Financial Toxicity Among Women with Breast Cancer Varies by Age and Race

Abstract

Introduction.

Financial toxicity negatively affects clinical outcomes in breast cancer. Underrepresented demographics may be at higher risk for financial toxicity. We characterized disparities on the basis of age and other factors.

Patients and Methods.

Surveys completed by women with stage 0–IV breast cancer treated at Memorial Sloan Kettering Cancer Center between 06/2022 and 05/2023 were analyzed. The comprehensive score for financial toxicity (COST) scale was used to assess financial toxicity. Descriptive statistics were calculated for differences in financial toxicity/related factors, and outcomes by age and race. Associations between variables of interest and COST scores were analyzed using linear regression.

Results.

Of 8512 respondents (75% white, 9.3% Asian, 8.4% Black), most (68%) had clinical stage 0/I disease. Stratified by age, young Black women had higher financial toxicity than young white or Asian women (p < 0.001). On multivariable analysis, women age < 45 years experienced higher financial toxicity than older women (coefficient − 2.0, 95% CI − 2.8 to − 1.1, p < 0.001). Compared with white women, financial toxicity was greater among Black (coefficient − 6.8, 95% CI − 7.8 to − 5.8) and Asian women (coefficient − 3.5, 95% CI − 4.4 to − 2.5). Cost-related medication non-adherence was more frequent among Black and Asian women (p < 0.001). Asian women more often paid for treatment with savings than white and Black women (p < 0.001). Young women reported using savings for treatment-related costs more than older (45% vs. 32%); p < 0.001).

Conclusions.

Racial minorities and young patients are disproportionately affected by financial toxicity. Further studies are planned to determine how financial toxicity evolves over time and whether referral to financial services effectively reduces toxicity.

In the USA, costs associated with breast cancer treatment are estimated to be upward of $30 billion¹ and account for more spending than any other cancer care.^2,3 Given increases in cost-sharing between payers and patients, this can translate to a significant financial burden for patients. Out-of-pocket expenses for breast cancer patients totaled $3.14 billion in 2019.⁴ Patient expenditures associated with oncologic treatments compromise access to care and are a source of distress to patients.⁵ This cost-related stress, termed financial toxicity, has been shown to negatively affect quality of life, treatment adherence, and overall survival.⁶

Among those with breast cancer, certain populations may be at especially high risk for financial toxicity.⁷ While there are conflicting data with respect to whether racial minorities have higher degrees of financial toxicity per se,⁷ multiple studies have demonstrated that the financial impact of breast cancer care is more significant for Black women.⁸ These findings may also in part reflect differences in tumor biology. Black women more often present with unfavorable tumor subtypes and more advanced disease, requiring more aggressive therapies,⁹ which are associated with higher treatment costs. Age may also influence the relationship between race and financial toxicity. Although adults younger than 39 years are disproportionately affected by economic strain associated with cancer treatment in general,¹⁰ the influence of age on financial toxicity in breast cancer specifically, has not been well studied.

The goal of this study was to understand how age and race influence financial toxicity among women with breast cancer treated at a tertiary cancer care center using the comprehensive score for financial toxicity (COST) scale.¹¹

PATIENTS AND METHODS

Design, Subjects, and Recruitment

After obtaining Memorial Sloan Kettering Cancer Center (MSK) (New York, NY, USA) Institutional Review Board (IRB) approval, we conducted a pilot survey study¹² to assess financial toxicity during cancer treatment. Beginning in June 2022, surveys assessing financial toxicity and need for referral to patient financial services were administered to individuals receiving treatment for cancer at MSK, including both new patients and those who were actively being treated or being seen in follow-up. Patients with COST scores < 20 on initial screening and who expressed interest on the basis of survey responses were referred to MSK Patient Financial Services, our institution’s financial counseling and assistance program. Subsequent to initial screening, participants were requested to complete the survey again every 3–4 months through MSK Engage, our institution’s web-based platform for monitoring patient-reported outcomes.

Women diagnosed with clinical stage 0–IV breast cancer between June 2022 and May 2023 were included in this analysis. Registry patients were recruited for financial toxicity screening through emails generated 1 week prior to their scheduled medical, surgical, or radiation oncology appointment via MSK Engage. Patients who did not complete the survey prior to their appointment were able to do so at their appointment either independently on an iPad, or with the assistance of a nurse in clinic. Informed consent was not obtained, as this was an operational pilot study with emphasis on referring patients to financial services experiencing financial toxicity; however, permission was obtained from the MSK IRB to analyze the data presented herein.

Survey Components, Data Abstraction, and Outcomes of Interest

Financial toxicity was assessed using the comprehensive score for financial toxicity (COST) scale, a validated metric for oncology patients 18 years of age and older.¹¹ COST comprises 11 items interrogating financial distress within the 7-day period preceding survey administration. Each item is scored using a 5-point Likert scale (0–4, with 0, experiencing the item not at all and 4, experiencing the item very much). A composite score, including specific questions that are reverse-scored, is summed resulting in a numerical output from 0 to 44, with higher scores indicating better financial well-being and lower scores indicating more severe financial distress.^13,14

Additional items were included in the survey to complement COST data and to elucidate sources of financial strain. These non-COST items were investigator-designed by the MSK Affordability Working Group with input from the MSK Patient and Community Education division and the MSK Immigrant Health and Cancer Disparities Service. Patients were asked to report on a scale of 1–5 how often they took less medication than prescribed due to concerns over prescription cost (1, never; 5, all the time) and whether they had to use their savings to pay for treatment costs (yes/no). Desire to be referred to MSK Patient Financial Services was assessed by a single yes/no question. Age, insurance status, and clinical disease stage at time of diagnosis were abstracted retrospectively from chart review.

Statistical Analysis

Descriptive statistics were used to characterize cohort demographics using medians and interquartile ranges for continuous variables, and counts and percentages for categorical variables. Continuous variables were compared between races and age groups using the Wilcoxon rank-sum test or the Kruskal–Wallis test, and categorical variables were compared using Fisher’s exact test or the chi-squared test. While the American Association for Cancer Research classifies adolescent and young adults with cancer as those 15–39 years of age,¹⁵ our institution expands this criterion to include women younger than 45 years of age. Therefore, we considered age as a binary variable: < 45 years or 45 years and older. Insurance status was categorized as government-based only (Medicaid and/or Medicare), private only, or mixed. COST scores were treated as a continuous outcome.

Linear regression analysis was used to determine associations between explanatory variables of interest and financial toxicity in the generalized estimating equation (GEE) framework, which accounts for repeated measures. The identity link function was used with a correlation structure of “independence,” which was determined by minimizing the quasi-likelihood information criterion. Items that demonstrated significance on univariate analysis at a prespecified type I error rate of 0.05 were included as covariates in a multivariable linear regression model. The final model was selected using a backward selection algorithm. All analyses were conducted using R 4.3 (R Foundation for Statistical Computing, Vienna, Austria).

RESULTS

Of the 35,400 patients from the MSK tumor registry who completed survey screening for financial toxicity, 11,865 had a diagnosis of breast cancer. Of the 8512 women included in this study, 6369 (75%) identified as white, 791 (9.3%) as Asian, and 714 (8.4%) as Black. The majority of patients were of non-Hispanic ethnicity [7249/8512 (92%)]. Of the 1125/8512 (13%) who were < 45 years of age, 110 (9.7%) identified as Black, 163 (14.5%) as Asian, and 736 (65.4%) as white. Tables 1 and 2 present cohort characteristics by race and age, respectively. Compared with white and Asian women, Black women were more often insured by Medicare and/or Medicaid alone [43/590 (7.3%) versus 30/641 (4.7%) of Asian women and 125/4897 (2.6%) of white women; p < 0.001]. Women < 45 years of age were primarily privately insured [824/910 (91%)]; insurance coverage of women 45 years of age or older was more evenly distributed between private insurance alone [3360/5721 (59%)] and Medicare/Medicaid with private supplement [2142/5721 (37%)] (p < 0.001). Black women more frequently presented with higher-stage disease than white or Asian women (p < 0.001). Higher disease stages were also more common among young women (p < 0.001) overall.

TABLE 1.

Cohort characteristics stratified by race; categorical data presented as n (%) and continuous data as median (interquartile range)

Characteristic	White n = 6369	Asian n = 791	Black n = 714	Other n = 638	p *
Age					< 0.001
<45 years	736 (12%)	163 (21%)	110 (15%)	115 (18%)
≥ 45 years	5633 (88%)	628 (79%)	604 (86%)	523 (82%)
Clinical stage					< 0.001
0	688 (11%)	110 (14%)	76 (11%)	61 (9.8%)
I	3638 (58%)	411 (53%)	338 (48%)	340 (54%)
II	1234 (20%)	165 (21%)	189 (27%)	141 (23%)
III	353 (5.7%)	54 (7%)	64 (9.1%)	49 (7.8%)
IV	326 (5.2%)	36 (4.6%)	34 (4.9%)	34 (5.4%)
Unknown	130	15	13	13
Initial COST score	30 (22–37)	26 (17–34)	21 (11–30)	24 (15–31)	< 0.001
Unknown	569	50	72	70
Referral to financial services	286 (20%)	63 (27%)	110 (45%)	56 (31%)	< 0.001
Unknown	4972	559	469	458
Medication non-adherence**					< 0.001
Never	5074 (92%)	548 (82%)	470 (78%)	460 (86%)
Rarely	254 (4.6%)	71 (11%)	58 (9.7%)	36 (6.7%)
Sometimes	134 (2.4%)	42 (6.3%)	51 (8.5%)	29 (5.4%)
Often	26 (0.5%)	3 (0.4%)	12 (2%)	7 (1.3%)
All the time	11 (0.2%)	7 (1%)	8 (1.3%)	2 (0.4%)
Unknown	870	120	115	104
Use of savings	1742 (32%)	290 (44%)	198 (33%)	183 (34%)	< 0.001
Unknown	936	125	113	105
Insurance					< 0.001
Government	125 (2.6%)	30 (4.7%)	43 (7.3%)	31 (6.2%)
Private	2992 (62%)	482 (75%)	370 (63%)	337 (68%)
Mixed	1780 (36%)	129 (20%)	177 (30%)	129 (26%)
Unknown	1472	150	124	141

COST comprehensive score for financial toxicity scale

^*Statistical significance analyzed by Pearson’s chi-squared test, Kruskal–Wallis rank-sum test, or Fisher’s exact test

^**On initial survey

TABLE 2.

Cohort characteristics stratified by age; categorical data presented as n (%) and continuous data as median (interquartile range)

Characteristic	Age < 45 years n = 1125	Age ≥ 45 years n = 7393	p *
Stage			< 0.001
0	78 (7.1%)	868 (12%)
I	487 (44%)	4241 (59%)
II	332 (30%)	1400 (19%)
III	126 (11%)	394 (5.4%)
IV	79 (7.2%)	352 (4.9%)
Unknown	23	148
Initial COST score	25 (15–33)	29 (20–36)	< 0.001
Unknown	35	726
Referral to financial services	117 (32%)	398 (24%)	0.001
Unknown	754	5709
Medication non-adherence**			0.8
Never	889 (89%)	5666 (90%)
Rarely	61 (6.1%)	42 (0.7%)
Sometimes	41 (4.1%)	215 (3.4%)
Often	6 (0.6%)	42 (0.7%)
All the time	3 (0.3%)	25 (0.4%)
Unknown	125	1087
Use of savings	448 (45%)	1968 (32%)	< 0.001
Unknown	125	1156
Insurance			< 0.001
Government	10 (1.1%)	219 (3.8%)
Private	824 (91%)	3360 (59%)
Mixed	76 (8.4%)	2142 (37%)
Unknown	215	1672

COST comprehensive score for financial toxicity scale

^*Statistical significance analyzed by Pearson’s chi–squared test, Kruskal-Wallis rank-sum test, or Fisher’s exact test

^**On initial survey

Financial Toxicity and Secondary Outcomes of Interest

Financial toxicity scores were lower among Black women than white or Asian women (Table 1) and among young women compared with older women (Table 2). Only Black women had COST scores with an interquartile range low enough to be considered as having severe financial toxicity. The same racial differences in financial toxicity scores were seen in each age stratum (Table 3). Using savings to pay for treatment-associated costs was most common among Asian women (Table 1). Black women more frequently reported taking less medication than prescribed because of financial hardship compared with white or Asian women (Table 1). There were no age-based differences regarding need to take less medication secondary to inability to pay for prescriptions (p = 0.8), but a significantly higher proportion of young women [448/1000 (45%)] used their savings for treatment-related costs compared with older women [1968/6237 (32%); p < 0.001]. Significantly more Black women desired referral to patient financial services [110/245 (45%) vs. 63/232 (27%) of Asian women and 286/1397 (20%) of white women; p < 0.001]. Young women requested referral to patient financial services more frequently than women 45 years of age and older [117/371 (32%) vs. 398/1684 (24%); p = 0.001].

TABLE 3.

Financial toxicity stratified by race and age; comprehensive score for financial toxicity scale scores reported as median (interquartile range); for patients who completed multiple surveys, the median score was used

	Age < 45 years				p *
	White n = 736	Asian n = 163	Black n = 110	Other n = 115
COST score	27 (17–34)	24 (15–33)	18 (8–28)	22 (13–32)	< 0.001
	Age ≥ 45 years				p *
	White n = 5633	Asian n = 628	Black n = 604	Other n = 523
COST score	30 (22–37)	26 (18–34)	21 (13–30)	24 (16–31)	< 0.001

^*Statistical significance analyzed by Wilcoxon rank-sum test

Factors Associated with Financial Toxicity

On univariate analysis, younger age, non-white race, Hispanic ethnicity, and higher disease stages were associated with greater financial toxicity (Table 4). On multivariable analysis, age 45 years or older was associated with higher financial toxicity (coefficient − 2.0, 95% CI − 2.8 to − 1.1, p < 0.001). Although Black race was associated with the greatest financial toxicity (coefficient − 6.8, 95% CI − 7.8 to − 5.8), Asian women also had worse financial well-being relative to white women (coefficient − 3.5, 95% CI − 4.4 to − 2.5). Invasive cancer and Hispanic ethnicity remained independently associated with experiencing more financial toxicity (Table 4). Although interaction terms were evaluated to assess whether the association of race on financial toxicity varied by age or whether the impact of clinical stage on financial toxicity varied by race or age, these interaction terms were not significant and therefore not included in the final model.

TABLE 4.

Univariate and multivariable analysis of factors associated with financial toxicity as measured by the Comprehensive Score for Financial Toxicity (COST) scale

Variable	Univariate			Multivariable
	Coefficient	95% CI	p	Coefficient	95% CI	p
Age (years)			< 0.001			< 0.001
≥ 45	–	–		–	–
< 45	− 3.3	− 4.1 to − 2.5		− 2.0	− 2.8 to − 1.1
Race			< 0.001			< 0.001
White	–	–		–	–
Asian	− 3.4	− 4.3 to − 2.5		− 3.5	− 4.4 to − 2.5
Black	− 7.6	− 8.5 to − 6.6		− 6.8	− 7.8 to − 5.8
Other	− 4.4	− 5.4 to − 3.3		− 2.3	− 3.6 to − 1.1
Ethnicity			< 0.001			< 0.001
Non-Hispanic	–	–		–	–
Hispanic	− 6.3	− 7.3 to − 5.2		− 5.3	− 6.4 to − 4.2
Clinical stage			< 0.001			< 0.001
0	–	–		–	–
I	− 1.4	− 2.3 to − 0.60		− 1.7	− 2.5 to − 0.85
II	− 4.6	− 5.6 to − 3.7		− 4.2	− 5.2 to − 3.3
III	− 5.0	− 6.3 to − 3.6		− 4.3	− 5.6 to − 3.0
IV	− 7.4	− 8.8 to − 6.0		− 7.2	− 8.6 to − 5.7
Insurance type*			< 0.001
Government	–	–
Private	0.36	− 1.5 to 2.2
Mixed	3.1	1.3–5.0

CI confidence interval

^*Insurance dropped out of the final MVA model through backward selection

DISCUSSION

In this study of patient-reported financial distress after a diagnosis of breast cancer, Black and Asian women, and those younger than 45 years of age, had significantly worse financial toxicity than white women and those 45 years of age or older, respectively. Compared with white women, Black and Asian women also more frequently reported that expense of treatment influenced adherence to medication. Young and Asian women more frequently used their savings to pay for treatment. These data demonstrate that the expense of oncology care has behavioral ramifications that are an important facet of disparities in adherence to recommended care.

Existing data reporting the association between race and financial hardship resulting from breast cancer treatment are mixed. While some studies report race-based differences in the financial impact of treatment,⁸ this may not necessarily translate to a causal relationship between minority race and financial toxicity.^7,16 In their single-institution cross-sectional analysis of COST survey data from 568 women between 3 and 18 months after surgical management of breast cancer, Corkum et al. found that race-based differences in financial toxicity on univariate analysis were excluded from multivariable analysis during stepwise selection after adjusting for other factors (e.g., area deprivation index).¹⁷ In contrast, our findings, which are consistent with previously published literature demonstrating that racial minorities experience higher rates of health-related financial toxicity,^17–20 may reflect differences in sample population. Institutional practices may have contributed to selection bias; prior to 2022, patients with Medicaid only comprised a small percentage of patients treated at MSK. In addition, while items to assess social determinants of health are being incorporated in current iterations of MSK Engage surveys, surveys included in this study did not adequately capture this information. Forthcoming investigations will adjust for these factors.

The timeframe within which COST surveys are administered may further complicate the relationship between sociodemographic variables and financial toxicity. The COST scale assesses financial toxicity on the basis of patients’ experiences in the 7-day period preceding survey completion. As timing of COST surveys differs in existing studies and often occurs many months to year(s) after a cancer diagnosis, literature may reflect the durable effect of financial hardship beyond the treatment period itself. In their study of 988 cancer survivors, Hastert et al. reported that compared with white women, a higher proportion of Black patients experienced financial hardship and were in debt after their cancer treatment.²¹ Patients with breast cancer comprised 44% of the study population, for whom median time from diagnosis was 20.2 months. Our study includes survey responses collected at multiple timepoints after diagnosis and treatment, making our results somewhat generalizable.

The association between race and financial toxicity may have important implications for clinical outcomes. In this study, we found race-based differences in how cost can restrict treatment adherence. This is consistent with mediation analysis from others indicating that limiting care due to cost contributes to worse quality of life, and that this adverse financial outcome is more profound for Black patients than white patients.²¹ While a number of factors contribute to worse clinical outcomes among racial minorities with breast cancer, inconsistent oncologic treatment compromises quality of care and may provide an explanation for increased mortality associated with more severe financial toxicity.²² Insofar as disparate access to resources, financial services, and quality care may be manifestations of patient interactions with our healthcare system, they remain modifiable.

In addition to race, age has been shown to impact quality and access to care. Although < 10% of annual breast cancer diagnoses are made in adolescent and young adult women,²³ the American Association for Cancer Research identifies these patients as an underserved demographic that is particularly vulnerable to poor clinical outcomes.¹⁵ Further, adolescent and young adult patients comprise the largest cohort of oncology patients without insurance and may be more likely to experience financial strain, and consequently, lapses in care.²⁴ Even among those who are insured in the demographic, there are unique challenges; a greater proportion of young patients have high-premium coverage that may result in increased financial burden as a result of cost-sharing.²⁵ Additionally, young adults more commonly have low health-insurance literacy that may contribute to inability to effectively and appropriately utilize coverage.²⁶ While we did not observe age to have a moderating effect on the association between race and financial toxicity, being younger than 45 years of age at time of diagnosis was an independent risk factor for financial toxicity. Our data are consistent with that of others showing the significant financial burden of cancer treatment for young patients.²⁷ Worse clinical outcomes in breast cancer have been associated with cost-related undertreatment in young women, and in racial and ethnic minorities.^28,29 A higher degree of financial toxicity was seen in these populations in our study in spite of differences in insurance coverage. Our data indicated higher rates of government-based insurance among Black women compared with white and Asian women, but nearly all of the younger women in our sample had commercial insurance. Other investigations highlight that having insurance may ameliorate financial burden but does not eliminate it; even in young women with insurance, the high out-of-pocket cost associated with medications has been shown to contribute to non-adherence to endocrine therapy.²⁷ Treatment non-adherence may be one factor in worse clinical outcomes that drive up future costs of care. Investigations exploring interventions to mitigate financial toxicity and how these affect long-term clinical outcomes are forthcoming.^30–32

Although awareness of financial toxicity, and its consequences for morbidity and mortality are increasing among providers, challenges to mitigating financial hardship require further investigation. Financial assistance programs are often available at many cancer centers, but these are underutilized.³³ Interview-based research from gynecological oncology patients has identified lack of awareness, perceptions of ineligibility, fear of negative consequences, and feelings of being overwhelmed as reasons for their underuse.³⁴ From the provider side, clinician engagement remains inadequate.³⁵ In their study, Jagsi et al. found that both younger age and non-white race were associated with patient interest in discussing the influence of breast cancer treatment on finances.¹⁶ In concordance with these data, we demonstrated that a higher proportion of Black and young women vocalized desire for referral to patient financial services. Further studies are planned to determine whether referred patients utilized these services, and whether utilization mitigates financial toxicity during and after treatment.

One strength of our study is its reporting of data for Asian women specifically, in contrast to the preponderance of published investigations, which focus on financial toxicity among Black women. Though by no means exhaustive, by expanding the racial demographics considered, we highlight the importance of considering factors that moderate the association between race and financial toxicity. We can speculate on the basis of our findings that Asian women may have savings that they are able to utilize to pay for treatment costs, but their lack of insurance might offset their ability to avoid financial hardship. Black women, on the contrary, may contend with financial needs independent from their cancer diagnosis, as evidenced by a smaller proportion of Black women having private insurance in our study. Another strength is our determination of financial toxicity early after diagnosis, allowing for evaluation of baseline cost-related distress; this is also in contrast to the majority of existing studies, which query financial toxicity near the end or after completion of breast cancer treatment. Such early evaluation is important because subsequent decisions regarding treatment can themselves be influenced by cancer-related financial burden.^36,37

Our study also has several limitations. The classification of insurance type as government, private, or mixed may simplify complexities of out-of-pocket costs. Differences in prescription coverage, in particular between Medicaid and Medicare, may have diluted the relative contribution of insurance status to financial toxicity among young and Black women. Although this repeated-measures analysis incorporates data from surveys administered iteratively throughout the study periods, the number of participants with complete responses at subsequent timepoints after the initial screening survey may have been insufficient to adjust for baseline COST score. Given the lack of a predefined threshold for clinically meaningful difference in scores, statistically significant differences may have been secondary to robust sample size and may not reflect clinical importance. Further studies are necessary to demonstrate how distress from the cost of care varies over time. As the COST scale queries patients on the basis of the preceding 7-day period, there may be more appropriate metrics to evaluate long-term financial toxicity. Although we adjusted for clinical stage as a surrogate for chemotherapy, additional patient and tumor characteristics are necessary to fully understand how demographics and disease features influence cost of treatment. We did not have access to data regarding patient income or education level, both of which have been shown to influence financial toxicity. While an important aspect of our study is documenting disparities in financial navigation and referral to financial services, data regarding utilization of these services and their efficacy in alleviating financial toxicity are lacking. These are, however, active areas of investigation for the authors. Finally, the single-institution, tertiary care setting of this study may limit generalizability.

CONCLUSIONS

The financial burden of breast cancer treatment can cause distress that compromises quality of life and clinical outcomes. Black, Asian, and younger women experienced more severe financial toxicity than hite or older women in our study. Acknowledging that racial minorities and adolescent and young adult patients may be disproportionately impacted is the first step in designing interventions that mitigate cost-associated hardship. Further studies are necessary to elucidate how financial toxicity evolves over time, and whether interventions such as physician engagement or referral to financial services are effective.