Abstract
Purpose:
To investigate racial/ethnic-related disparities by insurance status in “forgoing needed medical care in the last year due to finances” in childhood cancer survivors.
Methods:
Our study included 3310 non-Hispanic/Latinx White, 562 non-Hispanic/Latinx Black, and 92 Hispanic/Latinx survivors from the St. Jude Lifetime Cohort Study. Logistic regression analyses, guided by Andersen’s Healthcare Utilization Model, were adjusted for “predisposing” (survey age, sex, childhood cancer diagnosis and treatment, and treatment era) and “need” (perceived health status) factors. Additional adjustment for household income/education and chronic health conditions was considered.
Results:
Risk of forgoing care was highest among non-Hispanic/Latinx Blacks and lowest among Hispanics/Latinxs for each insurance status. Among privately insured survivors, relative to non-Hispanic/Latinx Whites, non-Hispanic/Latinx Blacks were more likely to forgo care (adjusted OR: 1.82, 95% CI: 1.30–2.54): this disparity remained despite additional adjustment for household income/education (adjusted OR: 1.43, 95% CI: 1.01–2.01). In contrast, publicly insured survivors, regardless of race/ethnicity, had similar risk of forgoing care as privately insured non-Hispanic/Latinx Whites. All uninsured survivors had high risk of forgoing care. Additional chronic health condition adjustment did not alter these results.
Conclusions:
Provision of public insurance to all childhood cancer survivors may diminish racial/ethnic disparities in forgoing care that exist among the privately insured and reduce the risk of forgoing care among uninsured survivors to that of privately insured non-Hispanic/Latinx Whites.
Implications for Cancer Survivors:
Under public insurance, childhood cancer survivors had low risk of forgoing care, at the similar level to privately insured non-Hispanic/Latinx Whites, regardless of race/ethnicity.
Keywords: Healthcare disparities, race factors, health services research, health equity, cancer survivors, insurance
Introduction
In the United States (US), approximately 1 in 285 children are diagnosed with cancer prior to age 20 years [1]. With 5-year survival increasing to 85% [2], the number of childhood cancer survivors in the US will exceed half a million by the end of 2020 [3]. This growing population is at risk for acute, chronic, and late-occurring treatment-related morbidity and premature death [4, 5]. By the age of 50 years, survivors have on average experienced 17.1 chronic health conditions, of which 4.7 are severe/disabling, life-threatening, or fatal [5]. This cumulative burden is nearly twice that of age-sex matched community controls [5].
Given this substantial disease burden, childhood cancer survivors represent a vulnerable population for which health insurance is a critical requisite for the maintenance of overall health and wellbeing. Uninsured survivors of childhood cancer are more likely than their insured counterparts to report not receiving survivor-focused/cancer-related and/or usual medical care [6–10]. Health insurance coverage differs by race/ethnicity among childhood cancer survivors, with minority survivors more likely to be uninsured, and among those insured, more likely to have public insurance compared to non-Hispanic/Latinx White survivors [8, 11–13]. Healthcare utilization also differs by race/ethnicity among survivors. Compared to non-Hispanic/Latinx Whites, non-Hispanic/Latinx Black and Hispanic/Latinx childhood cancer survivors report less problem-based “general or nonspecific contact” with healthcare providers [6, 11], but report equitable or higher usage of general physical checkups [6, 11, 12]. Although inconsistent with earlier publications [6, 11], non-Hispanic/Latinx Black survivors have more recently been found to be less likely to receive cancer-related care compared to non-Hispanic/Latinx Whites and Hispanics/Latinxs [7, 12].
Underutilization of healthcare is detrimental to health owing to delays in diagnosis and treatment that culminate in poor health outcomes [14]. Racial/ethnic disparities in healthcare utilization according to insurance status have not been evaluated in childhood cancer survivors. Instead, previous analyses of this high-risk population by the multi-institutional Childhood Cancer Survivor Study have predominantly focused on characterizing the influence of either insurance status or race/ethnicity on healthcare utilization, while accounting for the other variable’s effect with stratification or adjustment—i.e., the possibility of interactions between race/ethnicity and insurance status on healthcare utilization has not been explored [6–9, 11, 12]. The goal of this analysis was to utilize data collected in the single-institution St. Jude Lifetime Cohort (SJLIFE) study to examine differences (disparities) in a specific utilization measure, namely, “forgoing needed medical care due to finances,” by race/ethnicity and insurance status jointly. In addition, we also evaluated if disparities were ameliorated under public insurance.
Methods
Study Population
Our analyses used cross-sectional data from SJLIFE. SJLIFE is a retrospectively constructed cohort study with prospective clinical assessment established to investigate the late effects of childhood cancer and its treatment among 5-year survivors treated for a pediatric malignancy at St. Jude Children’s Research Hospital (SJCRH). SJLIFE eligibility, recruitment, and ascertainment have been described elsewhere in detail [15, 16]. SJLIFE participants complete on-campus medical assessments and written or verbal surveys in English or Spanish every 2 to 5 years. The current analysis utilized the most recent assessment for each participant, whose completion dates ranged from November 24, 2007, to May 25, 2019. Non-US citizens/residents and those who did not complete the health status, insurance, and forgoing care survey questions were excluded (Fig. 1). All participants provided written informed consent in accordance with SJCRH Institutional Review Board approval [15].
Figure 1.
Consort diagram for the study population
Outcome and Primary Exposure Measures
The outcome of forgoing medical care was ascertained from participants’ responses to the survey question: “In the past 12 months, was there a time you needed to see a doctor or go to the hospital but did not go due to finances?” The exposures of interest were race/ethnicity and insurance status. Due to sample sizes, we focused on self-reported racial/ethnic groups of non-Hispanic/Latinx Black, Hispanic/Latinx, or non-Hispanic/Latinx White: survivors of other, unknown, or multiple racial/ethnic backgrounds were excluded (Fig. 1). Hereafter, “Black” and “White” will refer to non-Hispanic/Latinx members of these respective racial groups. Insurance status was determined by a two-part question querying whether a participant currently had health insurance and if yes, what type of insurance they had. Participants could mark multiple insurance types and were classified as (1) “uninsured” if they responded “no”; (2) “privately insured” if they only obtained insurance through their place of employment/education, a spouse or parent’s insurance policy, and/or a policy they purchased themselves; and (3) “publicly insured” if they were covered by Medicaid, Medicare, the Affordable Care Act (ACA), US military, or another governmental program. The two main exposures of interest, race/ethnicity (three categories) and insurance status (three categories), were combined to form a nine-category variable to evaluate effect modification (i.e., different patterns/degrees of disparity across race/ethnicity groups under different insurance statuses), using privately insured White survivors as the reference group.
Conceptual Framework
The conceptual framework for this analysis was the Andersen Healthcare Utilization Model, which has been broadly applied to understand the factors that may influence healthcare utilization [17]. This model posits that healthcare utilization is determined by three types of factors: “predisposing”; “enabling”; and “need” [17]. Our application of this model is shown in Fig. 2, which displays the variables we considered in our examination of disparities in forgone medical care among survivors. Our two main exposures of interest, race/ethnicity and health insurance, were respectively considered “predisposing” and “enabling” variables. For adjustment covariates, “predisposing” factors included age at survey completion, sex, education, and childhood cancer diagnosis group, treatment modality, major surgery, and treatment era. Household income was considered an “enabling” factor whereas perceived self-reported health status [18] and clinically ascertained counts of chronic health conditions were considered “need” factors.
Figure 2.
Adaptation of Andersen’s Healthcare Utilization Model to the present study of forgone medical care disparities among survivors of childhood cancer
Key Adjustment Variable Definition
Indicators of socioeconomic and health status were considered key adjustment variables owing to their strong influence on healthcare utilization behaviors. For socioeconomic status, household income and educational attainment were self-reported in SJLIFE surveys [15]. Household income brackets spanned less than $20,000, $20,000–$39,999, $40,000–$59,999, $60,000–$79,999, $80,000–$99,999, $100,000 or higher, and undetermined for survivors who either reported not knowing their income or did not answer the survey question. Educational attainment was categorized into eight groups: no high school diploma; completed high school/Tests of General Educational Development; training after high school other than college; some college; college graduate; post-graduate level; and undetermined for those who did not answer the survey question.
Both self-reported health status and clinically ascertained chronic health conditions were considered for adjustment of health status. Perceived health status was queried by asking participants if, in general, they considered their health “excellent,” “very good,” “good,” “fair,” or “poor.” For the presence and severity of clinically ascertained chronic health conditions, cumulative burden, which is the number of clinically ascertained chronic health conditions experienced by a survivor, was used. Cumulative burden was classified using the SJCRH–modified National Cancer Institute’s Common Terminology Criteria for Adverse Events version 4.03, which assigns scores of mild (grade 1), moderate (grade 2), severe/disabling (grade 3), life-threatening (grade 4), or fatal (grade 5) [5, 19, 20]. Chronic health conditions with a score of moderate or higher were counted whereas mild events were not; by design, fatal events were not counted since participants had to be alive at the time of survey completion. Cumulative burden was then categorized as 0, 1–3, 4–6, and 7 or more chronic health conditions.
Detailed definition of all other adjustment variables is provided in the Supplementary Methods.
Statistical Analysis
Descriptive statistics compared Black, Hispanic/Latinx, and White survivors with respect to all demographic, cancer-related, and health variables included in our analysis. To describe forgoing medical care by race/ethnicity and insurance status, the observed percentage of those who reported forgoing care was calculated for each of the nine race/ethnicity-insurance groups. A main multivariable logistic regression model was constructed based on Andersen’s Model [17] to estimate the odds of forgoing care by race/ethnicity-insurance group. To account for the potential influence of demographic, cancer-related, and health characteristics on the association between racial/ethnic-insurance group and forgoing care, we adjusted for age at survey completion, sex, treatment era, childhood cancer diagnosis group, treatment modality, major surgery, and self-reported health status. In the US, race/ethnicity and insurance are related to household income and educational attainment for historical, political, and social reasons. Therefore, controlling for income and education in evaluating the total disparity across race/ethnicity-insurance groups is arguably inappropriate because it would take away some of the disparity as “explained” by the factors that are actually components of the disparity. However, it is also meaningful to estimate income-and-education-adjusted disparity so that the estimates reflect differences across race/ethnicity-insurance groups that are not attributable to their differences in income and education (e.g., effects of contextual factors such as policies, mistrust, and discrimination). Therefore, a secondary model was fit adjusting further for household income/education. To assess whether self-reported health status adequately represents the need for healthcare, the cumulative burden of moderate to life-threatening chronic health conditions was added to the main model in a supplementary analysis. Additionally, details of a supplementary analysis on insurance stability are provided in the Supplementary Methods. The Hosmer-Lemeshow test was used to assess the goodness of fit for all logistic regression models. The threshold for statistical significance was set to a two-sided Wald-test p-value less than 0.05. All statistical analyses were conducted in R version 3.6.1 [21] using RStudio [22] and bar graphs were created using ggplot2 [23].
Results
Population Characteristics
Of the 5017 survivors who completed a SJLIFE survey, 3964 fulfilled inclusion criteria (Fig. 1). Table 1 provides the demographic, cancer-related, and health characteristics. The study population is comprised of 562 Blacks, 92 Hispanics/Latinxs, and 3310 Whites. Across the three groups, the proportion of males to females was similar: females comprised 52.5% of Black, 50.0% of Hispanic/Latinx, and 47.7% of White survivors. Only 24.7% of Blacks completed college/post-graduate education compared to 39.1% of Hispanics/Latinxs and 37.2% of Whites. A third of Black survivors reported an annual household income of less than $20,000, which was twice the percentages observed for Hispanics/Latinxs and Whites. Conversely, 40.1% of Whites reported annual household incomes of $60,000 or higher, which was twice the percentages observed for Blacks and Hispanics/Latinxs. Disproportionately more Blacks reported having fair/poor health compared to Hispanic/Latinx and White survivors. For each race/ethnicity group, about half of all survivors had four or more clinically assessed moderate to life-threatening chronic health conditions. Private health insurance was most commonly utilized by all three race/ethnicity groups, although Blacks more frequently reported having either no or public insurance compared to Hispanics/Latinxs and Whites; both observations also applied to age-defined subgroups of each race/ethnicity (Table S1). Although uninsured Blacks and Hispanics/Latinxs were just as likely as uninsured Whites to gain insurance, insured Blacks had 2.62 (95% CI: 1.78–3.87) times higher adjusted odds of losing insurance compared to insured Whites (Table S2).
Table 1.
Demographic, cancer-related, and health characteristics of study participants by racial/ethnic group
| Characteristic | Black (N=562) N (%) |
Hispanic/Latinx (N=92) N (%) |
White (N=3310) N (%) |
|---|---|---|---|
| Demographic characteristics | |||
| Age at survey | |||
| <20 years | 32 (5.7%) | 5 (5.4%) | 128 (3.9%) |
| 20-29 years | 249 (44.3%) | 60 (65.2%) | 1118 (33.8%) |
| 30-39 years | 162 (28.8%) | 22 (23.9%) | 1118 (33.8%) |
| 40-49 years | 90 (16.0%) | 4 (4.3%) | 703 (21.2%) |
| ≥50 years | 29 (5.2%) | 1 (1.1%) | 243 (7.3%) |
| Sex | |||
| Female | 295 (52.5%) | 46 (50.0%) | 1578 (47.7%) |
| Male | 267 (47.5%) | 46 (50.0%) | 1732 (52.3%) |
| Education | |||
| No high school diploma | 72 (12.8%) | 9 (9.8%) | 269 (8.1%) |
| High school diploma | 110 (19.6%) | 17 (18.5%) | 593 (17.9%) |
| Training after high school, other than college | 29 (5.2%) | 3 (3.3%) | 154 (4.7%) |
| Some college | 153 (27.2%) | 20 (21.7%) | 826 (25.0%) |
| College graduate | 111 (19.8%) | 21 (22.8%) | 910 (27.5%) |
| Post-graduate level | 28 (5.0%) | 15 (16.3%) | 322 (9.7%) |
| Undetermined a | 59 (10.5%) | 7 (7.6%) | 236 (7.1%) |
| Household income | |||
| <$19,999 | 164 (29.2%) | 13 (14.1%) | 493 (14.9%) |
| $20,000-$39,999 | 134 (23.8%) | 19 (20.7%) | 604 (18.2%) |
| $40,000-$59,999 | 57 (10.1%) | 20 (21.7%) | 533 (16.1%) |
| $60,000-$79,999 | 41 (7.3%) | 8 (8.7%) | 431 (13.0%) |
| $80,000-$99,999 | 28 (5.0%) | 5 (5.4%) | 287 (8.7%) |
| ≥$100,000 | 30 (5.3%) | 7 (7.6%) | 610 (18.4%) |
| Undetermined b | 108 (19.2%) | 20 (21.7%) | 352 (10.6%) |
| Health insurance | |||
| None | 144 (25.6%) | 20 (21.7%) | 460 (13.9%) |
| Public | 186 (33.1%) | 17 (18.5%) | 718 (21.7%) |
| Private | 232 (41.3%) | 55 (59.8%) | 2132 (64.4%) |
| Time since diagnosis | |||
| 5-9 years | 29 (5.2%) | 5 (5.4%) | 105 (3.2%) |
| 10-19 years | 207 (36.8%) | 40 (43.5%) | 990 (29.9%) |
| 20-29 years | 196 (34.9%) | 44 (47.8%) | 1096 (33.1%) |
| 30-39 years | 93 (16.5%) | 2 (2.2%) | 815 (24.6%) |
| 40-49 years | 32 (5.7%) | 1 (1.1%) | 283 (8.5%) |
| ≥50 years | 5 (0.9%) | 0 (0%) | 21 (0.6%) |
|
| |||
| Cancer-related characteristics | |||
| Treatment era | |||
| <1980 | 67 (11.9%) | 1 (1.1%) | 639 (19.3%) |
| 1980s | 153 (27.2%) | 11 (12.0%) | 984 (29.7%) |
| 1990s | 196 (34.9%) | 51 (55.4%) | 1063 (32.1%) |
| ≥2000 | 146 (26.0%) | 29 (31.5%) | 624 (18.9%) |
| Cancer diagnosis group | |||
| Acute lymphoblastic leukemia | 100 (17.8%) | 38 (41.3%) | 1048 (31.7%) |
| Acute myeloid leukemia | 32 (5.7%) | 12 (13.0%) | 108 (3.3%) |
| Chronic myeloid leukemia | 3 (0.5%) | 3 (3.3%) | 22 (0.7%) |
| Central nervous system | 72 (12.8%) | 10 (10.9%) | 425 (12.8%) |
| Hodgkin lymphoma | 67 (11.9%) | 8 (8.7%) | 419 (12.7%) |
| Non-Hodgkin lymphoma | 34 (6.0%) | 4 (4.3%) | 237 (7.2%) |
| Ewing sarcoma family of tumors | 3 (0.5%) | 2 (2.2%) | 111 (3.4%) |
| Osteosarcoma | 32 (5.7%) | 5 (5.4%) | 106 (3.2%) |
| Retinoblastoma | 32 (5.7%) | 2 (2.2%) | 92 (2.8%) |
| Rhabdomyosarcoma | 29 (5.2%) | 1 (1.1%) | 110 (3.3%) |
| Soft tissue sarcoma | 30 (5.3%) | 1 (1.1%) | 91 (2.7%) |
| Neuroblastoma | 19 (3.4%) | 4 (4.3%) | 140 (4.2%) |
| Germ cell tumor | 29 (5.2%) | 0 (0.0%) | 63 (1.9%) |
| Wilms tumor | 51 (9.1%) | 1 (1.1%) | 187 (5.6%) |
| Nasopharyngeal carcinoma | 13 (2.3%) | 1 (1.1%) | 9 (0.3%) |
| Other | 16 (2.8%) | 0 (0.0%) | 142 (4.3%) |
| Treatment modality | |||
| No chemotherapy or radiation | 72 (12.8%) | 3 (3.3%) | 278 (8.4%) |
| Chemotherapy only | 210 (37.4%) | 46 (50.0%) | 1146 (34.6%) |
| Radiation only | 40 (7.1%) | 0 (0.0%) | 266 (8.0%) |
| Chemotherapy and radiation | 240 (42.7%) | 43 (46.7%) | 1620 (48.9%) |
| Major surgery | |||
| No | 116 (20.6%) | 41 (44.6%) | 931 (28.1%) |
| Yes | 446 (79.4%) | 51 (55.4%) | 2379 (71.9%) |
|
| |||
| Health characteristics | |||
| Self-reported health status | |||
| Excellent | 66 (11.7%) | 13 (14.1%) | 370 (11.2%) |
| Very good | 136 (24.2%) | 42 (45.7%) | 998 (30.2%) |
| Good | 189 (33.6%) | 23 (25.0%) | 1205 (36.4%) |
| Fair | 147 (26.2%) | 12 (13.0%) | 590 (17.8%) |
| Poor | 24 (4.3%) | 2 (2.2%) | 147 (4.4%) |
| Clinically ascertained chronic health conditions | |||
| 0 conditions | 22 (3.9%) | 8 (8.7%) | 151 (4.6%) |
| 1-3 conditions | 238 (42.3%) | 36 (39.1%) | 1099 (33.2%) |
| 4-6 conditions | 138 (24.6%) | 31 (33.7%) | 868 (26.2%) |
| ≥7 conditions | 132 (23.5%) | 17 (18.5%) | 953 (28.8%) |
| Did not complete evaluation | 32 (5.7%) | 0 (0.0%) | 239 (7.2%) |
|
| |||
| Study outcome variable | |||
| Forgone care due to finances in last year | |||
| No | 360 (64.1%) | 77 (83.7%) | 2542 (76.8%) |
| Yes | 202 (35.9%) | 15 (16.3%) | 768 (23.2%) |
“Undetermined” refers to those who did not complete the survey question on education.
“Undetermined” refers to those who either reported not knowing their income or did not answer the survey question on income.
Observed Percentage of Survivors Forgoing Care by Race/Ethnicity and Insurance Status
Thirty-six percent of Black, 23.2% of White, and 16.3% of Hispanic/Latinx survivors reported forgoing needed medical care in the past 12 months due to finances (Table 1). The distribution of survivors forgoing care by race/ethnicity and insurance status is provided in Fig. 3. In each insurance category, Black survivors more often reported forgoing care, whereas Hispanics/Latinxs were the least likely to report forgoing care. The racial/ethnic disparity in forgoing care was smaller among the publicly insured compared to the privately insured, with publicly insured Black survivors only forgoing care 2.3% more than their White counterparts. Across all race/ethnicity groups, a higher proportion of those without insurance reported forgoing care, ranging from 45.0% among Hispanics/Latinxs to 60.4% among Blacks. For Hispanics/Latinxs and Whites, the privately insured were least likely to report forgoing care (5.5% and 16.0%), while privately insured Blacks did not forgo care any less than publicly insured Blacks (27.6% vs. 27.4%).
Figure 3.
Observed percentages of study participants who needed but did not seek medical care in the past year due to finances within each of the nine race/ethnicity-insurance strata
Racial/Ethnic Disparities in Forgoing Care by Insurance Status Adjusted for “Predisposing,” “Enabling,” and “Need” Factors
The main model in Table 2 shows the adjusted odds ratios (ORs) and corresponding 95% confidence intervals (CIs) of forgoing care by race/ethnicity-insurance status. The odds of forgoing care, relative to privately insured Whites, was the highest for Blacks and the lowest for Hispanics/Latinxs for each insurance status. For all racial/ethnic groups, publicly insured survivors had similar adjusted odds of forgoing care compared to privately insured Whites. Uninsured Blacks, Hispanics/Latinxs, and Whites had 7.52 (95% CI: 5.09–11.12), 2.98 (95% CI: 1.16–7.64), and 5.43 (95% CI: 4.29–6.87) times higher odds of forgoing care, respectively, compared to privately insured Whites. Although the observed proportion of privately and publicly insured Black survivors forgoing care was equal (Fig. 3), only privately insured Blacks were statistically more likely to forgo care compared to privately insured Whites (OR: 1.82, 95% CI: 1.30–2.54). Accounting for chronic health conditions did not change adjusted race/ethnicity-insurance disparity and self-reported health status estimates (Table S3).
Table 2.
Adjusted odds ratios (ORs) of forgoing care by race/ethnicity and insurance status relative to White survivors with private health insurance
| Race/ethnicity | Insurance | Main model a |
Main model + income + education b |
||
|---|---|---|---|---|---|
| OR (95% CI) | P-value | OR (95% CI) | P-value | ||
| Black | None | 7.52 (5.09–11.12) | <0.001 | 4.41 (2.93–6.64) | <0.001 |
| Public | 1.25 (0.86–1.82) | 0.25 | 0.73 (0.49–1.08) | 0.12 | |
| Private | 1.82 (1.30–2.54) | <0.001 | 1.43 (1.01–2.01) | 0.04 | |
| Hispanic/Latinx | None | 2.98 (1.16–7.64) | 0.02 | 2.18 (0.84–5.66) | 0.11 |
| Public | 0.88 (0.24–3.28) | 0.85 | 0.57 (0.15–2.14) | 0.40 | |
| Private | 0.34 (0.10–1.14) | 0.08 | 0.30 (0.09–1.01) | 0.05 | |
| White | None | 5.43 (4.29–6.87) | <0.001 | 3.53 (2.75–4.55) | <0.001 |
| Public | 1.08 (0.86–1.36) | 0.50 | 0.69 (0.53–0.89) | 0.004 | |
| Private | Reference | Reference | |||
Abbreviations: OR, odds ratio; CI, confidence interval
Adjusted for age at survey completion, sex, treatment era, childhood cancer diagnosis group, treatment modality, major surgery, and self-reported health status.
Adjusted for household income/education in addition to the covariates in the main model.
Additional adjustment for income/education generally attenuated racial/ethnic-insurance disparities in forgoing care (Table 2). However, although attenuated, accounting for the influence of income/education did not fully explain the racial/ethnic disparity observed in the main model among the privately insured. Even within the same income/education levels, privately insured Blacks had 43% higher odds of forgoing care (OR: 1.43, 95% CI: 1.01–2.01), relative to privately insured Whites. The equalizing effect of public insurance observed in the main model was further enhanced in the income/education-adjusted model. All publicly insured race/ethnicity groups became even more protective against forgoing care, relative to privately insured White survivors, within the same income/education levels: publicly insured Blacks, Hispanics/Latinxs, and Whites had 0.73 (95% CI: 0.49–1.08), 0.57 (95% CI: 0.15–2.14), and 0.69 (95% CI: 0.53–0.89) times the odds of forgoing care, respectively. Table S4 shows the same models as Table 2 with changes in the reference group to allow for direct race/ethnicity comparisons within the same insurance status. Table S5 shows the estimates for all covariates in the two models.
Discussion
From our analysis of forgone care among survivors of childhood cancer within the context of race/ethnicity and health insurance, we observed three key findings. First, racial/ethnic disparity in forgoing care exists among privately insured survivors, with Blacks forgoing care the most. Second, although attenuated, this racial/ethnic disparity persisted despite adjustment for income/education, suggesting that it is not solely attributable to differences in socioeconomic status. Third, racial/ethnic disparity in forgoing medical care is absent under public insurance among childhood cancer survivors.
An encouraging observation, the third key finding above, is that public insurance seems to not only alleviate the racial/ethnic disparity in forgoing care, but equalizes all race/ethnicity groups to privately insured Whites. Relative to privately insured Whites, all publicly insured survivors, regardless of race/ethnicity, have similar adjusted odds of forgoing care. Further adjustment for socioeconomic status indicated that, among those with similar income/education, public insurance may actually protect Whites against forgoing care, relative to privately insured Whites; publicly insured Black and Hispanic/Latinx survivors showed the same level of protection, relative to privately insured Whites, although not statistically significant due to their small sample sizes. This is consistent with previous findings that publicly insured survivors utilize survivor-focused healthcare more than privately insured survivors [8, 9]. Public insurance may not only ameliorate racial/ethnic disparities but also appears equal to or better than private insurance in enabling care.
While our analysis is the first to suggest that public insurance reduces racial/ethnic healthcare utilization disparities among survivors of childhood cancer, this phenomenon has been documented in the general population. The passage and partial implementation of the ACA resulted in historic coverage gains for all racial/ethnic groups but particularly Black and Hispanic/Latinx adults [24]. With increases in public insurance accessibility, racial/ethnic disparities in not having a usual source of care or forgoing needed care due to finances have narrowed [25]. Compared to White adults in Medicaid non-expansion states, Black adults in Medicaid expansion states reported better coverage rates, less forgoing care due to finances, and similar utilization [25]. It is noteworthy that this impact of public insurance on racial/ethnic disparity is also clearly observed in the high-risk population of childhood cancer survivors.
The first key finding above suggests that private insurance does not equally benefit all racial/ethnic groups. Privately insured Blacks have nearly twice the adjusted increased odds of forgoing needed care as privately insured Whites, and income/education explains this disparity only partially. The inability of income/education to explain this disparity fully emphasizes that other factors are important contributors to disparity among the privately insured. Underinsurance, which is generally defined as insurance that does not adequately meet an individual’s healthcare-related financial needs [26], is known to be associated with socioeconomic status and forgoing needed care due to finances in childhood cancer survivors [27] as well as in the general population [28], and may provide a partial explanation. Other explanations for Black-White healthcare inequities have been proposed, including interpersonal discrimination, psychosocial burden of perceived racism/distrust, and residential segregation [29]. It is likely that eliminating the Black-White disparity we observed in forgoing care among the privately insured will require the targeting of such barriers jointly with income/education inequities. Indeed, it has been argued that a focus on structural racism, which is “the totality of ways in which societies foster racial discrimination through mutually reinforcing systems of housing, education, employment, earnings, benefits, credit, media, healthcare, and criminal justice,” as a root cause of persisting Black-White disparities in the US healthcare system is required for advances in racial/ethnic health equity [29]. However, if we accept that structural racism had indeed been responsible for the persistent Black-White disparity among the privately insured survivors, would it not also influence the publicly insured survivors? It is not entirely clear why the racial/ethnic disparity observed among the privately insured was not observed under public insurance. However, the ability of public insurance to alleviate and equalize forgoing care to the same or better levels as privately insured White survivors emphasizes that public insurance is an effective intervention for racial/ethnic disparities in forgoing care. Furthermore, the independence of this association from income/education suggests that public insurance addresses causes of disparity not entirely attributable to income/education, implying that public insurance circumvents other barriers.
Survivors of childhood cancer without insurance are known to underutilize healthcare significantly [6–10]. Compared to privately insured Whites, uninsured survivors in our study sample had 2.98–7.52 times the odds of forgoing needed care in the main model. Uninsured survivors had equally high risk of forgoing care regardless of race/ethnicity within this high-risk population of childhood cancer survivors. Provision of public insurance to this highest-risk subgroup is a pressing public health issue for improving their extraordinary level of underutilization of needed healthcare.
Differences in the apparent and adjusted associations of public insurance and forgoing care warrant discussion. Without any adjustment, publicly insured survivors of all racial/ethnic groups reported forgoing care more frequently than privately insured Whites. However, with consideration of “predisposing,” “enabling,” and “need” factors, the adjusted odds of forgoing care between publicly insured survivors of all racial/ethnic groups and privately insured Whites did not differ. In particular, self-reported health status was strongly associated with forgoing care (Table S5); each decrement in self-reported health status was associated with a large increase in the odds of forgoing care; a similar finding was made in a study of predominantly adult cancer survivors [30]. Since publicly insured childhood cancer survivors of all racial/ethnic groups report poorer health status than privately insured White survivors (Figure S1), adjustment for self-reported health status has a significant influence on the association between race/ethnicity-insurance group and forgoing care.
Accounting for cumulative burden of chronic health conditions did not change adjusted race/ethnicity-insurance disparity, consistent with the notion that self-reported health status better reflects survivors’ healthcare needs since this measure considers personal perceptions, adaptability to disease burden, and inclinations towards care-seeking and adherence. Indeed, perceptions of need can differ greatly between individuals with the same underlying conditions. To quote Andersen who pioneered the Healthcare Utilization Model, “Any comprehensive effort to model health services’ use must consider how people view their own general health and functional state, as well as how they experience symptoms of illness, pain, and worries about their health and whether or not they judge their problems to be of sufficient importance and magnitude to seek professional help” [17]. Similarly, perceptions of available healthcare are affected by the dimensions of affordability (income, ability to pay, insurance), acceptability (personal judgements of service/provider suitability, adequacy, or competency), and availability (convenience, transportation, temporal and geographic ease of access) [31]. Many of these factors may be associated with race/ethnicity and could lead to differences in perceived healthcare.
In our analysis, Hispanic/Latinx childhood cancer survivors were substantially less likely than Black and White survivors to forgo care. This result was somewhat unexpected considering that, in the general population, Hispanic/Latinx adults are known to face substantial barriers in accessing healthcare [32]. Compared to Black and White adults, Hispanic/Latinx adults have the highest levels of not having any insurance, forgoing needed care due to finances, and not having a usual source of care [25]. This discrepancy may be due to the similarity of our Hispanic/Latinx survivor population to Whites in socioeconomic status, which is related to forgoing care. The Hispanic/Latinx group had a similar proportion of college graduates/post-graduates, 39.1%, compared to Whites, 37.2%, which contrasts with 2019 US Census data indicating that 18.8% of US Hispanic/Latinx and 40.1% of White individuals have at least a Bachelor’s degree [33]. Overall, this finding warrants further investigation.
This research has several limitations, which need to be considered in interpreting our findings. First, data were collected from a single institution and may not be generalizable to all childhood cancer survivors in the US. Second, the small sample size for Hispanic/Latinx survivors may impact the precision of estimates for this ethnic group. However, with a participation rate of 86.6%, which is comparable to 88.1% of Blacks and 88.0% of Whites [16], participating Hispanic/Latinx survivors are similarly represented as the other race/ethnic groups. Third, the perception of needed care may vary across racial/ethnic groups and socioeconomic status. However, most survivors are comprehensively evaluated during their on-campus clinical assessment as part of their SJLIFE visits; thus, the need for medical care is well established and communicated. Fourth, since poorer minority participants without insurance are less likely to continue participating in survivorship research [34], our analyses may be biased due to differential participation and drop out. Additionally, the data used in the current analysis are cross-sectional and thus we are unable to determine causal relationships. Finally, we did not have data on insurance components such as deductibles, co-insurance, and out-of-pocket maximums that may influence forgoing care.
In summary, race/ethnicity and insurance status were associated with forgoing medical care due to finances despite the need among survivors of childhood cancer. Under public insurance, racial/ethnic disparities in forgoing care relative to privately insured White survivors were absent. Ensuring access to, and needed utilization of, health services is particularly important for adult survivors of childhood cancer in order to address their long-term medical needs associated with late effects of cancer therapy. Efforts towards diminishing disparities across racial/ethnic-insurance groups are pressingly needed, in particular, addressing the disparity in insurance stability (Table S2), a critical problem among the disadvantaged in the US [35], through, for example, enhancing the enrollment information system infrastructure and operating procedures of public health insurance for ease of retention [36].
Supplementary Material
Acknowledgements
The authors thank all individuals who participated in this study.
Funding
This work was supported by the US National Cancer Institute (P30-CA21765 and U01-CA195547) and the American Lebanese Syrian Associated Charities.
Footnotes
Conflicts of Interest/Competing Interests
No conflicts of interest or competing interests were disclosed.
Ethics Approval
The study was reviewed by the St. Jude Children’s Research Hospital Institutional Review Board and ethical approval was obtained on April 25th, 2007. The procedures used in this study adhere to the tenets of the Declaration of Helsinki.
Consent to Participate
All participants provided written informed consent for participation in the St. Jude Lifetime Cohort Study.
Consent for Publication
Not applicable.
Availability of Data and Material
Most data are accessible through the St. Jude Cloud (https://stjude.cloud). A few variables that were used in this paper but are not available on St. Jude Cloud are available upon reasonable request to the corresponding author.
Code Availability
Not applicable.
References
- 1.Ward E, DeSantis C, Robbins A, Kohler B, Jemal A. Childhood and adolescent cancer statistics, 2014. CA Cancer J Clin. 2014;64(2):83–103. 10.3322/caac.21219. [DOI] [PubMed] [Google Scholar]
- 2.Howlader N, Noone A, Krapcho M, Miller D, Brest A, Yu M et al. SEER cancer statistics review, 1975-2017. Bethesda, Maryland, USA: National Cancer Institute; 2020. https://seer.cancer.gov/csr/1975_2017/. [Google Scholar]
- 3.Robison LL, Hudson MM. Survivors of childhood and adolescent cancer: life-long risks and responsibilities. Nat Rev Cancer. 2014;14(1):61–70. 10.1038/nrc3634. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Oeffinger KC, Mertens AC, Sklar CA, Kawashima T, Hudson MM, Meadows AT, et al. Chronic health conditions in adult survivors of childhood cancer. N Engl J Med. 2006;355(15):1572–82. 10.1056/NEJMsa060185. [DOI] [PubMed] [Google Scholar]
- 5.Bhakta N, Liu Q, Ness KK, Baassiri M, Eissa H, Yeo F, et al. The cumulative burden of surviving childhood cancer: an initial report from the St Jude Lifetime Cohort Study (SJLIFE). Lancet. 2017;390(10112):2569–82. 10.1016/S0140-6736(17)31610-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Oeffinger KC, Mertens AC, Hudson MM, Gurney JG, Casillas J, Chen H, et al. Health care of young adult survivors of childhood cancer: a report from the Childhood Cancer Survivor Study. Ann Fam Med. 2004;2(1):61–70. 10.1370/afm.26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Nathan PC, Greenberg ML, Ness KK, Hudson MM, Mertens AC, Mahoney MC, et al. Medical care in long-term survivors of childhood cancer: a report from the Childhood Cancer Survivor Study. J Clin Oncol. 2008;26(27):4401–9. 10.1200/jco.2008.16.9607. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Casillas J, Castellino SM, Hudson MM, Mertens AC, Lima IS, Liu Q, et al. Impact of insurance type on survivor-focused and general preventive health care utilization in adult survivors of childhood cancer: the Childhood Cancer Survivor Study (CCSS). Cancer. 2011;117(9):1966–75. 10.1002/cncr.25688. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Mueller EL, Park ER, Kirchhoff AC, Kuhlthau K, Nathan PC, Perez GK, et al. Insurance, chronic health conditions, and utilization of primary and specialty outpatient services: a Childhood Cancer Survivor Study report. J Cancer Surviv. 2018;12(5):639–46. 10.1007/s11764-018-0700-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Park ER, Kirchhoff AC, Nipp RD, Donelan K, Leisenring WM, Armstrong GT, et al. Assessing health insurance coverage characteristics and impact on health care cost, worry, and access: a report from the Childhood Cancer Survivor Study. JAMA Intern Med. 2017;177(12):1855–8. 10.1001/jamainternmed.2017.5047. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Castellino SM, Casillas J, Hudson MM, Mertens AC, Whitton J, Brooks SL, et al. Minority adult survivors of childhood cancer: a comparison of long-term outcomes, health care utilization, and health-related behaviors from the Childhood Cancer Survivor Study. J Clin Oncol. 2005;23(27):6499–507. 10.1200/jco.2005.11.098. [DOI] [PubMed] [Google Scholar]
- 12.Bhatia S, Gibson TM, Ness KK, Liu Q, Oeffinger KC, Krull KR, et al. Childhood cancer survivorship research in minority populations: a position paper from the Childhood Cancer Survivor Study. Cancer. 2016;122(15):2426–39. 10.1002/cncr.30072. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Liu Q, Leisenring WM, Ness KK, Robison LL, Armstrong GT, Yasui Y, et al. Racial/ethnic differences in adverse outcomes among childhood cancer survivors: the Childhood Cancer Survivor Study. J Clin Oncol. 2016;34(14):1634–43. 10.1200/JCO.2015.66.3567. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Glasziou P, Straus S, Brownlee S, Trevena L, Dans L, Guyatt G, et al. Evidence for underuse of effective medical services around the world. Lancet. 2017;390(10090):169–77. 10.1016/S0140-6736(16)30946-1. [DOI] [PubMed] [Google Scholar]
- 15.Hudson MM, Ness KK, Gurney JG, Mulrooney DA, Chemaitilly W, Krull KR, et al. Clinical ascertainment of health outcomes among adults treated for childhood cancer. JAMA. 2013;309(22):2371–81. 10.1001/jama.2013.6296. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Howell CR, Bjornard KL, Ness KK, Alberts N, Armstrong GT, Bhakta N, et al. Cohort profile: the St. Jude Lifetime Cohort Study (SJLIFE) for paediatric cancer survivors. Int J Epidemiol. 2020. 10.1093/ije/dyaa203. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Andersen RM. Revisiting the behavioral model and access to medical care: does it matter? J Health Soc Behav. 1995;36(1):1–10. 10.2307/2137284. [DOI] [PubMed] [Google Scholar]
- 18.Klassen AF, Anthony SJ, Khan A, Sung L, Klaassen R. Identifying determinants of quality of life of children with cancer and childhood cancer survivors: a systematic review. Support Care Cancer. 2011;19(9):1275–87. 10.1007/s00520-011-1193-x. [DOI] [PubMed] [Google Scholar]
- 19.Hudson MM, Ehrhardt MJ, Bhakta N, Baassiri M, Eissa H, Chemaitilly W, et al. Approach for classification and severity grading of long-term and late-onset health events among childhood cancer survivors in the St. Jude Lifetime Cohort. Cancer Epidemiol Biomark Prev. 2017;26(5):666–74. 10.1158/1055-9965.Epi-16-0812. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Dong H, Robison LL, Leisenring WM, Martin LJ, Armstrong GT, Yasui Y. Estimating the burden of recurrent events in the presence of competing risks: the method of mean cumulative count. Am J Epidemiol. 2015;181(7):532–40. 10.1093/aje/kwu289. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.R Core Team. R: a language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. 2019. https://www.R-project.org/. [Google Scholar]
- 22.RStudio Team. RStudio: integrated development environment for R. Boston, MA, USA: RStudio, PBC; 2020. http://www.rstudio.com/. [Google Scholar]
- 23.Wickham H ggplot2: elegant graphics for data analysis. Springer-Verlag New York; 2016. https://ggplot2.tidyverse.org. [Google Scholar]
- 24.McMorrow S, Long SK, Kenney GM, Anderson N. Uninsurance disparities have narrowed for Black and Hispanic adults under the Affordable Care Act. Health Aff (Millwood). 2015;34(10):1774–8. 10.1377/hlthaff.2015.0757. [DOI] [PubMed] [Google Scholar]
- 25.Baumgartner JC, Collins SR, Radley DC, Hayes SL. How the Affordable Care Act has narrowed racial and ethnic disparities in access to health care: Commonwealth Fund 2020. https://www.commonwealthfund.org/sites/default/files/2020-01/Baumgartner_ACA_racial_ethnic_disparities_db.pdf.
- 26.Lavarreda SA, Brown ER, Bolduc CD. Underinsurance in the United States: an interaction of costs to consumers, benefit design, and access to care. Annu Rev Public Health. 2011;32(1):471–82. 10.1146/annurev.publhealth.012809.103655. [DOI] [PubMed] [Google Scholar]
- 27.Nipp RD, Kirchhoff AC, Fair D, Rabin J, Hyland KA, Kuhlthau K, et al. Financial burden in survivors of childhood cancer: a report from the Childhood Cancer Survivor Study. J Clin Oncol. 2017;35(30):3474–81. 10.1200/JCO.2016.71.7066. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Collins SR, Gunja MZ, Aboulafia GN. U.S. health insurance coverage in 2020: a looming crisis in affordability — findings from the Commonwealth Fund Biennial Health Insurance Survey, 2020: Commonwealth Fund 2020. https://www.commonwealthfund.org/sites/default/files/2020-08/Collins_looming_crisis_affordability_biennial_2020_sb.pdf. [Google Scholar]
- 29.Bailey ZD, Krieger N, Agénor M, Graves J, Linos N, Bassett MT. Structural racism and health inequities in the USA: evidence and interventions. Lancet. 2017;389(10077):1453–63. 10.1016/S0140-6736(17)30569-X. [DOI] [PubMed] [Google Scholar]
- 30.Weaver KE, Rowland JH, Bellizzi KM, Aziz NM. Forgoing medical care because of cost: assessing disparities in healthcare access among cancer survivors living in the United States. Cancer. 2010;116(14):3493–504. 10.1002/cncr.25209. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Tanner EC, Vann RJ, Kizilova E. Consumer-level perceived access to health services and its effects on vulnerability and health outcomes. J Public Policy Mark. 2020;39(2):240–55. 10.1177/0743915620903299. [DOI] [Google Scholar]
- 32.Velasco-Mondragon E, Jimenez A, Palladino-Davis AG, Davis D, Escamilla-Cejudo JA. Hispanic health in the USA: a scoping review of the literature. Public Health Rev. 2016;37(1):31. 10.1186/s40985-016-0043-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.United States Census Bureau. The Hispanic population in the United States, 2019: educational attainment of the population 25 years and over by sex, hispanic origin, and race (Table 5). Current Population Survey, Annual Social and Economic Supplement, 2019. 2020. https://www.census.gov/data/tables/2019/demo/hispanic-origin/2019-cps.html.
- 34.Casillas J, Oeffinger KC, Hudson MM, Greenberg ML, Yeazel MW, Ness KK, et al. Identifying predictors of longitudinal decline in the level of medical care received by adult survivors of childhood cancer: a report from the Childhood Cancer Survivor Study. Health Serv Res. 2015;50(4):1021–42. 10.1111/1475-6773.12282. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Short PF, Graefe DR, Schoen C. Churn, churn, churn: how instability of health insurance shapes America’s uninsured problem: Commonwealth Fund 2003. https://www.commonwealthfund.org/sites/default/files/documents/___media_files_publications_issue_brief_2003_nov_churn__churn__churn__how_instability_of_health_insurance_shapes_americas_uninsured_problem_short_churn_688_pdf.pdf. [PubMed]
- 36.Capoccia V, Croze C, Cohen M, O’Brien JP. Sustaining enrollment in health insurance for vulnerable populations: lessons from Massachusetts. Psychiatr Serv. 2013;64(4):360–5. 10.1176/appi.ps.201200155. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.