Financial Costs and Burden Related to Decisions for Breast Cancer Surgery

PURPOSE:

Financial toxicity is a well-recognized adverse effect of cancer care, yet little is known about how women consider treatment costs when facing preference-sensitive decisions for breast cancer surgery or how surgical treatment choice affects financial harm. We sought to determine how financial costs and burden relate to decisions for breast cancer surgery.

METHODS:

Women (≥ 18 years old) with a history of breast cancer were recruited from the Army of Women and Sisters Network to complete an 88-item electronic survey. Descriptive statistics and regression analysis were used to evaluate the impact of costs on surgical decisions and financial harm after breast cancer surgery.

RESULTS:

A total of 607 women with stage 0 to III breast cancer were included. Most were white (90%), were insured privately (70%) or by Medicare (25%), were college educated (78%), and reported household incomes of more than $74,000 (56%). Forty-three percent underwent breast-conserving surgery, 25% underwent mastectomy, 32% underwent bilateral mastectomy, and 36% underwent breast reconstruction. Twenty-eight percent reported that costs of treatment influenced their surgical decisions, and at incomes of $45,000 per year, costs were prioritized over breast preservation or appearance. Overall, 35% reported financial burden as a result of their cancer treatment, and 78% never discussed costs with their cancer team. When compared with breast-conserving surgery, bilateral mastectomy with or without reconstruction was significantly associated with higher incurred debt, significant to catastrophic financial burden, treatment-related financial hardship, and altered employment. Among the highest incomes, 65% of women were fiscally unprepared, reporting higher-than-expected (26%) treatment costs.

CONCLUSION:

Cancer treatment costs influenced decisions for breast cancer surgery, and comparably effective surgical treatments differed significantly in their risk of patient-reported financial burden, debt, and impact on employment. Cost transparency may inform preference-sensitive surgical decisions and improve patient-centered care.

INTRODUCTION

In the United States, approximately 300,000 women are diagnosed with breast cancer each year, resulting in health care costs that are estimated to reach $20 billion by 2020.^1-3 Increasing deductibles, copayments, and premiums have resulted in cancer treatment costs being shifted from the health care system to patients themselves.^4-6 This financial hardship is now recognized as a major adverse effect of cancer care and has been associated with reduced quality of life, nonadherence, and an increased risk of early mortality.^7-10 Overall, 16% to 78% of cancer survivors report experiencing treatment-related financial hardship.^11-13 In 2009, the American Society of Clinical Oncology formally acknowledged the magnitude of this problem and encouraged oncologists to discuss the costs of care with individuals embarking on treatment.¹⁴

Eligible women with breast cancer can choose between breast-conserving surgery (BCS) with radiation or mastectomy, and women are increasingly undergoing contralateral prophylactic mastectomy despite a lack of evidence demonstrating additional survival benefit.^3,15-17 Long-term clinical trial data and contemporary observational series have established comparable recurrence rates and overall survival across these surgical options.^2,3,18-21 Therefore, decisions for breast cancer surgery are highly preference sensitive.^22,23 Although the physical and emotional adverse effects of surgical treatment are routinely discussed in this context (ie, appearance, complications, length of recovery), the potential for financial harm is not explicitly addressed.^24-26

Numerous studies have characterized how women make decisions for breast cancer surgery; yet, in an era where patients increasingly shoulder a greater proportion of their health care expenses, little is known about how cost information may be factored into cancer treatment decisions.^26,27 Thus, we sought to determine how women consider cancer treatment costs when making decisions for breast cancer surgery and how these decisions relate to financial hardship.

METHODS

Study Population

We conducted a Web-based survey of women (≥ 18-years old) living in the United States with a prior breast cancer diagnosis. After approval from our institutional review board (protocol No. 00074944; December 29, 2016 to December 29, 2019) and the Scientific Advisory Committee, participants were recruited from the Dr. Susan Love Research Foundation’s Army of Women (AOW) and the Sisters Network of North Carolina from January 20, 2017 to June 16, 2017. Women received a call-to-action e-mail providing an overview of the study.²⁸ The AOW is a coalition of volunteers, both with and without a history of breast cancer, that aims to promote stakeholder participation in breast cancer research.²⁹ In an effort to improve diversity within our study cohort, the Sisters Network of North Carolina was approached for participation. The Sisters Network is a National African American Breast Cancer Survivorship Organization.³⁰

Survey Design

In collaboration with patient advocates, we developed an 88-item questionnaire including previously validated measures as well as novel, pretested questions developed for the purpose of this study.^31-34 Sociodemographic, clinical, and treatment information were self-reported by participants, including age, race/ethnicity, comorbidities, education and income level, insurance status, cancer stage at diagnosis, and the time of survey completion.

Survey items assessed women’s experiences with treatment-related costs and their consideration of health care costs when making decisions for breast cancer surgery. Additional items were included to determine how the costs of treatment compared with other factors commonly considered in surgical decision making for breast cancer, including recovery time, complications of treatment, sensory loss, fear of recurrence, future surveillance, opinions from the medical team, sexuality, appearance of the breast, and need for radiotherapy. Specific questions were developed to capture the estimated proportion of costs covered by participants’ insurance, out-of-pocket costs, and personal or family debt related to breast cancer care. Finally, survey items queried participants about their overall financial well-being, health behaviors, employment, and personal sacrifice as a result of their cancer care.³³ Estimated time of survey completion was approximately 30 minutes. The final survey was programmed using Qualtrics software (Qualtrics, Provo, UT).

Data Analysis

Descriptive statistics were used to summarize the sample. The continuous outcome—reported out-of-pocket costs—was described using medians and quartiles. Wilcoxon rank sum tests evaluated differences across surgery types, and subsequent pairwise testing compared other surgical groups with the BCS group. Categorical outcomes were described with percentages and binomial CIs. Specifically, the importance factors were defined by responses on a 1 to 7 scale, with 1 being not at all important and 7 being extremely important. Somewhat to extremely important included scores of 5 to 7. Associations with categorical outcomes were tested using χ² or Fisher’s exact with pairwise testing performed where appropriate. Income data were collected as a categorical range and then further collapsed. For modeling, income was categorized into three groups on the basis of the data distribution: $74,000 or more, $74,001 to $125,000, and more than $125,000. Experienced financial burden across surgery types was of particular interest; thus, this variable was dichotomized into none to somewhat versus significant to catastrophic and tested for differences with additional pairwise comparisons. For all pairwise tests, a Bonferroni adjustment was applied. All bivariable analyses were tested using χ² or Fisher’s exact test and Wilcoxon rank rum tests.

The five surgery types were collapsed into three broader categories (bilateral mastectomy [BM] with or without reconstruction, mastectomy with or without reconstruction, and lumpectomy) and compared using a χ² test. Twenty-five versions of the data set were created using the Markov chain Monte Carlo method to impute missing data.³⁵ After multiple imputation for missing variables, multivariable logistic regression was performed on this outcome using the collapsed surgery categories, age at diagnosis, stage, race, income group, insurance type, insurance coverage, chemotherapy receipt, radiation receipt, whether patients discussed costs, and time since diagnosis. All statistical analyses were performed in SAS 9.4 (SAS Institute, Cary, NC).

RESULTS

Participant Baseline Characteristics

At the time of the study initiation, all 109,168 members of the AOW members received a call-to-action e-mail notification of the survey, including individuals with a history of breast cancer and those without. By convention, the 1,414 individuals who responded to the survey link were considered invited participants. Of those, 970 (69%) agreed to participate.^28,36 To improve study diversity, 50 women were approached from the Sisters Network of North Carolina, and 23 (46%) responded. Overall, 1,464 women collectively received detailed information about the study, of whom 993 (68%) agreed to participate. Our response rate was consistent with other surveys reported through the AOW in recent years. The final study cohort consisted of 654 eligible women who completed the survey in its entirety. Of those, 47 patients reported not having undergone surgery, presumably related to active surveillance, stage in treatment, or diagnosis of de novo metastatic disease. Overall, 607 patients underwent initial surgery for stage 0 to III breast cancer, who comprised the analytic sample for the remainder of the reported data. Baseline demographic, clinical, and treatment characteristics of participants are listed in Table 1. The median age was 49.6 years (range, 26 to 76 years). The majority had early-stage breast cancer, with 85.7% reporting stage 0 to II disease at diagnosis. Median time from diagnosis was 6.7 years (0.1 to 37.1 years). Self-reported race/ethnicity was 90% white, 5% black/African American, 1% Asian, 1% American Indian or Alaskan Native, 2% multiracial, and 0.5% preferred not to answer. The distribution of reported insurance coverage was 70% private, 25% Medicare, and 5% Medicaid. Forty-two percent were actively receiving treatment at the time of survey completion. Overall, 60% underwent chemotherapy, and 65% underwent radiation. Twenty-six percent of women reported an annual household income more than $125,000, and 45% reported a graduate-level education.

TABLE 1.

Baseline Demographic, Clinical, and Treatment Characteristics (N = 607)

Cancer Treatment Costs Related to Decisions for Breast Cancer Surgery

Overall, 43% (95% CI, 39% to 47%) of women reported that they considered costs when making breast cancer treatment decisions. When asked specifically which variables influenced their decisions for breast cancer surgery, women reported prioritizing fear of recurrence (85% reporting somewhat to extremely important; 95% CI, 82% to 88%) and advice from the medical team (95% reported as somewhat to extremely important; 95% CI, 93% to 97%). Approximately one third (28%; 95% CI, 25% to 32%) of the cohort reported that treatment costs influenced their decisions for breast cancer surgery, with half of that group specifically describing costs as extremely important. Costs increased in importance as annual household income fell, with 16% of participants with an income of more than $125,000 per year reporting costs as somewhat to extremely important versus 41% of participants with an income of $74,000 per year (P < .01). Among women in the lowest income brackets (≤ $45,000/ year), cost was more influential on decisions for breast cancer surgery than loss of sensation, breast preservation or appearance, need for long-term surveillance, and avoiding radiation. Despite this, 78% of women (95% CI, 75% to 82%) reported never discussing cancer treatment costs with their medical team. Thirty-four percent of respondents reported costs were higher than expected, whereas 38% reported having no expectations at all. Among those with the highest incomes, those values were 26% and 39%, respectively. On bivariable analysis, later stage, lower percentage of insurance-covered health care, nonwhite race, less education, and discussing costs with the medical team were associated with participants reporting costs as somewhat to extremely important in surgical decisions (all P values < .05).

Median patient-reported out-of-pocket costs differed by surgical treatment, with BM with and without reconstruction associated with higher patient-reported out-of-pocket costs than BCS (P < .01 and P = .02, respectively). Likewise, compared with BCS, BM with reconstruction (BMR) and BM were associated with higher debt incurred from cancer (27% of patients with BMR reporting more than $5,000 v 12% of patients with BCS, P < .01; and 25% of patients with BM v 12% of patients with BCS, P = .04), higher rates of cancer-induced financial hardship (27% of patients with BMR v 14% of patients with BCS, P < .01; and 33% of patients with BM v 14% of patients with BCS, P < .1), higher rates of altered or reduced employment (70% of patients with BMR v 46% of patients with BCS, P < .01; and 67% of patients with BM v 46% of patients with BCS; P = .02), and higher rates of significant to catastrophic financial burden (23% of patients with BMR v 12% of patients with BCS, P = .02; and 30% of patients with BM v 12% of patients with BCS, P < .01). Median reported out-of-pocket costs were higher after breast reconstruction (P < .01) compared with women who did not undergo reconstruction.

Evaluating the Magnitude of Financial Burden

Overall, 35% of respondents reported experiencing a somewhat (18.5%), significant (13%), or catastrophic (3%) financial burden as a result of their cancer treatment. On unadjusted analysis, BM with or without reconstruction was associated with a greater reported financial burden compared with BCS. On regression analysis, BM with or without reconstruction was associated with higher odds of reporting somewhat to catastrophic burden than BCS (odds ratio [OR], 1.89; 95% CI, 1.07 to 3.33). Covariables associated with decreased odds of somewhat to catastrophic burden included increasing age at diagnosis (OR, 0.95; 95% CI, 0.92 to 0.97), higher household income of $74,001 to $125,000 (v ≤ $74,000; OR, 0.33; 95% CI, 0.19 to 0.56) and more than $125,000 (v ≤ $74,000; OR, 0.13; 95% CI, 0.07 to 0.24), increasing percentage of care covered by insurance (OR, 0.50; 95% CI, 0.40 to 0.63), and increasing years since diagnosis (OR, 0.65; 95% CI, 0.53 to 0.80). On bivariable analysis, nonwhite race, private insurance (compared with Medicare), less education, and receipt of chemotherapy were associated with higher experienced financial burden (all P values < .05).

The overwhelming majority of the sample reported that cancer-related treatment burden did not interfere with their receipt of other medical care. However, one in four of the most economically disadvantaged women reported avoiding doctors and medical testing as a result of their cancer diagnosis, and 29% of these women reported difficulty affording basic necessities. Moreover, depleted savings was the most consistently reported outcome of breast cancer treatment in the $45,000 or more (39%), $45,000 to $74,000 (23%), and $74,000 to $125,000 (18%) income groups.

Significant differences were detected across surgery types in reported difficulty affording daily living (P < .01), depleted savings (P < .005), and paying for cancer care (P < .001; Fig 1). Specifically, participants who underwent BM or BMR reported greater financial hardship than women who underwent BCS, including greater out-of-pocket costs, debt incurred, and altered employment. Table 2 lists the financial experiences by surgical treatment.

Fig 1.

Differences in sacrifice reporting by surgery type. (*) P < .05. (**) P < .005. (***) P < .001. BCS, breast-conserving surgery; BM, bilateral mastectomy; BMR, bilateral mastectomy with reconstruction; UM, unilateral mastectomy; UMR, unilateral mastectomy with reconstruction.

TABLE 2.

Experienced Financial Burden by Breast Cancer Surgery

DISCUSSION

To our knowledge, this study is the first to explore how health care expenditures influenced decisions for breast cancer surgery and how comparably effective surgical treatment options affected financial harm. Even in this relatively advantaged convenience sample, one third considered costs when making decisions for breast cancer surgery. Not surprisingly, costs were more influential in surgical choice as household income fell, and in the lowest income group (≤ $45,000/year), were more important than breast preservation or appearance. Overall, 35% of study participants reported significant to catastrophic financial burden as a result of their breast cancer care, and at higher incomes, depleted savings was a commonly reported finding. Notably, BM was associated with greater out-of-pocket costs, higher debt incurred, increased financial burden, and altered employment compared with other options. Cost transparency remains uncommon as women choose between oncologically equivalent surgical treatments. Thus, patients with breast cancer may unknowingly be guiding therapeutic decisions that increase the risk of financial harm.

Other authors have reported differences in the personal and health system costs related to breast cancer treatment. On review of national insurance claims of Medicare-covered and privately insured patients with breast cancer, Smith et al³⁰ demonstrated that mastectomy plus reconstruction was associated with a two-fold greater risk of complications, and an approximately $2,000 to $20,000 difference in health system costs when compared with breast conservation. In a longitudinal survey of 1,502 women after breast cancer, Jagsi et al³⁷ reported that 25% of participants experienced financial decline after diagnosis, and that younger, lower income, minority women were disproportionately at risk. Cost-related treatment delays, refusal, and nonadherence are more common among black women, who are also at greater risk for financial harm and subsequent mortality.^38,39 Consistent with our findings, more intensive breast cancer treatment has been associated with disrupted employment and greater health care burden.⁴⁰

Consistent with other literature, women in our study reported that the fear of cancer recurrence and recommendations from the medical team were the most influential factors on their decisions for breast cancer surgery.^36,41,42 Despite this, 28% of participants reported cancer treatment costs as at least somewhat important in surgical decisions, yet the majority (78%) reported never having discussed costs with their cancer team. Over the past few decades, the rates of contralateral prophylactic mastectomy (CPM) have tripled despite a lack of evidence supporting additional survival benefit.^15,43,44 Many women report pursuing CPM for symmetry, peace of mind, and the ability to forego continued surveillance.^45,46 Nevertheless, when compared with other options, BM has been associated with a greater risk of complications and delays in adjuvant treatment.^44,47,49 Reflecting this, recent guidelines from the American Society of Breast Surgeons and the Choosing Wisely campaign recommend that routine use of CPM be discouraged in average-risk women.^50,51 Unveiling the greater risk for financial burden associated with BM may further inform surgical treatment decisions, and ultimately, has the potential to affect these national trends.

Our study explored how decisions for breast cancer surgery related to financial costs and burden. Notably, the impact of costs on cancer treatment decisions and downstream financial harm are related but distinct. Our results showed that these variables were associated but remain unique. Among women who said costs were important in surgical decisions, 58% experienced meaningful financial burden compared with 25% of patients who did not rate cost as influential. Similarly, 53% of women who reported considering costs in breast cancer treatment decisions reported greater than slight financial burden compared with 20% of women who reported not considering treatment costs at all.

Our study has several important limitations that deserve to be acknowledged. First, although the AOW has been used extensively as a valuable cohort for many studies of patient-reported outcomes in breast cancer care,^46,52 it consists of a high rate of white women with advanced education, insurance coverage, and higher household incomes. For comparison, in 2018 the median national household income was approximately $61,000 per year, and 15% of the country’s population was covered by Medicare.⁵³ Furthermore, members of both the AOW and Sisters Network include women meaningfully engaged in research and survivorship efforts. Thus, our findings may not be generalizable and may underestimate the risk of treatment-related financial harm when compared with the overall population of women with breast cancer. Second, self-reported costs and burden are estimates and vulnerable to recall bias, and the reliability of these data is unknown. Although recall of exact expenditures may be difficult years from diagnosis, patient reports of debt, borrowing, and indirect costs (ie, travel burden, time off work, childcare expenses, etc) require self-report and have been previously described in other cohorts at similar times from diagnosis.⁵⁴ Compared with other participants, women in our study diagnosed ≥ 10 years from survey completion were more likely to report no financial burden related to their cancer care (51% v 38%); however, the magnitude of incurred debt did not differ over time. Importantly, we were unable to determine what costs were attributed to surgery itself versus other components of treatment. Thus, the financial burden reported by participants likely reflects the entire continuum of breast cancer care, including costly imaging and chemotherapy that may contribute more significantly to overall health care spending. Finally, we were unable to determine how patient-reported financial burden affected cancer outcomes, yet data suggest that financial harm affects adherence and survival.¹¹ Thus, financial insecurity resulting from decisions for breast cancer surgery has the potential to influence receipt of adjuvant treatment.

To date, patient out-of-pocket costs and subsequent risk of financial harm have not been routinely incorporated into shared decisions for breast cancer surgery, a process that has otherwise highly revered patient values. Cost transparency has practical and ethical challenges, and high-quality, patient-centered care, aimed at maximizing survival while minimizing harm, should remain the priority. Preferences for cost transparency likely differ across varied socioeconomic backgrounds and health literacy. Prior research has demonstrated that barriers to cost conversations include physician lack of knowledge, inability to intervene, and inadequate time and resources.²⁴ Successful clinical models have included financial care coordinators to help navigate these challenges. Nonetheless, future research is needed to determine the impact of cost transparency on preference-sensitive cancer treatment decisions. Cost transparency at the time of breast cancer diagnosis may inform decisions for breast cancer surgery and has the potential to reduce financial harm without compromising cancer outcomes.

AUTHOR CONTRIBUTIONS

Conception and design: Rachel A. Greenup, Laura Fish, Lisa Tolnitch, S. Yousuf Zafar, Evan R. Myers, E. Shelley Hwang

Financial support: Terry Hyslop, E. Shelley Hwang

Administrative support: E. Shelley Hwang

Provision of study materials or patients: E. Shelley Hwang

Collection and assembly of data: Rachel A. Greenup, Lisa Tolnitch, Evan R. Myers, E. Shelley Hwang

Data analysis and interpretation: Rachel A. Greenup, Christel Rushing, Laura Fish, Brittany M. Campbell, Lisa Tolnitch, Terry Hyslop, Jeffrey Peppercorn, Stephanie B. Wheeler, S. Yousuf Zafar, Evan R. Myers, E. Shelley Hwang

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Financial Costs and Burden Related to Decisions for Breast Cancer Surgery

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/jop/site/ifc/journal-policies.html.

Rachel A. Greenup

Honoraria: Novartis

Terry Hyslop

Consulting or Advisory Role: AbbVie

Travel, Accommodations, Expenses: AbbVie

Jeffrey Peppercorn

Employment: GlaxoSmithKline (I)

Stock and Other Ownership Interests: GlaxoSmithKline (I)

Research Funding: Pfizer

Stephanie B. Wheeler

Research Funding: Pfizer (Inst)

Travel, Accommodations, Expenses: Pfizer

S. Yousuf Zafar

Employment: Shattuck Laboratories (I)

Stock and Other Ownership Interests: Shattuck Laboratories (I)

Consulting or Advisory Role: Vivor, Family Reach Foundation, AIM Specialty Health, McKesson, RTI Health Solutions, Discern Health, Copernicus WCG

Research Funding: AstraZeneca (Inst)

Evan R. Myers

Consulting or Advisory Role: Merck, Allergan, Bayer, AbbVie, Urocure (I), NovoNordisk

No other potential conflicts of interest were reported.