Association Between Underinsurance and Access to Care Among Children With Special Health Care Needs in the United States

METHODS

The National Survey of Children With Special Health Care Needs was conducted by the MCHB in coordination with the National Center for Health Statistics of the Centers for Disease Control and Prevention. The survey was a list-assisted, random-digit-dial telephone survey that was built onto the existing State and Local Area Integrated Telephone Survey, which is conducted by the National Center for Health Statistics. The survey was administered to a representative sample of households in each of the 50 states and the District of Columbia to assess the prevalence and the impact of special health care needs among children at the state and national levels. The survey questionnaire was developed to provide a comprehensive array of data on the child's health and functional status and the child's and the family's experiences with the health care system.

The National Survey of Children With Special Health Care Needs was conducted in 3 parts. First, selected demographic information was collected for all children who were younger than 18 years. Next, the CSHCN Screener, developed under the auspices of the Foundation for Accountability, was administered “family style” to identify whether any of the children in the household had an existing special health care need.²³ When a child was identified as having a special health care need through the screening questions, a comprehensive battery of questions then was administered. When >1 child in a household was identified as having a special need, 1 of those children was randomly selected to be the subject of the comprehensive interview.

CSHCN were identified from a set of screening questions that were developed using the MCHB definition as the underlying conceptual framework.²⁴ From ∼200000 households with children, 373055 children were screened for special health care needs, and 38866 interviews were completed regarding children who screened in for having special health care needs as indicated by an increased need for services beyond that of other children or a functional limitation. CSHCN experience a variety of physical and mental health–related conditions. Although no condition-specific data were collected in the National Survey of Children With Special Health Care Needs, other representative national data indicate that asthma, attention-deficit/hyperactivity disorder, developmental delays, impairments that limit mobility, vision problems, and impairments that require special equipment are the most commonly reported conditions among CSHCN.²⁵ Among all US children, the overall prevalence of CSHCN was estimated to be 12.8%.²⁶

The respondent for the survey was the person who was most knowledgeable about the child's health and health care. In >96% of the interviews, that was a parent or legal guardian.²⁷ The survey was translated into 11 different languages for non-English speakers. Languages other than English accounted for 5.5% of the household screening interviews.²⁷ The interview completion rate for age-eligible households was 76.2%.²⁷ The detailed survey method and questionnaire are available on-line.²⁸

Level of insurance adequacy was derived from a series of 3 questions. Adequacy was assessed among children who had health insurance by the following questions: (1) Did the benefits meet the child's needs? (2) Were the costs not covered by insurance reasonable? (3) Did the health insurance plan allow the child to see the health care providers they needed? A summary measure of adequacy was drawn by combining these 3 variables. When a respondent answered “always” or “usually” to all 3 measures of adequacy, the child was considered to have adequate insurance coverage. When a respondent answered “sometimes” or “never” to any of the 3 questions, the child was considered to have inadequate coverage. We defined “underinsured” CSHCN as those who had inadequate coverage. We also examined continuity of insurance coverage, as it may also affect access and other health outcomes. Continuity was assessed through the following: (1) whether the child was covered by health insurance at the time of the survey and (2) if not, whether the child was covered by health insurance at any time in the 12 months before the interview. These variables were combined into a 4-level variable: CSHCN who were never insured during the past year, CSHCN who had periods without insurance in the last year, CSHCN who had continuous but inadequate coverage (the underinsured), and children who were covered continuously with adequate coverage (Fig 1). For each level of the variable described above, we examined 7 outcomes that measured different aspects of access to care and impact on the family: (1) whether the child had delayed or forgone care in the last year, (2) whether the child had any unmet need for specific health care services, (3) whether the child had a usual source of care, (4) whether the child had a personal doctor or nurse, (5) whether the family had difficulty receiving referrals for needed care for the child, (6) whether the child's health condition caused financial problems for the family. and (7) whether the child's condition affected the employment of any family members.

• Download figure
• Open in new tab
• Download powerpoint

Fig 1.

Estimates of underinsurance among CSHCN in the United States, 2001.

The data presentation is composed of 5 sections. The first part of the analysis shows the distribution and population estimates on measures of insurance coverage, adequacy of insurance, access to care, and impact on the family among CSHCN. The second part of the analysis examines the bivariate associations between sociodemographic and health factors (child's age, gender, race/ethnicity, poverty status, region of the country,²⁹ functional ability, type of special health care need, and number of special health care needs) and the 4-level insurance variable. Functional ability was measured by the frequency and the degree to which the child's condition affected his or her ability to do the things done by other children of the same age. Type of special health care need included 5 major components: use of or need for prescription medicines; use of or need for more medical care, mental health services, or education services than other children of the same age; use of or need for treatment or counseling for an emotional, developmental, or behavioral problem; limitation in the child's ability to do the things that most children of the same age do; and use of or need for special therapy, such as physical, occupational, or speech therapy. Number of special health care needs was used as a measure of severity and was categorized as 1 versus >1.

In the third part, a weighted logistic regression model was used to examine the factors associated with underinsurance. For this analysis, children without insurance and children with discontinuous coverage were excluded from the analysis. Therefore, children who had continuous coverage but were underinsured were compared with children with continuous and adequate coverage. After a bivariate examination of the association between the measures of access and underinsurance, the final part of the analysis uses a series of weighted logistic regression models to examine the effect of underinsurance on each of the 7 outcome measures. All models were adjusted for child's age, gender, race/ethnicity, family poverty level, region of the country, functional ability, type of special health care need, and number of needs. The survey questions that were used in deriving these outcome measures and explanatory variables are available at www.cdc.gov/nchs/about/major/slaits/cshcn.htm.

The data were collected through a complex, multistage sample design. Therefore, all analyses used sampling weights based on the 2000 Census and the 1999–2001 Current Population Survey to adjust for nonresponse bias and other discrepancies between the sampled children and known characteristics of the US noninstitutionalized population of children. The weighted data are representative of the noninstitutionalized population of children generally and of CSHCN specifically.²⁷ The analyses were conducted using the Survey Data Analysis software program (SUDAAN) to calculate accurate variances from complex sample surveys.³⁰ Adjusted odds ratios and 95% confidence intervals were computed by using the β coefficients and standard errors obtained from the logistic regressions. All differences described in the text are significant at the .05 level unless otherwise specified.

RESULTS

Figure 1 indicates that in this nationally representative survey of CSHCN, underinsurance is far more prevalent than having no insurance at all. Although the percentage of CSHCN who had no insurance was 5.2% and an additional 6.4% had a period with no insurance in the last year, approximately one third (32.3%) of the CSHCN were underinsured. The remaining 60% of CSHCN were considered to have both continuous and adequate coverage.

Table 1 provides measures of adequacy of insurance and access to care for children with continuous coverage. Even among the continuously insured, there were access and impact issues: 14% reported an unmet need for a specific care service, 6% reported that their child had delayed or forgone care, 20% had difficulty getting referrals for needed care, and 28% reported that the child's condition had affected the employment of at least 1 family member.

The correlates of underinsurance among CSHCN are presented in Table 2. Hispanics were more likely to be underinsured and have discontinuous coverage. Younger children were more likely to have discontinuous coverage, but there were no age differences for underinsurance. There was also a poverty gradient, with those most in poverty at the highest risk for being both underinsured and covered discontinuously. Families in which the child was most limited in their activities of daily living were more likely to report being underinsured, as were families in which the child needed physical, occupational, or speech therapy or the child was treated for an emotional, developmental, or behavioral condition.

The results of the weighted logistic regression that examined the factors that were associated with underinsurance are presented in Table 3 (CSHCN who were uninsured or discontinuously insured during the last year were not included in this analysis). We found results similar to the bivariate analysis: the odds of being underinsured were significantly higher for Hispanic individuals; those most in poverty; children who are most limited in their activities; children who needed physical therapy, and children who were treated for an emotional, developmental, or behavioral condition. In addition, the odds that CSHCN in the South were underinsured were significantly higher compared with CSHCN in the Northeast.

The bivariate associations of the measures of access and family impact with the level of insurance coverage then were examined (Table 4). Families of CSHCN who were underinsured were significantly more likely to report problems with access and have adverse family impacts for all outcomes, except for usual source of care and having a personal physician. The differentials were particularly striking for the likelihood of having delayed care, unmet needs for health care services, difficulties receiving referrals for specialty care, and financial problems for the family. It is interesting that the effect of both being underinsured and having discontinuous care were somewhat similar, except for delayed care and having an unmet need, which were more prevalent among CSHCN who had discontinuous coverage.

Table 5 presents the results of the multivariate analysis for the association between the 7 outcomes variables and level of insurance adjusted for sociodemographic and health factors. Uninsured CSHCN were generally at the highest risk for most access and family impact problems. However, CSHCN who were underinsured were at increased risk for every outcome, except for usual source of care and having a personal doctor or nurse. Congruent with the bivariate analysis, discontinuous coverage and underinsurance had similar effects on the 7 outcome measures, except for delayed and unmet needs, for which the effect was much stronger for discontinuous care.

DISCUSSION

Data from the nearly 40000 children in the National Survey of Children With Special Health Care Needs were used to examine 3 questions: (1) What is the extent of underinsurance in this population? (2) What are the characteristics of underinsured CSHCN? (3) What are the consequences of being underinsured? To our knowledge, this is the first nationally representative study that has described the impact of underinsurance on CSHCN.

We found that approximately one third of CSHCN were underinsured. The most commonly reported component of inadequate coverage was that the costs that were not covered by insurance were not considered reasonable. This is not surprising, because health care costs for CSHCN are much greater than for typically developing children.^18,31 However, significant proportions of the families of this population also reported that health insurance benefits did not meet the child's needs and that the coverage did not allow them to see needed providers. Our findings here are consistent with those of Davidoff,³² who noted that inadequate coverage for CSHCN may be most apparent for specialized services that are of particular importance to CSHCN.

We also found that >6% of CSHCN were insured at the time of the interview but had a period without health insurance in the past year. This study indicates that focusing only on insurance status at the time of the interview grossly underestimates the proportion of children who have less than complete coverage. Tang et al³³ found that ∼14% of all children were uninsured for part of the year preceding their study. Indeed, our study shows that the number of underinsured CSHCN is ∼6 times greater than the number of uninsured during the year. This sizable population of underinsured children is rarely discussed in the literature. Consequently, they represent the “hidden uninsured” of America's children.

The prevalence of underinsurance varied across different groups. Hispanic children, children living in poverty, and children who were most affected by their conditions were most likely to be underinsured, having higher rates of inadequate coverage than their counterparts. Adjustment for other factors attenuated the disparities by poverty level slightly. Previous research found that Hispanic individuals were also more likely to be uninsured among CSHCN and have periods without health insurance in the overall child population.^9,25,34 Being uninsured has also been associated with higher rates of poverty among CSHCN.²⁵ This study's results extend these earlier findings to additional risks, specific to CSHCN, such as the ability to gain access to specialty care.³⁵

A number of studies have examined access to care for CSHCN.^{32,34,36–39} We examined 5 indicators of access because access both is a multidimensional concept and certain aspects, such as the ability to obtain referrals for needed specialty care, are especially important for this population. Our findings were striking in their consistency. Children who were underinsured were generally at intermediate risk between CSHCN who never were insured and those who had continuous and adequate coverage. The 1 exception to this pattern was difficulty obtaining referrals for needed specialty care, a critical component for CSHCN, for which the underinsured were actually at higher risk than the uninsured. This may be because families of uninsured children may not even attempt to seek specialty care. Other research has shown that uninsured CSHCN are less likely to have access to care³⁷ and more likely to have unmet needs for routine and specialty care.^40,41 Our study indicates that there are gradations of risk for gaining access to care, even among those who are insured continuously.

We examined 2 measures of impact on the family: whether the child's health care caused financial problems for the family and whether the child's condition caused family members to reduce or stop work. The families of underinsured CSHCN were almost as likely to report financial hardships as families with uninsured children. It is interesting that families of underinsured children were more likely to report that the child's condition caused them to reduce or stop work. It could also be that “adequate” insurance plans offer more ancillary support and, thus, reduce the need to alter work schedules. The families of children who had discontinuous coverage were also more likely to be affected. This could be reflective of a parent's need to quit work or work part time to care for their child, thereby increasing their chances of having a gap in coverage. Overall, 21% of families reported financial burdens. This is consistent with studies on the adult population, in which even among the insured, 20% of Americans had problems paying medical bills.^11,42

There are a number of strengths to our study, including its large, nationally representative sample, its broad representation of children from different segments of US society, and the detail of questions on adequacy of insurance and various measures of access to care. However, the study design is also subject to limitations. First, because the study was cross-sectional, causal directionality is a possible concern, particularly for outcomes such as the impact on parental employment. Second, because insurance status was asked about over the last year, recall bias is a possibility. Third, we did not include type of insurance in the analysis because we could not obtain information on whether a family changed insurance coverage during the year. Fourth, we could not differentiate insurance continuity from personal physician continuity. Fifth, we could not examine the effect of underinsurance on specific conditions, although our findings suggest that it is the severity of the condition that is important. Finally, there is no single agreed-on definition of “underinsurance.” Alternative conceptions of underinsurance might yield different results from those presented here.

In addition, the survey data set that was used in this analysis did not include a referent sample of children without special health care needs. The next iteration of this survey, planned for 2005–2006, will include a referent population. Thus, it will be possible to examine to what extent the prevalence and the impact of underinsurance differs between children with and without special health needs.

Policy discussions on insurance often focus on mechanisms to extend insurance coverage to those with no coverage. This study indicates that such discussions are incomplete and ignore a large segment of children who are underinsured. In many respects, these children experience access problems that are similar in magnitude to the uninsured. Other researchers have cautioned that underinsurance may be a particular concern for CSHCN on the basis of documentation of higher rates of supplemental insurance purchased by families to provide “wrap-around” services for this population,^32,43 as well as through documentation of the inadequacy of appropriate depth and breadth of covered services for privately insured CSHCN.^12,13 The present work serves to validate those concerns and suggests that underinsurance is actually a much larger problem than absence of insurance.