The Alabama ALL Kids Evaluation Study is an ongoing multiyear study of new enrollees and continuous enrollees in Alabama’s State Children’s Health Insurance Program (SCHIP). The findings reported in these analyses resulted from a statewide study of children who were enrolled in ALL Kids, the separate, free-standing portion of Alabama’s SCHIP (ie, the non-Medicaid expansion portion) for the first time in each of its first 3 years (1998–2001).
Alabama implemented ALL Kids, the separate, free-standing portion of its SCHIP, in October 1998. ALL Kids covered otherwise-uninsured children <19 years old living in families with incomes <200% of the federal poverty level (FPL) and above the age-related Medicaid income thresholds: 133% of the FPL for children <6 years old and 100% of the FPL for children 6 to 18 years old. In an effort to curb substitution of ALL Kids for private coverage, children were not eligible to enroll in ALL Kids for 3 months after dropping private insurance coverage. Administrative checks were conducted on previous and current insurance coverage. Premiums were charged to families with incomes >150% FPL with children in ALL Kids at the rate of $50 per child per year, with a family maximum of $150 per year. Copayments, $5 for nonpreventive visits and $1 for generic drugs, were charged.
The reference time frame for questions about previous experience with insurance and health care was any time before SCHIP enrollment for each of the 3 cohorts. To compare enrollee groups by previous experience with the health care system at any time before SCHIP, we examined commonly used measures of access to care and unmet need (needed but could not get medical, dental, vision, or specialty care or prescriptions) and having a usual source of care.
Family income, race, age of child, education level of respondent, and an urban/rural measure were chosen to describe characteristics of the sample for each year. By using data from administrative SCHIP records, family income was categorized as 100% to 150% of the FPL (the no-fee group) and 151% to 200% of the FPL (the fee-paying group). Race was a self-declared classification of white, African American, or other, because Alabama has few minority children that are not African American (<3%). The age of the child, education level of the respondent, and perception of the child’s overall health were also obtained from self-reported survey data. Approval for this study was obtained from the Institutional Review Board for Human Subjects at the site of the evaluation study.
SAMPLE DESIGN AND SURVEY ADMINISTRATION
The data for analyses were collected through 3 mail surveys and administrative records. The same survey questions were asked each year, and the same procedures were followed. The first-year (October 1998 through September 1999) survey was a retrospective assessment of a random sample of all first-year ALL Kids’ enrollees. Every fourth enrollee on a randomly arranged list was sampled, resulting in a sample of 6200 of the 25 748 first-year enrollees.
The second- and third-year surveys (October 1999 through September 2000) were sent to families of all newly enrolled children during the first quarter of their enrollment period. In the second year of ALL Kids (1999–2000), there were 11 075 new enrollees and a total of 29 064 enrollees including new and continuous enrollees. In year 3 (2000–2001), 14 776 children were newly enrolled, and there were a total of 39 240 new and continuous enrollees for the year. The surveys focused on one single child’s health care experience per family. Thus, when there was more than one child in the family participating in SCHIP, one newly enrolled child was chosen randomly from the family as the target child for that interview (year 2: 7258 sent; year 3: 9658 sent).
If surveys were not returned, the standard follow-up methods were used: 1) mailing a reminder postcard; 2) sending the survey 2 more times; and 3) attempting to conduct the survey by telephone. For the first-year survey, the return rate was 60% (n = 3740). For the second and third years, the return rates were 63% (n = 3219) and 54% (n = 4110), respectively. A comparison was made of the responders and nonresponders in the first year survey based on the administrative information and a brief pediatric health history gathered when the child enrolled. No significant differences were found in responders versus nonresponders.
For descriptive analyses, we used t tests to evaluate differences in continuous variables across subgroups and χ2 tests for dichotomous and categorical variables. Multivariate relationships were evaluated by using logistic and multinomial logistic regression methods. All analyses were conducted by using SPSS 10.0 (SPSS, Inc, Chicago, IL).