Center for Studying Health System Change

Providing Insights that Contribute to Better Health Policy

Advanced Search Instructions

You can refine your search with the following modifiers:

* Use an to perform a wildcard search.Example: prescript* would return "prescription", "prescriptions" etc.
"" Use quotes to match a phrase.Example: "prescription drug" only returns results where the words are next to each other.
+ Use a plus sign to perform a search where the additional term MUST be part of the page.Example: prescription +drug
- Use a minus sign to perform a search where the additional term SHOULD NOT be part of the page.Example: prescription -drug
< > Use a < > sign to perform a search where the additional term should be of greater or lesser importance in the search.Example: prescription >drug
Find pages with the word precription with additional importance for the word drug.
( ) Use parentheses to group different search terms together.Example: prescription (+medicare -drug)

Insurance Coverage & Costs Costs The Uninsured Private Coverage Employer Sponsored Individual Public Coverage Medicare Medicaid and SCHIP Access to Care Quality & Care Delivery Health Care Markets Issue Briefs Data Bulletins Research Briefs Policy Analyses Community Reports Journal Articles Other Publications Surveys Site Visits Design and Methods Data Files

Printable Version

Implications for Tailoring CHIP Programs to Local Conditions.

s indicated earlier, we find a fivefold difference in the rates of uninsurance among children in twelve randomly selected metropolitan areas. We conducted some multivariate analyses to explain the likelihood that a child will go without health insurance. Explanatory variables included characteristics of the child and his family: family income, number of children, age of mother, age of child, race/ethnicity, whether the interview was conducted in Spanish, mother’s education level, number of parents working, whether a parent has an offer of insurance through their employer, and whether the child is eligible for Medicaid. Applying the results of this equation, along with variations in the characteristics of children and their families across the 12 metropolitan areas allows us to explain about half of the observed variation between the metropolitan areas. The four factors that consistently explain the greatest cross-site variation, in descending levels of importance, include: 19

  • The percentage of parents with an offer of insurance from their employer

  • the percentage of children from Spanish speaking families

  • The percentage of children eligible for Medicaid (controlling for income levels)

  • The distribution of income among families with children.

These four factors clearly relate to the four reasons children go without health insurance. Offer rates by employers are clearly related to the first reason, that parents lack access to employer sponsored insurance. For example, only 54% of children in Miami have a parent who is offered health insurance through an employer. This compares with 73% among children nationally.20 If offer rates in Miami increased to equal the national average, the uninsurance among children would decline by four percentage points. Congress clearly decided to pursue a strategy of publicly funded insurance over market reforms or regulatory approaches that would entice or require employers to offer health benefits. Whether this approach alone is sufficient to address the problem of uninsured children remains to be seen.

Although the percentage of children from Spanish speaking homes may be related to the percentage of children who are eligible for public benefits, it is most likely related to problems getting Medicaid-eligible Hispanic children to enroll. It is important to note that it is the language barrier, rather than being of Hispanic heritage, that predominantly affects uninsurance rates among children. Higher than average percentages of Spanish speaking families contribute to the high uninsurance rates among children in Miami and Orange County, while lower than average rates in Indianapolis, Little Rock, Syracuse, and Lansing lead to lower uninsurance among children. Although outreach efforts need to be directed to all uninsured children who are eligible for public benefits, this finding illustrates the vital role that outreach efforts will need to play if CHIP is to realize its full potential. More specifically, it highlights the importance of multi-lingualism and cultural sensitivity in these efforts.

The fact that Medicaid eligibility affects uninsurance rates is gratifying in that it suggests that expansions in eligibility for publicly subsidized insurance resulting from CHIP will have an impact. However, at the margin, we find that after controlling for other factors, being eligible for Medicaid decreases the likelihood of a child being uninsured by only 15%. Prior to the enactment of CHIP, variations in state Medicaid eligibility rules for children accounted for as much as a two percentage point variation in the overall uninsurance rate among children in the 12 sites. This suggests that states will need to work hard at informing targeted families about these new programs and at designing the Medicaid and Title XXI programs so that they will be attractive to these families.

Finally, we find that the income distribution among lower income families with children in cities, particularly the percentage living in poverty, explains a significant portion of the variation in local uninsurance rates among children. Children in Seattle, for instance, tend to come from higher income families than the national average. This accounts for over two percentage points in Seattle’s relatively low rate of uninsurance among children. In contrast, greater concentrations of poor children in Little Rock and Miami account for some of their relatively high rates of uninsurance among children. Since these estimates control for local variations in Medicaid eligibility and employer offer rates, an explanation is that lower income persons are less likely to avail themselves of private insurance when it is available to them. This again argues for limiting cost sharing requirements to the poor and near poor.

19 The multivariate equation results are available upon request from the Center for Studying Health System Change. For each explanatory variable, regression coefficients were multiplied with the difference between the site and national mean of that explanatory variable. This allows us to decompose the explained variation in uninsurance rates among children among the 12 metropolitan areas.

20 Since the 12 metropolitan areas are drawn from sites with populations over 200,000, the national mean also reflects the population living in metropolitan areas with populations over 200,000.

Table Of Contents | Next Section

Back to Top