|Home | About | Journals | Submit | Contact Us | Français|
Gaps in health insurance coverage compromise access to health care services, but it is unclear whether the length of time without coverage is an important factor. This article examines how coverage gaps of different lengths affect access to health care among low-income children.
We conducted a multivariable, cross-sectional analysis of statewide primary data from families in Oregon’s food stamp population with children presumed eligible for publicly funded health insurance. The key independent variable was length of a child’s insurance coverage gap; outcome variables were 6 measures of health care access.
More than 25% of children reported a coverage gap during the 12-month study period. Children most likely to have a gap were older, Hispanic, lived in households earning between 133% and 185% of the federal poverty level, and/or had an employed parent. After adjusting for these characteristics, in comparison with continuously insured children, a child with a gap of any length had a higher likelihood of unmet medical, prescription, and dental needs; no usual source of care; no doctor visits in the past year; and delayed urgent care. When comparing coverage gaps, children without coverage for longer than 6 months had a higher likelihood of unmet needs compared with children with a gap shorter than 6 months. In some cases, children with gaps longer than 6 months were similar to, or worse off than, children who had never been insured.
State policies should be designed to minimize gaps in public health insurance coverage in order to ensure children’s continuous access to necessary services.
Point-in-time estimates reveal that approximately 9 million children are without health insurance. When including children with gaps in coverage during the year, the number of uninsured children almost doubles.1–4 Although many uninsured children are eligible for public insurance programs, children are more likely than adults to have episodes without coverage.5–7 Therefore, it is crucial to study how these health insurance coverage gaps affect access to health care, especially among children eligible for Medicaid and the State Children’s Health Insurance Program (SCHIP).7
Current reports are mixed about how insurance gaps affect children’s access to care. Gaps are associated with discontinuities in receipt of recommended primary care and a higher likelihood of delayed medical care.8–12 In contrast, some evidence suggests that insurance gaps do not predict worsening of specific outcome markers, such as emergency department utilization rates or hospitalization for asthma.2,13 Less is known about how the duration of time without coverage affects health care access for low-income children.
Establishing the significance of the length of time without insurance among children eligible for public coverage has important policy implications. Oregon is a key state to highlight in this endeavor because Oregon has a higher proportion of children with coverage gaps compared with other larger, more populous states for many reasons.4 First, most states (36 in July 2006)14 require a waiting period after loss of private insurance before a child can receive SCHIP coverage. These mandatory regulations were implemented to discourage parents from migrating to public coverage unnecessarily (“crowd-out”). In July 2006, Oregon was among 16 states with a 6-month uninsured waiting period; only 1 state had a longer period.14 Oregon also required SCHIP renewal applications every 6 months, whereas most states had set renewal periods at 12-month intervals.14 In this article, we explore associations between duration of coverage gaps and differential reports of children’s access to health care among a subsample of Oregon families.
We identified all families enrolled in Oregon’s food stamp program at the end of January 2005. At that time, eligibility criteria for food stamps in Oregon required a household income of less than 185% of the federal poverty level (FPL) and US citizenship. These income and citizenship requirements were essentially the same for children’s Medicaid or SCHIP; therefore, for the purposes of this study, Oregon children receiving food stamps were presumed eligible for publicly funded health insurance.
We selected a representative sample of 10 175 from a total of 84 087 households with at least one child older than 1 year. (Families with only children who were younger than 1 year were excluded because infants younger than 1 year old have different eligibility requirements.) The stratified, random sample was divided evenly between families with at least one publicly insured child and those with no children enrolled in Oregon’s Medical Assistance Programs during the preceding 2 months. We used oversampling techniques to augment the sample in rural areas and then randomly selected a focal child from each household. A total of 8636 households were eligible to participate. (We excluded families who had moved out of state and those with no current address.)
We designed the survey instrument to collect information about the focal child’s insurance status and access to health care services during the previous year. Most survey questions were adapted from widely accepted national surveys.15–18 For validity, cognitive interviews were conducted with low-income parents. We had the surveys translated into Spanish and Russian, and then independently backtranslated to ensure fidelity of translation. The instrument was a self-report, mail return survey containing 63 items. We used a 4-wave methodology—2 surveys and 2 reminder postcards—over 6 weeks. Because of budgetary constraints, telephone follow-up was not possible. The Oregon Health and Science University Institutional Review Board approved all aspects of the study (OHSU eIRB 1717).
Six outcome variables pertained to compromised health care access, including the following: unmet health care needs in the last 12 months (medical care not received, prescriptions not refilled because of their high cost, and big problems getting dental care); no ambulatory visits in 12 months; no usual source of care; and delayed urgent care (rarely or never received immediate care as soon as it was needed).
We used 3 survey questions to construct the key health insurance coverage gap independent variable: (1) About how long has it been since YOUR CHILD last had health insurance coverage?” (included option for “My child currently has health insurance”); (2) “In the last 12 months, about how many months was YOUR CHILD without any health insurance coverage?”; and (3) “What are the main reasons YOUR CHILD went without health insurance some time in the last 12 months?” (Both (2) and (3) included an option for “My child was covered by health insurance for all of the past 12 months.”). All 3 included an option for “My child has never had health insurance.” We created 5 health insurance gap categories: no coverage gap, a coverage gap shorter than 6 months, a coverage gap between 6 and 12 months, a coverage gap longer than 12 months, or never had health insurance.
There was <1% inconsistency among responses to the 3 questions; however, to avoid misassignment of a continuously covered child as having a gap, we assigned inconsistent responses to no coverage gap. Additional independent variables for the multivariable analyses included: age (from self-report—1–4 years, 5–9 years, 10–14 years, 15–18 years), race/ethnicity (from self-report and administrative data—white/non-Hispanic, Hispanic/any race, non-white/non-Hispanic), household income (as a percentage of FPL—derived from self-reported household size and administrative income data), parental employment (not employed, employed/self-employed—based on self-report), urban or rural residence (based on zip code designations from the Oregon Office of Rural Health), and whether or not the child was reported to have a special health care need (based on self-report).
We used SPSS version 14.0 (SPSS, Chicago, Ill) with the complex samples module to conduct our analyses, which accounted for the complex sampling design. We first identified demographic characteristics associated with children’s insurance coverage gaps through χ2 bivariate analyses and individual logistic regression models. Then, to assess how different children’s coverage gap lengths were independently associated with health care access outcomes, we performed multivariable logistic regression models. We found no significant interactions (P <.15) between the key independent variable and each of the covariates in the original logistic regression models, so the final model was constructed without interaction terms and included only the key independent variable and covariates, as described above.
We received completed surveys from 2681 of the 8636 eligible households (31% response rate). This response rate is consistent with other similar surveys of Medicaid-eligible populations.19–21 Survey respondents had similar characteristics to the total population, as assessed by comparisons of race/ethnicity, gender, age, geographic region, household income, and current enrollment status in a public insurance program. We used administrative data from respondents and nonrespondents to weight survey responses back to the original population. We further adjusted the final weights by using a raking ratio estimation process to account for nonresponse.22,23 This article reports weighted results unless otherwise specified.
Nearly one-fourth of the population (23.5%) described themselves as Hispanic. Most respondents reported their child’s race as white (70%). Approximately two-thirds of the children lived in single-parent households. Almost 42% of parents reported current employment. Approximately 45% of the households had one or more uninsured adults. Over 12.8% of households had zero income, and most households had monthly earnings below 100% of the FPL (Table 1). These numbers are in reference to the population with known insurance status (2468 out of 2681).
Almost 11% of the children were uninsured at the time of the survey, and 25.4% of all children had an insurance coverage gap sometime during the preceding 12 months. Overall, approximately 17.5% of the children had a gap of shorter than 6 months, 1.5% had a gap between 6 and 12 months, 3.1% had a gap longer than 12 months, and 3.3% had never been covered by health insurance. As previously reported in the literature, children more likely to have coverage gaps were teenagers over age 14; were of Hispanic origin; lived in families with zero income or an income slightly above the FPL; and/or had an employed parent.
Among children in Oregon’s food stamp population, coverage gaps of any length were associated with compromised access to health care. In most cases, increasing length of insurance coverage gaps tended to coincide with increasing odds of having an unmet need (Tables 2 and and3).3). In particular, this relationship appeared to be true for unmet need for dental care, delayed urgent care, and having no usual source of care.
Among this population of Oregon children presumed eligible for public health insurance, 1 of 4 had a gap in coverage during a 12-month period. Explanations for children’s insurance instabilities are beyond the scope of this article; however, Fairbrother and colleagues4 observed that Oregon’s requirement to renew coverage every 6 months, instead of every 12 months, may contribute to higher levels of churning.
Our study confirms previous findings about the importance of continuous health insurance coverage.9,24 It goes beyond past studies in showing a possible tendency for children with longer gaps in insurance coverage to have greater odds of having an unmet need. State policy discussions should acknowledge that children with short gaps of only a few months experienced significant unmet need. Specifically, further efforts should simplify public insurance enrollment, shorten or eliminate SCHIP waiting periods, and extend coverage qualification periods in order to retain eligible children.4,6,25 Furthermore, because access further worsens for children with coverage gaps longer than 6 months, the design of state and federal policies should ensure fluidity of coverage, despite family income fluctuations.
Interpreting data presented here requires consideration of several important factors. First, families already connected to at least one system of public benefits (food stamps) may encounter fewer health insurance coverage gaps compared with a more general low-income population. The data from this study can only be generalized to Oregon’s food stamp population, so these results may understate the prevalence of insurance instabilities among all low-income families. This study, however, does capture the relationship between coverage gaps of different lengths and compromised access to care.
Second, for budgetary reasons, the survey was only administered in 3 languages; and telephone follow-up was not possible. Although our response rate is comparable to other similar studies of Medicaid-eligible populations, response bias remains an important consideration. The comprehensive food stamp administrative database and the raking ratio estimation adjustments allowed us to account for slight demographic differences between respondents and nonrespondents.
Finally, recall bias is always a concern with self-reported data. To minimize recall bias, we used validated questions from national surveys, asking respondents to recall occurrences over only the past 12 months, and several questions pertained to similar topics. In constructing the key independent variable, we noted few inconsistencies. For this independent variable, it was possible to capture the number of months that the child did not have coverage, but it was not possible to determine whether the total time of uninsurance represented one continuous gap or if the child had more than one coverage gap during the 12-month period.
Policy makers should pay close attention to the strong association between coverage gaps and children’s unmet health care needs. Frequent renewal requirements and mandatory waiting periods of uninsurance as a prerequisite for public insurance eligibility may affect a child’s ability to access needed care. High rates of churning, even if coverage gaps are short, may limit access to health care for children. Every effort should be made to minimize barriers to continuous health insurance coverage for children.
Supported in part by a grant obtained by the Office for Oregon Health Policy and Research from the US Health Resources and Services Administration (HRSA). Jennifer DeVoe’s time on this project was supported by grant numbers 5-F32-HS014645 and 1-K08-HS16181 from the Agency for Health care Research and Quality (AHRQ).
We thank the Office for Oregon Health Policy and Research (OHPR), the Oregon Department of Children, Adults and Families (CAF—food stamp office), and the Oregon Department of Medical Assistance Programs. We are grateful for contributions from Moira Ray, Janne Boone, Jessica Miller, James Oliver, Rebecca Ramsey, Pooya Naderi, Ron Taylor, and Jeff Tharpe. We thank Tina Edlund for her survey design expertise. We also acknowledge our 3 anonymous reviewers for their insights and helpful advice.