Financial Toxicity Among Patients With Breast Cancer Worldwide

1Financial toxicity (FT) is the negative impact that the monetary burden of medical care can have on patients’ well-being, fiscal security, and overall health. Contributors to FT include direct medical, direct nonmedical, and indirect costs. Direct medical costs include hospitalizations, pharmaceutical bills, laboratory charges, and outpatient treatment–related expenses. Direct nonmedical and indirect costs include transportation, food and lodging, and expenses linked to disease-related employment loss. Nonmedical and indirect expenses are not generally covered by insurance and financial aid programs and contribute to out-of-pocket costs (OOPCs). Catastrophically high OOPCs are known to negatively impact medical care and psychosocial well-being.

2Compared with other chronic conditions, patients with cancer are at risk for higher OOPCs.,, Breast cancer care in particular may be associated with high FT given the need for screening and diagnosis, multidisciplinary care, and longitudinal follow-up; notably, gender also affects financial security., However, the financial burden of breast cancer care worldwide is not well-characterized. Most studies examining breast cancer–related FT focus on high-income countries (HICs), with little attention given to low- and middle-income countries (LMICs)., Poverty is a major barrier to accessing care in LMICs, where patients often bear a substantial portion of the financial costs of treatment while also facing greater poverty. Patient financial concerns are associated with delayed breast cancer diagnoses and higher mortality. Establishing the global extent of FT and comparing the economic burden of disease in different populations is imperative to help policy makers prioritize funding of breast cancer care infrastructure.

3To address this gap, we quantified the FT rate among patients with breast cancer in high-, middle-, and low-income countries through a meta-analysis. The comparison of FT across countries with different economic and health care landscapes may shed light on the efficacy of varying health care financing strategies and social safety net systems on FT. Developing this understanding is the first step to creating effective interventions that not only reduce financial stress on low-income patients but also increase access to care and improve breast cancer outcomes worldwide.

4We conducted a systematic review and meta-analysis of studies exploring FT in individuals with breast cancer. This study did not require institutional review board approval because it is a review article using publicly available data sources. Informed consent was not needed because no patient data were used, in accordance with 45 CFR §46. A protocol for this review was developed using the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) reporting guideline and registered in PROSPERO (CRD42021228015).

5Original, full-text, English-language articles about FT related to breast cancer screening, diagnosis, and treatment were included. Records focusing on cost to the health system, rather than to individuals, or lacking stratified breast cancer data were excluded. Abstracts, editorials, commentary, reviews, and articles without full text were excluded.

6Four databases—Embase (Elsevier), PubMed, Global Index Medicus, and Global Health (EBSCO)—were queried from inception to February 2021. A robust search strategy was developed by the medical librarian (M.B.) and the lead author (C.A.W.). Relevant articles for the concepts “breast cancer” and “out of pocket finances/financial stress” were retrieved (eAppendix in Supplement 1). All records were imported into Covidence for screening.

7A standardized data extraction tool was developed and validated by 2 authors (C.A.W. and A.M.) independently extracting a subset of records to ensure consistency and minimize bias. Deduplication was conducted manually and using Covidence algorithms. Two authors (C.A.W. and A.M.) independently performed title and abstract review of all records, followed by full-text review of remaining records. Conflicts were resolved by third author (K.R.). The primary outcome was rate of FT among patients with breast cancer.

8Multiple definitions of FT were used across the studies. To synthesize heterogeneous data, quantitative and qualitative categories were established (Table 1). Quantitative methods included rate of patients experiencing FT (18 studies).,,,,,,,,,,,,,,,,, Qualitative methods included validated patient-reported outcome measures (PROM; 10 studies),,,,,,,,,, patient-reported severity of FT rated by the patient on a 3-point or 5-point scale (7 studies),,,,,,, qualitative interviews (4 studies),,,, and other (2 studies)., A study falling into multiple categories was analyzed in all categories.

9Data analysis was performed from March to December 2022. Rate of FT, as defined and reported by each study, was used where available. In studies that measured patient-reported severity only, ratings were reported on a 3-point (1 = minimal, 2 = moderate, 3 = severe) or 5-point (1 = none, 2 = a little bit, 3 = somewhat, 4 = quite a bit, 5 = a lot) scale. To calculate a FT rate from these studies, we defined FT as a moderate to severe rating (score of 2-3 on a 3-point scale or 4-5 on a 5-point scale), because these cutoffs were used consistently across studies to determine FT., The percentage population exceeding the cutoff determined FT rate in these studies.

10Statistical combination of the individual studies was performed to obtain a pooled estimate of the FT rate among patients with breast cancer in HICs and LMICs. We used a random-effects model by the DerSimonian and Laird method to account for heterogeneity between studies. Visual representation was achieved with a forest plot. The FT rate along with 95% CIs were calculated for quantitative estimation of FT and converted to percentages for better representation.

11Heterogeneity was assessed using 2-sided χ2 test for homogeneity and I2 index for variability. Cutoff points for heterogeneity in the I2 index were as follows: I2 value less than 25%, low heterogeneity; 25% to 75%, moderate heterogeneity; and greater than 75%, high heterogeneity. If P <.10, the presence of heterogeneity was considered in the χ2 test. To account for heterogeneity, subgroup analysis was performed among HIC studies for different study quality ratings, because the number of studies within LMICs was limited (4 studies).

12Study quality was assessed using a preexisting risk of bias tool. Fourteen criteria were independently assessed by 2 authors (C.A.W. and A.N.E.), and an overall quality rating was determined, corresponding to good, fair, or poor. Small study effects were assessed visually using funnel plot and statistically using Begg and Egger tests. The trim-and-fill method was used to account for any potential small study bias, given this was the source of asymmetry in the funnel plot. Statistical analysis was performed using Comprehensive Meta-Analysis statistical software version 4 (Biostat, Inc).

13A total of 11 086 citations were retrieved. After deduplication, 9005 remained and were screened. Following title and abstract screening, 462 articles underwent full-text assessment, and 34 articles were included (Figure 1).

14The sample size of included studies ranged from 5 to 2445 people (Table 1). All studies were published between 2008 and 2021. The majority were undertaken in HICs (24 studies), largely in the US (19 studies), followed by Canada (3 studies), South Korea (1 study), and Denmark (1 study). A total of 10 studies were from LMICs, including China (2 studies), India (2 studies), Haiti (1 study), Kenya (1 study), Nigeria (1 study), Egypt (1 study), Iran (1 study), and 1 study examining multiple Southeast Asian countries. Of the 34 studies, 26 were cross-sectional, 7 were prospective cohort, and 1 was retrospective cohort. Table 1 and Table 2 show additional details of the study characteristics.

15Most studies (20 studies [58.8%]) reported health insurance status. Universal public insurance was available in China, South Korea, Kenya, and Canada. A mix of private and public insurance was seen in the US and Iran. In India and Southeast Asia, few patients had insurance, and most paid entirely out-of-pocket. Income was reported in most studies (24 studies [70.6%]). Reporting was heterogeneous, with highly variable definitions of the lowest reported income bracket. Definitions ranged from descriptive including “being below the poverty line for the country” or “low income,” to quantifiable including earning less than 25% of mean national household income or earning less than a specific number (ranging from <$12 000 US to <$50 000 US annual household income). Overall, 24.3% fell into the lowest income bracket. Among US studies reporting race,,,,,,,,,,,,,,,,, an average of 73.7% of patients were White. Race was not reported in other countries. All studies evaluated FT during diagnosis and treatment for breast cancer, and none did so during screening.

16Many studies reported a FT rate (18 studies); there was significant heterogeneity in the definition of FT across these studies. Five studies had specific numerical criteria for defining FT. This included medical cost exceeding 40% of household capacity to pay or potential income (3 studies),,, OOPC exceeding 30% of annual household income (1 study), and moving more than 1 decile down the national income distribution (1 study). Four studies used PROM instruments, further described later.,,, Five studies used patient self-identification based on subjective financial difficulty criteria. Here, FT was defined as an affirmative answer to having financial difficulty, trouble paying medical bills, or having to pay more for medical care than is affordable,, or if patients brought up financial burden as a theme in qualitative interviews. In 4 studies, FT was defined according to patients’ report of specific, objective financial consequences of care, including losing income or a job; change in marital status; having to borrow money or go into debt; having difficulty paying for food, rent, or transportation,,; or having to forgo any type of medical treatment because of cost.

17Ten studies used validated PROM instruments. Of these, 4 calculated a FT rate. These 4 studies used the Comprehensive Score for Financial Toxicity (COST), though with varying cutoff scores to define FT (2 studies),, the InCharge Financial Distress/Financial Well-Being Scale (1 study),, and the Cota Patient Assessed Symptom Score-7 instrument (1 study). The COST is a validated instrument with a score range of 0 to 44, with lower scores associated with greater FT. Economic burden scores were calculated using the Breast Cancer Finances Survey or another validated tool containing questions about crucial aspects of cancer-related economic burden.,

18Six studies,,,,, used patient-reported severity of financial burden rated on a 3-point or 5-point scale. These studies asked patients to rate the degree of financial impact (1 study), difficulty (2 studies), or burden (3 studies) posed by breast cancer care. Four qualitative studies,,, offered representative quotations describing patient experiences, including high OOPCs, financial worry, and negative impact on patient and family financial well-being. Two other studies, used varied methods. Pezzin et al defined FT as a rating of somewhat difficult or very difficult on a question about difficulty paying medical bills when comparing 2 medication groups and compared the odds of FT among individuals in 2 treatment groups. Gany et al reported the impact of financial consequences on several quality of life measures, such as food insecurity, financial assistance, and savings.

19Eighteen studies,,,,,,,,,,,,,,,,, were eligible for meta-analysis, 14 from HICs and 4 from LMICs. The rate of FT among patients with breast cancer was 35.3% (95% CI, 27.3%-44.4%; I2 = 95.95%; P for heterogeneity, <.001) in HICs and 78.8% (95% CI, 60.4%-90.0%; I2 = 94.78%; P for heterogeneity, <.001) in LMICs (P <.001 for HICs vs LMICs) (Figure 2). A subgroup analysis based on study quality ratings was performed among 14 studies focusing on FT in HICs. The rate of FT among patients with breast cancer was 49.9% (95% CI, 34.7%-65.1%; I2 = 96.0%; P for heterogeneity, <.001) in studies rated fair in quality assessment and 28.5% (95% CI, 20.2%-38.6%; I2 = 95.7%; P for heterogeneity, <.001) in studies rated good in quality assessment (P <.001 for fair vs good quality assessment) (eFigure 1 in Supplement 1).The funnel plot visually showed slight asymmetry toward the right of the pooled point estimate for HIC studies; however, both Begg rank correlation test and Egger linear regression test indicated no small study bias. Conditioning on small study being the source of asymmetry, the trim-and-fill method revealed an overall imputed rate estimate of 42.7% (95% CI, 33.7%-52.3%), which was not materially different from 35.0% (95% CI, 27.0%-45.9%) (eFigure 2 in Supplement 1).

20FT in studies without a reported or calculatable FT rate varied. The full COST was reported in 4 studies. Mean COST ranged from 21.2 in Jin et al to 28.64 in Asaad et al, and the median COST ranged from 22 in Wan et al to 25 in Williams et al. One study used a 4-question subset of the COST scored out of 16, with a mean (SD) score of 6.2 (4.8) 1 year after breast surgery. An alternate PROM used in 2 studies, was an economic burden score, with a mean of 2.25 to 2.94 economic burden items per patient at various times after diagnosis. The qualitative studies used interviews with patients (3 studies),, and/or breast cancer assistance programs representatives (1 study) to understand FT impact. One study using odds ratios to compare FT between 2 adjuvant treatment groups found that patients receiving aromatase inhibitors, compared with those receiving tamoxifen, experienced higher FT rates (odds ratio, 1.4), although absolute FT rates were not reported. The study assessing quality of life measures found that 47.5% of patients were food insecure, 66.0% needed financial assistance, 34.0% used savings to pay for treatment, and 41.2% lacked savings altogether.

21Race and FT had no clear association.,,,,,,,,,,,,,,,,, Patients’ education level,,,,,,,,,,,,,,,,,,,,,,,,,,, or insurance status,,,,,,,,,,,,,,,,,,, were related to the income status of the country. The employment status of the patient was related to age, with older patient populations having more retirees, as well as to prevalence of homemakers compared with women working outside the home.,,,,,,,,,,,,,,,,, No clear association was found between these factors and FT. Few studies reported comorbidities (8 studies),,,,,,, or urban vs rural status (6 studies),,,,,; thus, conclusions could not be drawn. Cancer stage and treatments were extremely heterogeneous across studies; no clear association was found between either factor and FT.

22This systematic review and meta-analysis found a high rate of FT among individuals undergoing breast cancer treatment globally. Overall, 35.3% of patients in HIC studies and 78.8% of patients in LMIC studies experienced FT. To our knowledge, this study is the first to systematically characterize the global burden of FT. Typical FT rates across all health conditions in LMICs range from 6% to 12%.,, Our study demonstrates that FT prevalence among patients with breast cancer is substantially higher, with vulnerability compounded in LMICs.

23A high global burden of FT due to malignant neoplasms is unsurprising given the need for expensive, multidisciplinary treatment at specialized facilities, prolonged follow-up, transportation costs, and time off work. This burden has often been associated with adverse clinical outcomes, including mortality., However, the financial burden of breast cancer is even greater than those for other cancers, with substantially higher OOPC than colorectal, lung, and prostate combined. At a systems level, the Lancet Oncology Commission demonstrated that breast cancer was the most expensive cancer in the US in 2010, accounting for US $16.5 billion, or 13% of all cancer-related spending., Individually, direct medical costs of breast cancer care range up to US $100 000.

24Gaps remain in determining the cause of increased costs and FT for breast cancer specifically. In HICs, this may be attributable to many novel and costly cancer care interventions, services overutilization, increased willingness to pay, and varying insurance coverage. Individuals in LMICs encounter delayed diagnoses due to limited access to screening and high-quality diagnostic services, leading to higher prevalence of later-stage diagnoses requiring more extensive treatments. Lower baseline income, limited insurance coverage, and greater travel distances for specialized cancer care also contribute., The high incidence of FT among patients with breast cancer, especially in LMICs and compared with other diseases, presents an important target for cancer control policies and programs. Because higher FT is associated with worse patient outcomes, addressing FT is an important intervention to reduce the substantial, growing rates of breast cancer mortality in LMICs. Highest priority patient populations have low baseline education and socioeconomic status, lack health insurance, and live in low-resource areas. These populations are not exclusive to LMICs, but also exist in low-resource HIC areas. On the basis of our findings, we recommend 4 potential intervention areas to reduce FT and improve outcomes among patients with breast cancer: targeted educational campaigns to raise awareness about the signs and symptoms of breast cancer and the importance of early diagnosis and treatment; expansion of health care coverage to minimize direct medical OOPC; programs to assist with direct nonmedical and indirect costs, such as transportation to and lodging near treatment centers, childcare, and other family expenses incurred as a result of treatment; and funneling resources toward improving screening, referral, and treatment infrastructure for breast cancer care. Our data further suggest that there are lower FT rates in countries with universal health coverage, demonstrating efficacy of this health care policy approach. However, there are no comparative studies, and conclusive outcomes cannot be determined with existing data.

25Our study highlights the need for more studies assessing FT. There are limited LMIC data, and most HIC data are from the US; fully understanding the burden in different regions and contexts is integral to executing potential solutions. There is also a clear need to standardize FT measurement to improve generalizability across studies. Definitions for FT ranged from patient reports to a single survey question to numerical expenditure-based thresholds. Other studies used patient-reported experiences of FT without providing a quantifiable estimate. We recommend a combined approach using both qualitative and quantitative variables to measure FT. Quantitatively, FT can be measured as catastrophic health expenditure, defined using the total OOPC as a percentage of household income or discretionary (nonfood) expenditure. The most widely used metric of FT is medical cost greater than 40% of total household nonfood expenditure. Qualitative measures include validated PROMs on food scarcity, joblessness, and childhood education.

26Our study has important limitations. We were limited to English-language articles because of personnel available to conduct the study; however, we still captured a wide range of international studies. There was significant heterogeneity between studies in FT definition and demographic data reporting; in addition, income, rural vs urban status, and FT definitions were not reported uniformly. Nevertheless, we explored country income and study quality as sources of heterogeneity when possible. Other inconsistently reported variables were employment status, stage of breast cancer, comorbidities, and treatments. Consistent reporting of these variables is necessary to examine their potential contribution to FT and facilitate geographic comparisons. FT patterns and interventions vary with health care systems, and in LMICs particularly, financial planning is integral to treatment planning and national goals. Our study did not account for different health care systems or control for health care–dedicated gross domestic product. Our calculation, therefore, is an important global overview of FT among patients with breast cancer and represents a starting point for future analyses.

27FT is a substantial burden of cancer treatment among patients with breast cancer. The incidence of breast cancer–related FT is 78.8% in LMICs and 35.3% in HICs. The definition and measurement of FT varied across studies, and reporting of covariates was inconsistent. Further efforts, ideally led by transnational organizations, are required to thoroughly characterize FT burden in a standardized fashion across different settings. Policy interventions to improve breast cancer education, expand health care coverage, enhance patient support for nonmedical costs, and invest in health care infrastructure could improve both financial and health outcomes among patients with breast cancer at risk for FT.