Introduction

In the United States, more than 150,000 men and women are currently living with metastatic breast cancer (mBC), a number that is expected to increase by nearly 100,000 within the next 10 years [1]. While individual prognosis varies, treatment advances have resulted in more patients with mBC living longer. However, improvements in survival mean that patients often manage ongoing treatment for extended durations, which can impact physical, psychosocial, and emotional health [2].

To better understand the experience of living with mBC, the Dr. Susan Love Foundation for Breast Cancer Research initiated the Metastatic Breast Cancer Collateral Damage Project questionnaire, where “collateral damage” was defined as the “psychological, social, vocational, financial” impact of an mBC diagnosis [3, 4]. Findings demonstrated that patients younger than 50 years at diagnosis, those who experienced financial challenges, and those with children younger than 18 years of age reported higher levels of collateral damage [4]. Further, collateral damage was independently associated with physical and psychological morbidity [4]. The results of this initial study provided key insights regarding the prevalence of and risk factors for collateral damage among adults with mBC. However, greater than 90% of participants in that study were non-Hispanic White (NHW) and 75% had a college degree, limiting generalizability of the study’s findings to individuals with mBC from other racial, ethnic, and socio-economic backgrounds. To address this limitation and identify gaps in care and future opportunities for intervention, we sought to gain an in-depth understanding of collateral damage in mBC patients treated at four New York City hospitals, focusing on Black and Hispanic/Latinx adults and individuals from lower socio-economic backgrounds.

Methods

Study population, recruitment, and enrollment

Adults age ≥18 years with a confirmed diagnosis of mBC, either de novo stage IV breast cancer or a metastatic recurrence of early-stage breast cancer, were eligible to be contacted to participate in a one-time survey. At three New York Presbyterian (NYP) affiliated sites—Weill Cornell Medicine (WCM)-Manhattan, NYP-Brooklyn Methodist Hospital, and NYP-Queens—potentially eligible patients were identified by research staff through the screening of clinic lists. Following confirmation by a treating provider, patients were contacted either by phone, email, or approached in the clinic. Purposive sampling was conducted to meet a pre-planned accrual target of at least 50% Black or Hispanic/Latinx participants as well as to identify patients with Medicaid insurance to include representation of socio-economically disadvantaged individuals.

At a fourth NYP site, Columbia University Irving Medical Center (CUIMC), potentially eligible patients were identified through the screening of clinic lists by research staff and approached in the clinic where they were provided with information about the study, including a flyer. Patients were instructed to contact WCM research staff directly by email or phone if they were interested in participating; interested patients were then enrolled in the study at WCM. Patients who completed the survey received a $20 gift card.

Patients completing the survey online via REDCap provided electronic informed consent prior to being emailed a link to the survey. Patients completing a paper version of the survey provided written informed consent. The survey was available in English and Spanish. In addition to the recruitment that took place at WCM and CUIMC, information about the study, including the study flyer, was shared with patient advocate partners for dissemination. The WCM (IRB# 21-12024213) and the CUIMC (IRB# AAAU1910) Institutional Review Boards reviewed and approved this study.

Survey measures

Demographics and social determinants of health

Patients were asked to self-report their current age, age when they were first diagnosed with mBC, sex, race, ethnicity, and fluency in English, whether they spoke a language other than English at home, whether they were United States born, number of years living in the United States, highest attained level of education, marital status, health insurance status and type, and occupational status. Additional questions assessed financial comfort (How would you describe your household’s financial situation right now?) [5, 6]; housing situation in the last 6 months, frequency (over the past 30 days) of worry about food running out, whether a lack of reliable transportation kept them from medical appointments, meetings, work or from getting to things needed for daily living, and whether there was a time that when they needed to see a doctor, but could not because of cost, in the past 12 months [7].

Survey of Health, Impact, Needs, and Experiences (SHINE)

The 36-item SHINE questionnaire was developed with the input of mBC patient advocates as part of the Metastatic Breast Cancer Collateral Damage Project [4]. The SHINE includes nine independent subscales measuring the following domains: (1) employment concerns, (2) financial concerns, (3) insurance problems, (4) mortality/uncertainty, (5) activity disruption, (6) interpersonal concerns, (7) social isolation/withdrawal, (8) self-concept disruption, and (9) benefit finding. For items encompassing the first five domains, participants were asked to rate on a 5-point scale (“not at all” to “very much” or “I have not experienced this”) how distressed or bothered they were as a result of their mBC. For the remaining domains, participants were asked to rate their level of agreement on a 5-point scale (strongly disagree to strongly agree) for each item describing their mBC experience. Scores for the employment concerns, financial concerns, insurance problems, mortality/uncertainty, and activity disruption subscales range from 0 to 4 and for the interpersonal concerns, social isolation/withdrawal, self-concept disruption, and benefit-finding subscales, range from 1 to 5. Subscale scores are calculated based on the mean of the four items included in the subscale, with higher scores indicative of higher endorsement of the experience measured in that subscale.

Patient-Centered Communication in Cancer Care (PCC-Ca)-6

The PCC-Ca-6 is a validated 6-item scale to capture patient-reported experiences around communication with their healthcare providers and decision making [8]. Each item corresponds to a single PCC domain: (1) exchanging information, (2) making decisions; (3) fostering healing relationships; (4) enabling patient self-management; (5) managing uncertainty; and (6) responding to emotions. Each item includes a 1–5 response scale, scores range from 1 to 5, with higher scores indicating better communication and an overall score calculated based on the mean of the six items.

Patient-reported Outcomes Measurement Information System (PROMIS)-29 Profile (PROMIS-29)

The 29-item PROMIS profile includes the 4-item short form versions for the following seven domains: (1) anxiety, (2) depression, (3) fatigue, (4) pain interference, (5) physical function, (6) sleep disturbance, and (7) social roles (each item with five response options) plus a single item assessing pain intensity (scale of 0 [no pain] to 10 [worst imaginable pain]) [9]. Raw PROMIS scores were converted to a T-score in which 50 is the mean and 10 is the standard deviation of the reference population.

Statistical analysis

Descriptive statistics were used to characterize the study sample using N (%) for categorical variables, and median (interquartile range [IQR]) for continuous variables. Fisher’s exact test and the Kruskal–Wallis rank-sum test were used to examine the association between socio-economic variables/survey measures and race/ethnicity. All p values were two sided with statistical significance evaluated at the 0.05 alpha level. All analyses were performed in R Version 4.2.1 (R Foundation for Statistical Computing, Vienna, Austria).

Results

Study population characteristics

Of the 232 mBC patients contacted and invited to participate in the study, 87 participants completed a survey between May 2022 and May 2023 (response rate: 38%). Study population characteristics are summarized in Table 1. Of the 87 study participants, 14% self-identified as Hispanic/Latina or Spanish origin, 28% as non-Hispanic Black (NHB), 7% as Asian American/Pacific Islanders (AAPI), and 41% as NHW; for 10%, race was categorized as other, more than one race, do not know, or was unknown (N = 2) Most participants (94%) reported speaking English “well” or “very well”; 35% reported speaking a language other than English at home; and 41% were born outside of the United States. Regarding education, 41% did not have a college degree. At the time of the survey, 32% of participants were employed, 19% unemployed, 27% were retired, and 13%, disabled. Thirty-eight percent were insured by Medicare or Medicaid. Among participants where age information was available, median age at initial diagnosis (N = 73) was 50 (IQR: 38, 60) years, and median age at the time of survey completion (N = 72) was 59 (IQR: 49, 68) years.

Table 1 Study population characteristics
Full size table

Social determinants of health

Prevalence of food and housing insecurity, transportation challenges, and financial well-being among study participants are shown in Table 2. When describing their household’s current financial situation, overall, 36% indicated that they had enough money for special things after paying the bills, 30% indicated they had enough money to pay their bills, but little spare money for extras, 13% reported they had money to pay bills but only because they have cut back on things, and 20% reporting difficulty paying bills no matter what. There was a statistically significant difference in proportion of race and ethnicity groups reporting difficulty paying bills (p < 0.001). NHB participants more frequently indicated having difficulty paying bills (52%) than Hispanic (25%), NHW (3%), and AAPI participants (17%). While 100% of Hispanic, NHW, and AAPI participants reported having stable housing, 29% of NHB participants were worried about losing housing in the future (p = 0.002). Regarding food insecurity, 42% of Hispanic and 46% of NHB participants (vs. 8%, NHW and 0%, AAPI, p = 0.005) reported that over the past 30 days, they worried sometimes or often that food would run out before having money to buy more. Regarding transportation concerns, 18% of Hispanic and 17% of NHB adults indicated lack of reliable transportation in the last year that kept them from medical appointments, meeting, work, or from getting to things needed for daily living (vs. 0%, in NHW and AAPI, p = 0.033).

Table 2 Social determinants of health by race and ethnicity
Full size table

Collateral damage

Median overall SHINE scores were high for the financial concerns (2.00, IQR: 0.56, 3.38) and mortality/uncertainty subscales (2.75, IQR: 1.06, 3.31), indicating high levels of distress corresponding to these two domains (Supplemental Table 1). For the interpersonal, social isolation/withdrawal, self-concept disruption, and benefit-finding subscales (Supplemental Table 1), overall median scores ranged from 2.00 (IQR: 1.50, 2.94) for the social isolation/withdrawal subscale, indicating that participants experienced low levels of isolation and loneliness to 4.00 (IQR: 3.56, 4.50) for benefit finding, indicating that participants experienced appreciation, hope, gratitude, and a sense of closeness with friends. Across all nine SHINE subscales, differences by race and ethnicity were only statistically significant (p = 0.014) for the financial concerns subscale, with median scores lowest for NHW participants (0.75, IQR: 0.38, 2.08) and highest for NHB (2.63, IQR: 1.75, 3.81) and AAPI (3.63, IQR: 2.00, 3.94) participants. Figure 1a–e depict the proportion of participants who indicated they were “at least somewhat” distressed or bothered for individual SHINE items composing the employment concerns, financial concerns, insurance problems, mortality/uncertainty, and activity disruption subscales. Figure 2a–d show the proportion indicating agreement with each SHINE item composing interpersonal, social isolation/withdrawal, self-concept disruption, and benefit finding subscales.

Fig. 1
figure 1figure 1figure 1

a–e Proportion of participants who were at least somewhat distressed or bothered - SHINE employment concerns, financial concerns, insurance problems, mortality/uncertainty, and activity disruption subscales

Full size image
Fig. 2
figure 2figure 2

a–d Proportion of participants in agreement - SHINE interpersonal, social isolation/withdrawal, self-concept disruption, and benefit finding subscales

Full size image

Physical and psychosocial well-being

Median PROMIS-29 T-scores are summarized in Table 3. Overall median physical (43, IQR: 37, 49) function and social roles and activities (46, IQR: 44, 57) scores were less than one standard deviation (10) below the population norm score of 50, indicating poorer health relative to the general population. Overall, median anxiety (60, IQR: 51, 64), depression (53, IQR: 41, 59), fatigue (55, IQR: 49, 59), sleep disturbance (54, IQR: 52, 56), and pain interference (56, IQR: 42, 61) ranged from less than one to one standard deviation above the population norm score of 50, also indicating poorer health relative to the general population. Median sleep disturbance (p = 0.029) and pain interference scores (p = 0.009) differed by race and ethnicity: median sleep disturbance scores were highest in Hispanic (57, IQR: 55, 60) participants relative to NHW (53, IQR: 52, 55), NHB (55, IQR: 53, 56), and AAPI (51, IQR: 49, 54) participants while median pain interference scores were highest in Hispanic (60, IQR: 52, 61), AAPI (60, IQR: 57, 61), NHB (57, IQR: 53, 64) relative to NHW (50, IQR: 42, 56) participants. Other PROMIS-29 sub-domains were not significantly different by race and ethnicity.

Table 3 PROMIS-29 by race and ethnicity
Full size table

Communication

The median overall PCC score was 4.0 (IQR: 3.5, 4.5), indicating that participants were generally satisfied with the quality of communication that they have had with their healthcare providers. In addition, there were no statistically significant differences in median scores or individual question responses by racial or ethnic group (data not shown).

Discussion

In a diverse population of adults living with mBC in New York City, NHB and Hispanic/Latina patients reported higher levels of financial concern and were more likely to report food and transportation insecurity than NHW patients. While overall physical and mental well-being were modestly poorer among our study participants compared to healthy United States population norms, differences by race and ethnicity were only observed for sleep disturbance and pain interference domains. Participants also generally rated their experiences around communication and decision making highly across all racial and ethnic groups. With more than 40% of participants identifying as a member of a historically under-represented racial or ethnic minority group, our study provides novel insights into physical, financial, and psychosocial sequelae experienced after an mBC diagnosis.

Despite underrepresentation of certain racial and ethnic groups in many studies, disparities in financial and employment outcomes have previously been documented among mBC patients of color. In a study of over 1000 mBC patients, including 9% who identified as NHB, 7% as Hispanic, and 10%, AAPI, 54% of patients reported ceasing their cancer treatment due to cost [10]. This was most common among NHB (87%), AAPI (94%), and American Indian/Alaska Native (99%) patients compared to NHW patients (39%). Nearly two thirds of patients also reported that they stopped working, with this most frequent among Hispanic (84%), AAPI (78%), and American Indian/Alaska Native patients (98%) [10]. Compared to NHW patients, Hispanic patients were more likely to report borrowing money from friends or family and not paying other medical expenses while Black mBC patients were more likely to have taken unpaid leave from a job [10]. A separate analysis inclusive of these patients found that several domains of financial insecurity were more common among those who were uninsured vs. insured [11]. These included interacting with a collections agency (92% among uninsured vs. 30% among insured), inability to meet monthly expenses (48% among uninsured vs. 30% among insured), needing to file for bankruptcy (50% among uninsured vs. 41% in insured), and not having sufficient savings/assets to meet the costs of treatment (53% among uninsured vs. 41% among insured) [11]. In contrast, several domains salient to financial distress were more common among those who were insured, including higher than expected out-of-pocket costs (60% among insured vs. 25% among uninsured), worry about future financial challenges (77% among insured vs. 47% among uninsured), and worry about the financial stress impact on family (54% among insured vs. 25% among uninsured) [11]. Differences by race and ethnicity were observed as well: although Black and Hispanic/Latina patients were less likely to report being contacted by collections compared to White women, they were more likely to have filed for bankruptcy [11]. Further, while Black women were less likely to report being worried about future financial problems, Hispanic women were more likely to report being financially stressed and being worried about financial stress on their family, compared to White women [11].

Unmet social needs, including food, housing, and transportation insecurity, have previously been documented to negatively impact cancer care. In a study inclusive of over 1000 Black and Latino patients (with early stage or metastatic cancer) treated in New York City-based clinics, nearly 70% reported unmet transportation needs and 42% of Latino and 53% of Black participants reporting a food-related unmet need [12]. While unmet housing needs were overall less common (reported by 14% and 17% of Latinos and Black patients, respectively), Black patients who indicated an unmet housing need were three times more likely to miss an appointment [12]. In a New Mexico-based study of adults with early-stage breast, colorectal, and prostate cancer, 26% were food insecure both prior to and remained so following their diagnosis, while 10% of participants became (newly) food insecure post-diagnosis, with Hispanic patients more than twice as likely to be food insecure compared to non-Hispanic patients [13]. Further, patients who were food insecure were more likely skip, postpone, or alter their cancer care due to cost [13]. Advanced cancer patients with unmet social needs may be particularly vulnerable to having their care impacted, given the frequency of appointments, costs of ongoing treatment, and the extended period of time that they must manage symptoms and side effects. Not unexpectedly, quality of life may also be impacted: in a cohort study of Black cancer survivors, participants who indicated they were food, housing, or transportation insecure had significantly lower FACT-G scores compared to those who were not [14].

There is recognition that multi-level strategies at the policy, systems, institutional, practice, community, and patient levels are vital to addressing unmet social needs among both patients with early stage as well as advanced cancer [15,16,17]. Examples of approaches that have been tested or implemented at cancer centers include working with ride-sharing companies and providing transportation vouchers to assist with transportation challenges, cancer-center-based food pantries, and patient navigation/care coordination programs [15, 16, 18,19,20]. However scaling, disseminating, and sustaining effective interventions across the cancer care trajectory, while a critical need, is also a challenge given both structural and individual level barriers and complexities of care.

Our study includes some notable limitations. Study participants were recruited from four sites at two urban, academic cancer centers, though the catchment areas of these cancer centers encompass racially, ethnically, and socio-economically diverse neighborhoods as demonstrated from our population. The experience of patients treated in community practices as well as rural settings may differ substantially and would not have been captured in our study. We also acknowledge that we were likely underpowered to observe potential differences across multiple racial/ethnic groups, particularly given the small sample size of certain subgroups (e.g., only six AAPI participants). In addition, while we captured many demographic characteristics, we did not collect information about whether participants had young or school age children living at home, though we acknowledge caregiving responsibilities likely contribute to the psychosocial burden experienced by younger patients. Nonetheless, our findings highlight notable disparities in social and financial concerns by race, ethnicity, and economic status among adults living with mBC. Routine identification of individuals with unmet needs through screening followed by efficient follow-up and delivery of resources to address unmet needs is increasingly recognized as a priority, with the goal of mitigating psychosocial and financial sequelae among those most vulnerable to experiencing these challenges following a mBC diagnosis.