and C. S. Bieler. 1995. SLJDAAN Users Manual: Software for Analysis of Correlated Data (release 6.40). Research Triangle Park, NC: Research Triangle Institute.
Simpson, G., B. Bloom, R. A. Cohen, and P. E. Parsons. 1997. "Access to Health Care. I. Children". National Center for Health Statistics. Vital Health Statistics 10 (196).
Weinick, R. M., and A. C. Monheit. 1999. "Children's Health Insurance Coverage and Family Structure, 1977-96." Medical Care Research and Review 56: 55-73.
Weissman, J. S., and A. M. Epstein. 1994. Falling Through the Safety Net. Baltimore, MD: The Johns Hopkins University Press.
Address correspondence to Katherine Heck, M.P.H., Health Statistician, Infant and Child Health Studies Branch, Office of Analysis, Epidemiology, and Health Promotion, National Center for Health Statistics, 714 P Street, Room 476, Sacramento, CA 95814. This article, submitted to Health Services Research on October 11, 1999, was revised and accepted for publication on September 6, 2000.
|Printer friendly Cite/link Email Feedback|
|Author:||Heck, Katherine E.; Parker, Jennifer D.|
|Publication:||Health Services Research|
|Date:||Feb 1, 2002|
|Previous Article:||Helpful or harmful? The impact of strategic change on the performance of U.S. urban hospitals.|
|Next Article:||Development of a scale to measure patients' trust in health insurers. (Methods).|