Financial Hardship by Age at Diagnosis Including in Young Adulthood among African American Cancer Survivors

Financial hardship is common among cancer survivors (1), particularly those with fewer financial resources at the time of diagnosis (2). By some estimates, approximately 50% of cancer survivors experience financial hardship related to cancer (1). This can take the form of material financial hardship (MFH), including direct out-of-pocket medical expenses and indirect costs such as missed work and lost income; behavioral financial hardship (BFH), or coping mechanisms survivors employ to manage the costs of care, such as limiting treatment due to cost; and psychological financial hardship (PFH), or feelings of distress or worry about the costs of cancer (1). Financial hardship is associated with not getting all recommended treatments due to cost concerns (2, 3), not receiving recommended follow-up screenings (4), lower health-related quality of life (5, 6), and higher mortality (7).

African American adults are at particular risk for financial hardship due to cancer. On average, household incomes are 37% lower among African American compared with white adults (8), and median net worth among African American households was only 13% that of white households in the United States in 2019 (9). Previous work has consistently reported higher prevalence of financial hardship among African American compared with white cancer survivors (2, 5, 10–12).

Young adults (YA) in the United States have the lowest levels of wealth, job security, and employment-provided benefits of any age group, placing them at risk for financial hardship due to cancer as well (13). A cancer diagnosis can also affect YAs' ability to continue to attend school or work and has implications for their long-term career development and potential earnings (14), as well as future access to employer-sponsored health insurance (15). Previous work suggests that YAs with cancer experience greater financial hardship, including excess medical expenditures and productivity losses (16) and cost-related medication nonadherence (17), than their cancer-free peers; however, it is not clear whether financial hardship is more common among YA cancer survivors (ages 20–39 at diagnosis) compared with those diagnosed at later ages.

Younger age at diagnosis has been identified as a risk factor for financial hardship among cancer survivors, although these findings typically compare working-age with older adults (often ages <65 vs. 65+; refs. 18–20). Little is known about the financial outcomes of cancer survivors who were YAs relative to middle aged and older at diagnosis, despite dramatic differences in financial resources by age in the United States (21). In addition, despite financial hardship being more common among African American than white cancer survivors as a whole (2, 5, 10–12), we are unaware of existing work describing the financial effect of cancer on African American YA survivors.

The purpose of this study is to estimate associations between age at diagnosis, including a separate category for YA survivors (ages 20–39 at diagnosis compared with survivors diagnosed at 40–54, 55–64, and 65–79), and MFH, BFH, and PFH among African American cancer survivors. We hypothesized that more YA survivors will report MFH and BFH, and that PFH will be more severe among YA relative to older survivors.

Detroit Research on Cancer Survivors (ROCS) is a population-based cohort designed to understand medical, behavioral, financial, and psychosocial outcomes among African American cancer survivors in Metropolitan Detroit (22, 23). Survivors were eligible to join the cohort if they were diagnosed with invasive breast, colorectal, lung, or prostate cancer between the ages of 20 to 79 since January 1, 2013, or with endometrial cancer between the ages of 20to 79 or any other invasive cancer between the ages of 20 to 49 since January 1, 2016, and identified through the Metropolitan Detroit Cancer Surveillance System (MDCSS) and self-reported African American or Black race. At ROCS enrollment, participants completed a baseline interview online, over the phone, or by completing and returning a paper survey. Participants received a $25 gift card for enrollment survey completion. The Institutional Review Board at Wayne State University approved this research.

Results include cross-sectional data from the first 3,888 ROCS participants with available data, including 662 who completed the first version of the ROCS questionnaire, and 3,226 who completed a revised questionnaire with additional measures including the Comprehensive Score for financial Toxicity (COST), measuring PFH. Demographic and cancer-related characteristics of participants completing each survey version are presented in Supplementary Table S1. ROCS enrollment is ongoing as of July 2021. Results described here reflect responses to the initial survey from March 2015 to June 2017 and the revised survey from June 2017 through April 2021. At the time the analytic dataset was created, the overall response rate was approximately 40%, excluding those pending recruitment (i.e., still eligible for further study invitations). Response was highest among breast cancer survivors, women, and those ages 50 to 59 at diagnosis.

ROCS participants self-reported demographic, socioeconomic, and cancer-related factors including sex, educational attainment, household income, marital status, employment status, health insurance coverage, treatments received, and treatment status. Additional information including age at diagnosis, cancer site, stage, and date of diagnosis was obtained via linkage with MDCSS. All measures reflect responses at the time of ROCS enrollment unless otherwise noted.

Participants were considered to have experienced MFH if they reported a decrease in income, using assets, borrowing from family or friends, or experiencing lasting debt related to cancer care. Decreases in income were identified by an item asking whether participants' income changed since diagnosis, with options for increase, decrease, or no change. Survivors were asked whether, in order to pay bills related to their cancer treatment, they had to: (i) refinance or take out a second mortgage on their home, (ii) sell their home, (iii) sell stocks or other investments, or (iv) withdraw money from retirement accounts. Participants who answered in the affirmative to any of these items were considered to have used assets to pay for cancer care. Borrowing and debt were assessed by items asking whether they or any member of their family had to borrow money from friends or other family members to help pay for cancer treatment; and whether they were currently in debt due to expenses related to their cancer at the time of ROCS enrollment.

Participants also had the opportunity to specify “other” forms of MFH. After review and adjudication by two authors (T.A. Hastert and J.J. Ruterbusch), participants who indicated other responses consistent with MFH measures were considered to have experienced MFH. The most common other forms of MFH included using personal assets not specified in the ROCS survey (e.g., selling a vehicle or pawning possessions), bankruptcy, and receiving donations or charity assistance.

Three items assessed whether participants experienced BFH: (i) “Was there a time in the past 12 months when you needed to see a doctor but could not because of cost?”; (ii) “Did you ever turn down treatments (chemotherapy, radiation, pain medications, anti-nausea medications, anti-diarrhea medications, or other recommended cancer treatments) because you were concerned about the cost?”; and (iii) “Did you ever skip doses of prescribed medication to save money?”

Although the MFH and BFH measures have not been validated against objective measures of financial outcomes and forgone care; they are based on previously-used measures of financial hardship (24) and items from the Behavioral Risk Factor Surveillance System (25) and have demonstrated expected associations with health-related quality of life (5).

PFH was measured using version 2 of the COST measure of financial wellbeing (26, 27). The COST measure includes 11 statements (e.g., “I feel financially stressed”) and participants are asked to indicate how much each statement applies to the past seven days with options ranging from “not at all” to “very much.” COST has demonstrated excellent internal consistency (Cronbach α = 0.92) and appropriate convergent validity with psychological distress and income (27). Score values range from 0 to 44, and responses are coded so that higher scores indicate better financial wellbeing (28).

All statistical analyses were completed using SAS version 7.4, and figures were drawn using R software (29). The distribution of select clinical and demographic variables was summarized by age group at diagnosis (20–39, 40–54, 55–64, and 65–79) using counts and percentages. We compared the proportion of survivors reporting MFH and BFH by age group using a general chi-square test. COST score distributions were summarized by box plots and scores were compared using analysis of variance. We estimated prevalence ratios (PR) and 95% confidence intervals (CI) for each MFH and BFH category by age group with 65 to 79 as the reference using modified Poisson regression with a log-link function and robust standard errors (30, 31). Adjusted least squares COST score means were calculated for each age group from a multivariable linear regression model. To account for multiple testing, P values < 0.002 should be considered statistically significant to preserve an overall type I error rate of 0.05.

All multivariable regression models were adjusted for covariates selected a priori based on the authors' knowledge of predictors of financial hardship and associations between age and socioeconomic resources, and included sex, Surveillance, Epidemiology and End Results (SEER) summary stage, marital status, employment status, whether participants were on disability, type of health insurance, cancer treatments received, treatment status at ROCS enrollment, and months from diagnosis to ROCS enrollment, using categories in Table 1. We observed higher reported prevalence of MFH in survivors ages 40 to 64 in interviewer-administered versus self-administered surveys (Supplementary Table S2) and also include mode of survey completion in these models. Cancer site varied widely by age and was not included in the models; however, we include results limited to breast cancer survivors as a sensitivity analysis of the association between age at diagnosis and financial hardship limited to the same form of cancer. Because income at ROCS enrollment is potentially on the pathway through which age at diagnosis can affect financial outcomes, it was not included in the adjusted models.

Table 1 displays participant demographic, socioeconomic, and cancer-related characteristics, overall and by age group. Women accounted for 55% of participants. Most participants completed at least some college. A plurality of participants in each age category reported household incomes of less than $20,000 per year; however, this lowest income level was most common among YA survivors (47%) and least common among survivors ages 65 to 79 at diagnosis (36%). A total of 37% of participants were married or living with a partner, and this varied from 29% of YA survivors to 40% of survivors ages 65 to 79. Just over half of YA and middle-aged participants reported being employed full or part time compared with 11% of survivors ages 65 to 79. Medicaid was the most common form of health insurance coverage among YA survivors, whereas Medicare was most common among survivors ages 65 to 79, and private insurance was most common among survivors ages 40 to 64.

Cancer-related factors also varied by age group (Table 1). Breast and prostate cancer each accounted for more than one-third of participants, colorectal and lung each accounted for 11%, and other sites accounted for 8%. Lymphoma (N = 26), kidney (N = 22), thyroid (N = 25), and myeloma (N = 15) were the most common other cancers. Breast cancer accounted for 54% of cancers among YA survivors, and sites other than breast, colorectal, lung, or prostate accounted for 37%. Overall, 24.4 months elapsed between diagnosis and ROCS enrollment.

Table 2 presents prevalence of overall and specific forms of MFH and BFH, and Fig. 1 presents prevalence of MFH and BFH by age group. Overall, 48% of ROCS participants reported any MFH, including a majority of both YA survivors (72%) and those ages 40 to 54 (62%) compared with 49% of 55- to 64-year-olds and 33% of those ages 65 to 79 (P_trend < 0.001; Fig. 1A). Income decreases were the most common form of MFH (30%) followed by debt (19%), borrowing (11%), and using assets (5%). Age at diagnosis was inversely associated with all forms of MFH.

Approximately one-quarter (23%) of ROCS participants reported BFH. BFH was most common among YA survivors (26%) and least common among those 65 to 79 (20%; P_trend = 0.089; Fig. 1B). Needing a doctor and not going due to cost was the most common form of BFH (14%), followed by skipping doses of prescribed medication to save money (9%) and not receiving recommended treatment due to cost (6%). Nearly twice as many YA survivors as those ages 65+ reported needing a doctor and not going due to cost (18% vs. 10%; P = 0.001). Prevalence of other forms of BFH did not differ by age.

Table 3 gives adjusted PRs and 95% CIs of overall and specific forms of MFH and BFH by age. Compared with survivors ages 65 to 79 at diagnosis, prevalence of any MFH was 31% higher in those ages 55 to 64 (95% CI, 16%–47%), 63% higher in those ages 40 to 54 (95% CI, 43%–86%), and 75% higher in YA cancer survivors (95% CI, 49%–104%; P_trend < 0.001). Decreases in income, debt due to cancer, and borrowing from friends and family were more than twice as common among YA survivors compared with those ages 65 to 79 (PRs between 2.26 and 2.33, all 95% CIs exclude 1.0) and exhibited inverse dose–response associations with age at diagnosis.

A similar, but less pronounced, age-related trend was evident for BFH. Compared with survivors ages 65 to 79 at diagnosis, prevalence of BFH was 22% to 26% higher among those ages 40 to 64, and 50% higher among YA cancer survivors (95% CI, 8%–108%; P_trend = 0.019). Although age at diagnosis was not associated with refusing treatment due to cost or skipping doses of prescribed medication to save money, prevalence of needing a doctor and not going due to cost was 85% (95% CI, 21%–184%) higher among YA and approximately 40% higher among middle aged (ages 40–64) compared with older survivors. These results should be interpreted cautiously in the context of multiple testing.

Table 3 also presents PRs of any MFH or BFH associated with age at diagnosis stratified by household income (<$20,000 vs. $20,000 or more). The association between age at diagnosis and MFH did not differ by household income. Age at diagnosis was not associated with BFH among survivors with household incomes below $20,000/year (P_trend = 0.31); however, among survivors with household incomes of at least $20,000, prevalence of BFH was 2.41 (95% CI, 1.43–4.06) times as high among YA survivors compared with those ages 65 or older at diagnosis (P_trend < 0.001).

Associations between age at diagnosis and MFH and BFH among breast cancer survivors were largely similar to associations for all cancers combined (Table 3).

Figure 2 presents COST score means and interquartile ranges by age group, and Table 4 gives least squares COST score means by age. Mean COST score was 24.5 overall and inversely associated with age at diagnosis (P_trend < 0.001). Compared with survivors ages 65 to 79 at diagnosis, COST scores were 2.4 (95% CI, 1.4–3.4) points lower among those 55 to 64, 5.2 (95% CI, 3.9–6.4) points lower among survivors ages 40 to 54, and 6.5 (95% CI, 4.5–8.4) points lower among YA cancer survivors in an adjusted model. Results were similar in a model limited to breast cancer survivors.

We estimated prevalence of MFH, BFH, and PFH among African American cancer survivors by age at diagnosis, including YA cancer survivors in their own category. Consistent with our hypothesis, overall MFH and BFH were inversely associated with age, and prevalence of each was higher among YA survivors than those between the ages of 65 and 79 at diagnosis. PFH as measured by the COST score was also inversely associated with age, and YA cancer survivors reported lower financial wellbeing than any other age group. These findings were largely similar when analyses were limited to breast cancer survivors, suggesting that the observed associations are not due primarily to differences in cancer site by age.

Younger age has consistently been identified as a risk factor for financial hardship after cancer (2, 18–20, 24, 32); however, earlier work often compared outcomes between survivors <65 with those 65+ (18–20). This categorization appropriately highlights the fact that working-age survivors face financial challenges related to their reliance on employment as a primary source of income and insurance that are less common in older survivors (33); however, YAs in particular face financial instability that is not captured using this approach.

In 2018, median household incomes in the United States were lowest among adults <25 ($33,389), and were more than twice as high among 25 to 44-year-olds ($68,817; ref. 8); however, this coincides with a time when YAs experience demands on their financial resources related to starting and raising families, purchasing homes, and paying for their education. These incomes do not translate into greater wealth for YAs—median net worth of households headed by an adult <35 was $11,000, compared with $224,000 among adults ages 65 to 74 in 2019, and home ownership, an important contributor to household wealth, is lowest for YAs (<35) and has declined steadily among both adults <35 and 35 to 44 in recent years (21). Substantial disparities in wealth exist by race, with median wealth among white families estimated at $188,200 compared with $24,100 among African American families in 2019 (9). African Americans <35 represent a particularly financially vulnerable group with median wealth of only $600 compared with $25,400 among white YAs (9).

Despite YAs' financial constraints, very little is known about how financial hardship in YA cancer survivors compares with those in older groups. Ramsey and colleagues reported risk of bankruptcy was 3.6 times as high for adults ages 20 to 34 and 3.1 times as high for adults ages 35 to 49 relative to adults ages 65 to 79 (32). Shankaran and colleagues reported adjusted ORs of MFH and BFH of 52.5 (95% CI, 10.3–267.4) and 8.9 (95% CI, 1.1–70.9), respectively, comparing colorectal cancer survivors <50 with those 75+; however, this study included only 55 survivors under the age of 50, resulting in imprecise estimates of association (24). Banegas and colleagues reported ORs comparing survivors ages 18 to 44 with those 55 to 64 of 2.07 for borrowing or going into debt and 1.81 for bankruptcy (33). The result for borrowing or debt is roughly consistent with our estimated PRs of 2.26 for borrowing and 2.39 for debt. The populations reflected in each of these studies were 83% to 88% white, with Ramsey and colleagues reporting 12% nonwhite participants without further detail, and Shankaran and Banegas each reporting Black survivors accounted for only approximately 2% of their study populations, potentially further limiting the comparability of these findings with the results presented here (24, 32, 33).

Comparing outcomes among YA survivors to those in the oldest age group at diagnosis results in the largest differences in financial hardship prevalence; however, all three financial hardship outcomes were also more common among those ages 40 to 54 and 55 to 64 than 65 and older at diagnosis. This is likely due at least in part to the relatively greater financial resources available to older (vs. younger) adults, as well as the availability of Medicare coverage. The role of specific assets and resources in the financial outcomes of cancer survivors should be examined more carefully in future work.

Our finding of much higher prevalence of MFH among YA versus older African American survivors is particularly striking given that previous work has consistently reported higher levels of financial hardship among African American than white cancer survivors (2, 5, 10–12). This is consistent with lower incomes and wealth in African American compared with white adults (8, 9, 34, 35); however, almost no information is available about financial outcomes among YA African American survivors. Disparities have been reported in treatments received (36), comorbidities (37), and survival by race and neighborhood socioeconomic status (38–40); however, we are unaware of any previous estimates of financial hardship among African American YA survivors.

Our observed overall associations between age and BFH are weaker than those for MFH and PFH, and we observed effect modification such that age at diagnosis was associated with BFH among survivors with incomes of at least $20,000, but not among those with incomes <$20,000. This could be because access to Medicaid coverage allows survivors to get recommended care but may not be enough to prevent other financial consequences of cancer. Notably, this work was conducted in Michigan, which expanded Medicaid in 2014. Medicaid was the most common form of insurance reported by YA survivors (51%). Despite the strong inverse association observed between age and MFH and PFH, outcomes in ROCS may be better than in states where cancer survivors with limited financial resources are not able to access Medicaid coverage.

Compared with survivors ages 65 to 79, fewer YA ROCS participants reported owning assets including a home (23% vs. 58%), a car (48% vs. 68%), stocks (5% vs. 12%), or having a savings account (40% vs. 56%); however, more YA survivors still reported using assets to pay for cancer care (8%) than any other age group (4%–7%; P_trend = 0.001). In addition, specifying other forms of MFH not included in the ROCS questionnaire was most common in YA survivors (6% vs. 2% for all age groups). The most common written responses included selling a vehicle, pawning possessions, and receiving charitable donations. To the extent that other work has not included similar measures or allowed them to be expressed, prevalence of MFH in YAs may be underestimated. YA and working-age cancer survivors should be included in the development of future work describing financial outcomes after cancer in these groups to ensure it accurately reflects their experiences.

Strengths of this study include its use of a population-based cohort of African American cancer survivors, a group that is commonly underrepresented in cancer survivorship work. The detailed ROCS questionnaire allowed for the inclusion of several measures of MFH and BFH, and a validated scale as a measure of PFH (27), as well as several important potential confounding factors. Open-ended responses to the financial hardship questions also allowed us to identify other potential forms of financial hardship not specifically delineated in the questionnaire. The large sample size, including oversampling of survivors under the age of 50 at diagnosis, allowed us to calculate PRs of several forms of MFH and BFH comparing YA survivors with those in older groups.

Potential limitations should be considered when interpreting our findings. Because of the different incidence rates of specific types of cancer by age and the design of the ROCS cohort, the site distribution varies substantially by age in this population. This has implications for the types of treatments survivors received and their related costs and comorbidities; however, this mix of cases roughly represents the most common cancers, and the potential subsequent financial hardship associated with them, in younger versus older adults. Findings from analyses limited to breast cancer survivors are similar to our overall findings, suggesting observed associations are not due to differences in cancer site by age. Although the focus on outcomes among African American survivors provides important insight into the experiences of a population that is often underrepresented in related research, and a YA population that is particularly financially vulnerable, it may limit the generalizability of our findings to other populations. Although participants were asked to indicate whether they used any of several assets or took on any of several forms of debt related to cancer and prompted to report additional forms of debt and asset use, participants could have experienced additional forms of material and behavioral hardship not specifically included. On average, just over 2 years elapsed between cancer diagnosis and ROCS enrollment. In many cases, this provides sufficient time for survivors to have experienced financial hardship and to report on it; however, this also means that the sickest patients with cancer and those diagnosed with the most severe cancers are likely underrepresented in this work. In addition, the COST score was originally designed to capture the financial wellbeing of patients undergoing current treatment. Its validity among longer-term survivors is not known. Our findings could further be subject to recall bias.

The results described here suggest that younger age at diagnosis, and particularly diagnosis in young adulthood, represents a strong risk factor for financial hardship among African American cancer survivors when compared with those diagnosed at older ages. This is consistent with YAs' limited financial resources relative to older adults; however, to our knowledge, this has not been directly measured in previous work. Young adulthood at diagnosis should be considered a risk factor for cancer-related financial hardship and addressed in work designed to reduce the adverse financial impacts of cancer. Future work is needed to understand these associations in more diverse groups of YA cancer survivors and to understand how specific factors (e.g., assets, insurance, employment) may protect younger cancer survivors from the financial consequences of cancer.

T.A. Hastert reports grants from the American Cancer Society and NCI during the conduct of the study. J.J. Ruterbusch reports grants from NCI during the conduct of the study. A.S. Wenzlaff reports grants from NIH during the conduct of the study. A.G. Schwartz reports grants from NCI during the conduct of the study. No disclosures were reported by the other authors.

T.A. Hastert: Conceptualization, writing–original draft, writing–review and editing. J.J. Ruterbusch: Conceptualization, data curation, formal analysis, visualization, writing–original draft, writing–review and editing. J. Abrams: Supervision, methodology, writing–review and editing. M. Nair: Resources, writing–review and editing. A.S. Wenzlaff: Data curation. J.L. Beebe-Dimmer: Resources, project administration. S.S. Pandolfi: Data curation, writing–review and editing. A.G. Schwartz: Resources, supervision, funding acquisition, project administration, writing–review and editing.

T.A. Hastert was supported by funding from the American Cancer Society (MRSG-17-019). A.G. Schwartz, J.L. Beebe-Dimmer, and T.A. Hastert were supported by funding from the NCI (all three supported by U01CA199240, A.G. Schwartz and J.L. Beebe-Dimmer by P30CA022453, and additional support from HHSN261261201300011I to A.G. Schwartz). A.G Schwartz also supported this work through funding from the General Motors Foundation and the Karmanos Cancer Institute.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.