Lack of health insurance is a key barrier to accessing care for chronic conditions and cancer screening. The influence of insurance type (private, public, none) on survivor-focused and general preventive health care in adult survivors of childhood cancer was examined.
The Childhood Cancer Survivor Study is a retrospective cohort study of childhood cancer survivors diagnosed between 1970 and 1986. Among 8425 adult survivors, the relative risk (RR) and 95% confidence interval (CI) of receiving survivor-focused and general preventive health care were estimated for uninsured (n = 1390) and publicly insured (n = 640), compared with for the privately insured (n = 6395)
Uninsured survivors were less likely than those privately insured to report a cancer-related visit (adjusted RR, 0.83; 95% CI, 0.75-0.91) or a cancer center visit (adjusted RR, 0.83; 95% CI, 0.71-0.98). Uninsured survivors had lower levels of utilization in all measures of care in comparison with privately insured. In contrast, publicly insured survivors were more likely to report a cancer-related visit (adjusted RR, 1.22; 95% CI, 1.11-1.35) or a cancer center visit (adjusted RR, 1.41; 95% CI, 1.18-1.70) than were privately insured survivors. Although publicly insured survivors had similar utilization of general health examinations, they were less likely to report a Papanicolaou test or a dental examinations
There are in excess of 325,000 survivors of childhood cancer in the United States.1 This number continues to grow as the 5-year survival rate for pediatric cancers approaches 80%.2 Despite improved survival rates, this population has an increased risk of premature mortality and diminished health status because of the late effects or long-term complications of cancer treatment, particularly in the adult years.3-6 Because the incidence and severity of many late effects may be reduced with prevention and early detection, the Institute of Medicine recommends lifetime follow-up for all childhood cancer survivors.7 Fewer than 50% of adult survivors of childhood cancer who reside in the United States and Canada are receiving cancer-related follow-up care.8 Only 32% are reporting survivor-focused care, including risk reduction counseling or late-effects screening.9
Lack of health insurance is a key barrier to accessing medical care in the United States for adults with chronic conditions.10 Lack of health insurance is associated with lower cancer screening utilization.11 Among childhood cancer survivors, risk factors for being uninsured include younger age at cancer diagnosis, lower educational level, income less than $20,000, marital status (widowed, divorced, or separated), being a current or former smoker, and cranial radiation treatment.12 In addition to insurance status, ethnicity is associated with different rates of follow-up care for childhood cancer survivors. When adjusting for having health insurance, Hispanic female survivors are more likely to follow up at a cancer center than are white, non-Hispanic female survivors but are less likely to report standard cancer screening such as Papanicolaou (Pap) tests.13 These findings are in line with studies examining cancer screening practices and cancer care of the general US population.14, 15 Studies conducted within health maintenance organizations or managed care practices have demonstrated lower utilization rates of preventive care by specific groups of ethnic minorities despite the same access to services.16, 17
We sought to better understand the complex relationship between health insurance status, racial/ethnic group status, and health care utilization in this high-risk population. Using the large, geographically and socioeconomically diverse Childhood Cancer Survivor Study (CCSS) cohort, we examined the influence of insurance type (private, public, and no insurance) on survivor-focused and general preventive health care within 3 racial/ethnic groups of adult survivors of childhood cancer.
MATERIALS AND METHODS
Childhood Cancer Survivor Study
The CCSS is a multi-institutional study of individuals who have survived ≥5 years after treatment for childhood cancer. Childhood cancer diagnoses include leukemia, brain tumor, Hodgkin lymphoma, non-Hodgkin lymphoma, Wilms tumor, neuroblastoma, soft-tissue sarcoma, and bone tumors. The participants were diagnosed before 21 years of age and treated at 1 of 26 collaborating CCSS institutions between January 1, 1970, and December 31, 1986. A detailed description of the CCSS study design and conduct methodology has been reported.18-20 The study was approved by the institutional review board at each of the participating institutions, and informed consent was obtained from each participant.
The current analysis included 8425 survivors who completed the CCSS Baseline Questionnaire, were ≥18 years old, and completed information for race/ethnicity and insurance information (questionnaire can be downloaded from http://ccss.stjude.org). Because of differences in health care systems, participants who lived in Canada were excluded (n = 735). Only survivors ≥18 years were included, as this is the age period when insurance coverage is typically lost for adult survivors of childhood cancer because of aging out of parental or public insurance programs. As shown in Figure 1, of the 20,691 childhood cancer survivors included in the cohort, 3058 (14.8%) were lost to follow-up because of unsuccessful tracing due to incorrect last known address provided by the treating institution. The participants and those lost to follow-up were similar in sex, cancer type, treatment received, age at diagnosis, and age at study participation (or for those lost to follow-up, the age at which the cohort was assembled). Among the 17,633 subjects located, 14,357 (81.4%) completed the questionnaire, including 10,398 participants who were ≥18 years. The cohort examined here was based on subjects' self-report of race/ethnicity. Those patients (n = 1238) with missing race or insurance information were excluded from the analysis.
Primary Outcome Measures
Survivor-focused health care
Two measures defined survivor-focused health care received in the previous 2 years: cancer-related visits and cancer center visits. As previously described, participants were asked how many visits to a physician's office were related to their previous cancer (cancer-related visit) and whether any of the visits were at an oncology center (cancer center visit).8 These outcomes were not mutually exclusive.
General preventive health care
Two measures defined general preventive health care for both men and women: having a general physical examination ≤2 years ago and a dental examination ≤1 year ago. For women, general preventive health care included a clinical breast examination ≤1 year ago and a Pap test ≤3 years ago.21
Health insurance was classified as private, public (Medicaid, Medicare, or other public assistance programs), or no health insurance. Based on self-reported race/ethnicity, participants were categorized into 1 of 3 groups: non-Hispanic white (NHW); black, non-Hispanic; and Hispanic. Other independent variables included sex, highest level of educational attainment, household income, and cancer diagnosis. To assess the influence of comorbid health conditions on health care utilization, we included the prevalence and severity of a chronic health condition. As previously described, the severity of chronic health conditions was based on the National Cancer Institute Common Terminology Criteria for Adverse Events (version 3) and classified as mild (grade 1), moderate (grade 2), severe (grade 3), or life-threatening or disabling (grade 4).4
Descriptive statistics were computed within each category of the 3 insurance categories and compared across insurance categories, using the chi-square test for categorical variables and ANOVA for continuous variables. Prevalence was estimated for sociodemographic variables, medical variables, and type of survivor-focused and general preventive health care received. Using log-binomial regression with an interaction term of race and insurance types, we estimated the relative risk (RR) of each outcome of interest (survivor-focused and general preventive health care) in survivors with no insurance and those with public insurance, relative to those with private insurance, within each ethnic group. These RR estimates were adjusted for age, sex, household income, highest level of education attained, and having a severe, life-threatening, or disabling chronic condition (grade 3 or 4). For each RR estimate, a corresponding 95% confidence interval (CI) was obtained using the standard large-sample inference method for generalized linear models. In the case in which log-binomial regression did not converge, the COPY method was employed.22 Data were analyzed with SAS statistical software (version 9.1; SAS Institute Inc, Cary, NC).
The mean age at interview for the entire cohort was 28.1 years for the privately insured, 27.2 years for the publicly insured, and 26.3 years for the uninsured. The age range at interview for the entire cohort was 18-48.9 years. The mean interval from time of diagnosis for the entire cohort was 17.5 years for the privately insured, 17.4 years for the publicly insured, and 17.4 years for the uninsured.
Table 1 reports the additional characteristics of the 8425 participants who were adult survivors of childhood cancer. There were more men who were uninsured (59.8%) than were men with private and public insurance within the entire cohort and for each ethnic group. In contrast, within the entire cohort, women were more likely to be publicly insured (59.8%) compared with the rates of women with private insurance or who were uninsured. Publicly insured survivors had lower educational attainment (63.4% were high school graduates or less) and lower household incomes (67.1% with incomes less than $20,000) than did privately insured survivors or those without health insurance. There were higher rates of being uninsured among leukemia and non-Hodgkin lymphoma survivors. The publicly insured group was disproportionately represented by brain tumor survivors and survivors with a serious chronic health condition (grade 3 or 4). These trends were similar across all 3 ethnic groups.
Table 1. Characteristics of Adult Survivors of Childhood Cancer by Race/Ethnicity and Insurance Type
Entire Cohort (n=8425)
Black Non-Hispanic Survivors (n=417)
Hispanic Survivors (n=477)
White Non-Hispanic Survivors (n=7531)
Private (n=6395), %
Public (n=640), %
None (n=1390), %
Private (n=256), %
Public (n=67), %
None (n=94), %
Private (n=296), %
Public (n=72), %
None (n=109), %
Private (n=5843), %
Public (n=501), %
None (n=1187), %
Abbreviations: HS or less, some high school or high school graduate; HS + some college, high school graduate with either some college courses or other training; CNS, central nervous system tumor; HL, Hodgkin lymphoma; NHL, Non-Hodgkin lymphoma.
Note: Percentages are based on the total with available data for each variable.
P < .05 from test of equality across the 3 insurance categories among all survivors.
P < .05 from test of equality across the three insurance categories among Black survivors.
P < .05 from test of equality across the three insurance categories among Hispanic survivors.
P < .05 from test of equality across the three insurance categories among White survivors.
Survivor-focused and general preventive health care reported by the entire cohort and across the 3 ethnic groups is provided in Table 2. Uninsured survivors, for the entire cohort and, for the most part, across each ethnic group demonstrated the lowest rates of survivor-focused and general preventive health care. The single exception was for uninsured black, non-Hispanic survivors whose rate of reporting survivor-focused care was not significantly different from their counterparts with private or public insurance. In contrast, publicly insured survivors reported the highest rates of survivor-focused health care (entire cohort and across each ethnic group except for black survivors). Although the proportion of publicly insured survivors who reported a general physical examination in the previous 2 years (68.6%) was similar to that of privately insured survivors (67.1%), publicly insured survivors reported lower rates of other types of general preventive care (dental care, clinical breast examinations, Pap tests). Publicly insured Hispanic women had lower rates of reporting a Pap test (68.9%) compared with those with private insurance (76.5%). Uninsured Hispanic women had even lower rates of reporting a Pap test (53.3%). Among NHW females, both public and uninsured survivors had lower rates of Pap test use (72.2% and 72.9%, respectively) compared with privately insured women (81.9%). The proportion of uninsured black women reporting a Pap test was similar to those who had private or public insurance.
Table 2. Percentage of Black, Non-Hispanic, Hispanic, and White, Non-Hispanic Adult Survivors of Childhood Cancer Who Reported the Following Types of Health Care, by Insurance Type
Table 3 reports the adjusted relative risks of reporting survivor-focused and general preventive health care among survivors with public insurance or no insurance in comparison with those with private insurance (referent). After adjustment for race/ethnicity, age at interview, sex, household income, highest level of education attained, and presence of a serious (grade 3 or 4) chronic health condition, uninsured survivors were less likely to report a cancer-related visit (RR, 0.83; 95% CI, 0.75-0.91) and also less likely to report a cancer center visit (RR, 0.83; 95% CI, 0.71-0.98) than were privately insured survivors. In contrast, publicly insured survivors were more likely to have had survivor-focused health care (both cancer-related and cancer center visits) than were privately insured survivors (RR, 1.22; 95% CI, 1.11-1.35; and RR, 1.41; 95% CI, 1.18-1.70, respectively). Within the 3 ethnic groups, uninsured NHW survivors were less likely to report survivor-focused care (RR, 0.82; 95% CI, 0.74-0.90, for a cancer-related visit; RR, 0.79; 95% CI, 0.66-0.94 for a cancer center visit) compared with privately insured NHW survivors. Both uninsured and publicly insured NHW female survivors were less likely to report Pap test use (RR, 0.93; 95% CI, 0.87-0.99; and RR, 0.87; 95% CI, 0.80-0.95, respectively).
Table 3. RR and 95% CI of the Likelihood of Reporting Survivor-Focused or General Preventive Health Care Among Survivors With Public Insurance or No Insurance Compared With Among Those With Private Insurance (Reference)
Uninsured survivors were less likely than privately insured survivors to report all 4 measures of general preventive health care. However, in contrast with the above-described trends for survivor-focused health care, publicly insured survivors were not more likely to report general preventive care than privately insured survivors. Although the likelihood of a general physical examination was similar in both groups, publicly insured survivors were less likely to report a dental examination (RR, 0.90; 95% CI, 0.82-0.99), and publicly insured female survivors were less likely to report a Pap test (RR, 0.90; 95% CI 0.84-0.97) than were privately insured women.
Adult survivors of childhood cancer are at risk for myriad late effects related to their cancer treatment, including early-onset cardiovascular disease, stroke, and second malignancies.23-27 Much of morbidity and mortality related to the childhood cancer therapy received occurs during young adulthood, with a long latency after the initial exposure.3-6 It is important for adult survivors of childhood cancer to have access to long-term follow-up care and cancer screening, with the intent of preventing or lessening future morbidity and mortality. Affordable health insurance plans and/or public programs are an important factor for ensuring survivor-focused health care.
To our knowledge, this is the first large study of adult survivors of childhood cancer from across the United States to examine the influence of insurance type by 3 ethnic groups on survivor-focused and general preventive health care utilization. There were several notable findings. Despite being a significantly more disadvantaged group with high rates of poverty, survivors with public health insurance reported utilizing survivor-focused health care at rates higher than survivors with private health insurance. This suggests that Medicaid/Medicare services are providing much needed access to care for high-risk survivors with serious health conditions related to their previous cancer treatment. Our findings are similar to other studies conducted among low-income populations that have found that having any type of health insurance coverage, including public, has a significant impact on access to needed health care services.28
In contrast, a substantial proportion of uninsured survivors with serious chronic diseases did not report utilization of survivor-focused or general preventive health care. Large population-based studies have demonstrated significant benefits of public health insurance programs on the receipt of quality health care, as well as improvements in the continuity of care and receipt of preventive health services.29-32 Similar policy initiatives, in which both the federal and state governments finance a public insurance plan for low-income, at-risk childhood cancer survivors, could provide a vital safety net to improve access to health services.
We found significant differences in the rates of utilization of general preventive health care within the different categories of insurance coverage. We analyzed the 3 ethnic groups separately because health care utilization can vary by ethnicity for various factors, including cultural influences such as acculturation and nativity.33-35 Similar to other studies, we found lower reported rates of utilization of general preventive health care for the uninsured across all 3 ethnic groups.36 In the general population, uninsured adults are more likely to be diagnosed with advanced stage cancers because of poor access to cancer screening.37-40
We found lower probabilities of reporting cancer-screening practices (clinical breast examination and Pap test) for both uninsured and publicly insured NHW women. This finding of lower utilization rates in uninsured and publicly insured NHW survivors has similarly been observed in other cancer control studies. For example, a study examining the breast cancer screening practices of uninsured women found that uninsured NHW women (even after controlling for socioeconomic status factors) had lower reported utilization than did Black and Hispanic women.41 Community-level factors, specifically county-level proportions of uninsured women, affected breast cancer screening rates. Women who lived in counties with higher rates of uninsured were less likely to be screened. This county-level effect on screening rates, however, had little impact on those who had household incomes between $25,000 and $75,000. A second study, evaluating county-level covariates (including residence in health professional shortage areas, urban/rural setting, racial/ethnic composition, and number of health centers/clinics) found that Black women were more likely than NHW women to report Pap test. Among women who resided in urban areas with a smaller supply of primary care physicians, there were lower rates of Pap test use. Women in rural areas were also less likely to report Pap test use.42 Although the explanatory factors for the observed differences for NHW survivors in this CCSS study are not known, community-level factors that may affect health care utilization for diverse groups of survivors should be explored in future research.28, 43
We also found lower reported rates of dental care utilization for both the uninsured and publicly insured compared with privately insured survivors. In the general population, adult Medicaid beneficiaries have less utilization of dental services than privately insured adults.44 In states evaluating methods to enhance access to dental care services, improvements in utilization occurred when Medicaid programs reimbursed dental charges at rates comparable to private dental rates.45, 46 Given the significant risk for delayed and poor dental development in childhood cancer survivors, policy considerations to improve necessary dental care services for survivors are needed.47, 48
This study shows that public insurance programs result in high rates of reported utilization of survivor-focused health care. Specifically, we found that both Hispanic and NHW publicly insured survivors had a higher likelihood of reporting a cancer center visit compared with privately insured survivors. This finding was unanticipated, as we hypothesized that privately insured survivors would have the highest utilization of survivor-focused health care. There are several possible explanations. First, having a cancer diagnosis reflects a “teachable moment,” particularly for survivors with public insurance.49, 50 If cancer patients with a lower socioeconomic status enter the public health care system for the first time because of the need for cancer treatment, they may be more motivated to continue the recommended follow-up. They are now able to utilize health care services that may have not been available prior to their diagnosis of cancer. In contrast, those survivors with private insurance may have barriers to utilizing survivor-focused health care because of time missed from work, large out-of-pocket spending, higher deductibles, copayments, and/or lifetime caps on insurance coverage.51
A second explanation for our finding may be that the quality of care for publicly insured survivors is influenced by the type of hospital where they receive care. In the Medicare population, black or poor patients are more likely to receive care in urban teaching hospitals, which deliver higher quality of care.52 Although black and poor patients were found to receive lower quality of care, when adjusted for hospital type, the receipt of care in these urban teaching hospitals almost completely offsets the poor quality care they received within each hospital. The authors found, through the use of zip code data, that black or poor patients are almost 2 times more likely to receive care in urban teaching hospitals than in rural or nonteaching hospitals. It is also possible that adult childhood cancer survivors are more likely to receive their survivorship care in urban teaching hospitals, as the vast majority of pediatric oncology care is delivered within academic centers.
We found lower reported utilization of survivor-focused and general preventive care for uninsured survivors. We hypothesize these findings may be a result of lacking a usual source of care, being underinsured, and having lapses in insurance coverage.11, 28, 53 Because the primary causes of late mortality among adult childhood cancer survivors include cancer recurrence or development of a second malignancy, affordable access to general preventive care, including cancer screening, is critical for survivors.3 It will be important for future research to evaluate the effect of the recent national policy changes including the Patient Protection and Affordable Care Act (passed in the Senate, December 2009) and the major expansion of Medicaid.54
When interpreting our study results, there are the following limitations. Although this is the largest national cohort of childhood cancer survivors, the insurance status and socioeconomic status of nonparticipants is not known. As a result, there may be a selection bias in our sample, including a possible lower representation of uninsured survivors or those with a lower socioeconomic status, which could select for those survivors who are more likely to utilize care, thereby decreasing differences observed across insurance groups. Using self-reported data for the measurement of utilization of survivor-focused and general preventive health care can result in an overestimation of receipt of services because of recall bias.55, 56 Our findings demonstrate a statistically significant lower probability of reporting survivor-focused and general preventive health care for uninsured childhood cancer survivors. These findings further emphasize the need to develop targeted policy efforts to improve access to affordable health care options for adult survivors of childhood cancer. A third limitation is the use of current insurance status alone and lack of data regarding the continuity of insurance status for survivors in the sample. Previous population-based studies demonstrated that the uninsured and those unstably insured (ie, lapses in continuous insurance coverage) report nearly the same rates of poor access to care.43 Future research examining the effect of uninterrupted insurance coverage on access to survivor-focused and general preventive health care in young adult survivors is essential.
In summary, a significant proportion of uninsured adult survivors of childhood cancer, across all ethnicities have lower utilization rates of survivor-focused and general preventive health care. Targeted policy changes directed at greater access to affordable health care for all adult survivors of childhood cancer are critical given the significant burden of chronic disease due to cancer treatment at a young age.
CONFLICT OF INTEREST DISCLOSURES
Supported by Grant U24-CA-55727 (L.L. Robison, principal investigator) from the Department of Health and Human Services; funding to the University of Minnesota from the Children's Cancer Research Fund; funding to St. Jude Children's Research Hospital from the American Lebanese Syrian Associated Charities (ALSAC), the UCLA STOP Cancer Career Development Award, and the Harold Amos Medical Faculty Development Program of the Robert Wood Johnson Foundation.