Sage Journals: Discover world-class research

Abstract

Introduction

Discrimination exacerbates disparities among breast cancer survivors (BCS), yet how different reasons for experiencing perceived discrimination (e.g., race, age) influence health remains understudied. We explored the association between self-reported discrimination, psychosocial health, and quality of life (QOL), identified clusters based on reasons for perceived discrimination, and examined differences in QOL and psychosocial outcomes between these clusters.

Methods

In this cross-sectional study, we examined correlations between reasons for perceived discrimination (Everyday Discrimination Scale; EDS), QOL domains (cognitive, physical, social, emotional, and functional QOL measured with FACT-G), social dysfunction (Social Difficulties Inventory), and a psychological distress composite score (included measures of stress [Perceived Stress Scale], anxiety [PROMIS Anxiety], and depression [PROMIS Depression]), among 174 breast cancer survivors (stage 0-IV; ≥21 years). We used k-modes clustering to identify discrimination groups. Differences in demographics, clinical characteristics, and outcomes across clusters were assessed using Chi-square, analysis of variance, covariance, or non-parametric tests, followed by post hoc analyses.

Results

Overall, experiences of discrimination were associated with poorer QOL and psychosocial health (|0.306|<r<|0.452|, P < 0.001). Six distinct clusters emerged based on reasons for perceived discrimination from the EDS. Compared to Cluster 4 (no discrimination), participants in Cluster 1 (discrimination due to gender, age, and physical characteristics) had lower cognitive and physical QOL (4.3 < mean difference [MD]< 5.0, P < 0.001). Participants in Cluster 3 (discrimination due to physical characteristics) had poorer functional QOL, greater social disfunction, and higher psychological distress composite scores (0.3<MD <9.4, P < 0.001) than Cluster 4. Differences between Clusters 2 (discrimination due to gender) and 5 (discrimination due to gender, race/ethnicity) with all other Clusters were not statistically significant (P > 0.05).

Conclusion

QOL and psychosocial health scores varied between clusters based on reasons for perceived discrimination. Future interventions to improve QOL for breast cancer survivors should consider addressing stigma related to gender, physical appearance, and other forms of discrimination.

Keywords

psycho-oncology discrimination quality of life machine learning social phenotype

Introduction

Breast cancer is the most common cancer diagnosed in women, excluding non-melanoma skin cancers, in the US.¹ In the past decades, breast cancer survival rates have increased as a result of improvements in screening, detection, and treatment.^2,3 Nevertheless, breast cancer survivors often face various difficulties during and after breast cancer treatment, including changes in their psychosocial health that negatively impact their quality of life even decades after treatment completion.^4,5 The influence of cancer and its treatment on quality of life and psychosocial health is not uniform. Each person living with and beyond cancer experiences unique circumstances that shape survivorship quality of life and overall health.^6,7

Discrimination refers to the unequal treatment of individuals based on social and demographic characteristics.⁸ Some examples include age, gender, race, ethnicity, sexual orientation, disability, education, income, among others. Structural discrimination, in particular, describes the systemic and institutionalized policies and practices that disadvantage such groups.⁹ Acts of discrimination may involve verifiable events, such as being denied a job due to one’s identity.¹⁰ In parallel, individuals may hold perceptions of those acts, which may reflect their interpretations, lived experiences, and recognition of unfair treatment based on their identity or membership to marginalized social groups.¹¹ Taken together, discrimination at different levels has been associated with poorer health outcomes and overall lower quality of life.^12-17

Although breast cancer survivors are among the most studied cancer populations, studies analyzing the impact of perceived discrimination on survivorship outcomes are scarce.^18,19 Most research to date focuses on racial and ethnic groups, as well as sexual and gender minority (SGM) populations. Findings from these studies indicate that cancer survivors who perceive frequent exposure to racist events have significant deficits in post-cancer functioning, including clinically meaningful reductions in quality of life.^20,21 Additionally, perceived acts of discrimination experienced by SGM survivors are associated with poorer quality of life and greater levels of anxiety, depression, and psychological distress.^22-25

Existing evidence has shown how perceived discrimination relates to the health and quality of life of people living with and beyond cancer.^20-25 However, it is still unknown if and to what extent less commonly studied reasons for perceived acts of discrimination—such as physical disability, age, body size, or physical appearance—influence health and survivorship outcomes. Examining the factors that survivors attribute as sources for perceived discrimination may provide insight into individual processes that influence how discrimination is appraised and internalized.²⁶ In addition, exploring how less commonly studied reasons for perceived discrimination are linked to health is critical to identifying unique and overlooked psychosocial mechanisms of perceived vulnerability, which may shed light on pathways through which social inequities undermine psychosocial health and quality of life,²⁷ and inform intervention strategies to improve the well-being of women with breast cancer.

As noted, breast cancer survivors experience long-term negative impacts of cancer and its treatment that manifest in diminished psychosocial health and quality of life.^6,7 Examining quality of life and psychosocial dysfunction among breast cancer survivors could help us understand how perceived discrimination compounds survivorship challenges and contributes to inequities. Such findings can inform the development of tailored and equitable interventions that advance person-centered care. Thus, in this study we aimed to (1) assess the association between perceived everyday discrimination with quality of life (cognitive, physical, social, emotional, and functional) and psychosocial health (distress, social dysfunction) among breast cancer survivors, (2) identify clusters based on self-reported reasons for perceived experiences of discrimination (e.g., gender, race, physical characteristics, age, physical disability, socioeconomic status, among others), and (3) examine quality of life and psychosocial differences between perceived discrimination clusters.

Methods

Study Design

This is a cross-sectional secondary analysis of baseline data from two prospective studies assessing cognitive functioning in the context of everyday life among breast cancer survivors living in the US^28,29—one with early-stage (0-III) breast cancer survivors (n = 124) and the other with individuals living with metastatic (advanced-stage) breast cancer (n = 52). Convenience sampling was used to nationally recruit and enroll participants through breast cancer social networks (e.g., the Breast Cancer Resource Center, breastcancer.org, Keep A Breast©, METAvivor©) and the UCLA Clinical and Translational Science Institute between 2022 and 2024.

Eligible participants were at least 21 years of age, within six years of completing chemotherapy (for the early-stage sample), off active treatment (except for the advanced stage sample), physically and cognitively able to use digital technology, competent with smartphone/computer use, and with daily access to a personal smartphone and internet. Exclusion criteria for both samples included pregnancy; prior history of systemic cancer treatment unrelated to breast cancer; major sensory, neurological, or psychiatric conditions that would interfere with study participation; and, for the early-stage sample, evidence of metastases to the brain. Only participants that completed the everyday discrimination scale were included in the analysis.

Data collection was conducted remotely using REDCap for survey administration and BrainCheck for cognitive assessments. The University of Texas at Austin Institutional Review Board oversaw and approved the studies, including consenting procedures (STUDY00002393, initially approved on 03/08/2022).

Instrumentation

Clinical and Demographic Characteristics

Participants self-reported their date of birth, race, ethnicity, gender, marital status, years of education, employment, and income in a demographic survey. To measure the impact of cancer and its treatment on financial distress, we administered items 6, 8, and 10 of the Comprehensive Score for Financial Toxicity (COST; Cronbach’s α = 0.87). The three items were measured on a five-point Likert-type scale and summed to obtain a final score ranging between 0 and 12—lower values indicate lower financial distress. Participants also completed an instrument with cancer health history questions that included date of cancer diagnosis, type of breast cancer, treatments received, and comorbidities (all categorical level data). We calculated age and years since cancer diagnosis.

Everyday Discrimination

Participants completed the Everyday Discrimination Scale (EDS) to assess the frequency of experiences of discrimination in sociocultural settings.^30,31 The EDS has nine items (e.g., How often are you treated with less courtesy than other people? How often people act as if they think you are not smart?) answered on a six-point Likert scale ranging from never to almost everyday, and showed high internal consistency reliability in this sample (Cronbach’s α = 0.89). We used the chronicity approach proposed by Michaels and colleagues³² to obtain a final score, which could range between 0 and 2340. Higher values indicate more frequent experiences of discrimination.

When participants reported experiencing at least one act of discrimination a few times a year or more, they were given the option to specify the perceived reasons (one or more) for discrimination. Response options included: (1) ancestry or national origins, (2) race, (3) religion, (4) skin color, (5) tribe, (6) gender, (7) sexual orientation, (8) age, (9) height, (10) weight, (11) other aspects of the physical appearance, (12) education or income (‘SES’), (13) physical disability, and (14) others. We grouped response options 1-5 (‘race/ethnicity’), 6-7 (‘gender’), and 9-11 (‘physical appearance’) based on similarity and to align with potential broader factors that have been associated with health disparities (e.g., racism and cisheteronormativity).

Quality of Life

To measure quality of life, participants completed the Functional Assessment of Cancer Therapy – General (FACT-G).³³ FACT-G is measured in a five-point Likert scale and has 27 items assessing quality of life in four domains: physical, social and family, emotional, and functional quality of life. Internal consistency was high for all domains (0.78 < Cronbach’s α < 0.84). Subscale scores were obtained by summing individual items.³³ The quality of life subscale of the FACT-Cognitive Function (FACT-Cog) was administered to assess cancer-related cognitive impact on quality of life.³⁴ FACT-Cog quality of life has four items measured in a five-point Likert scale, which are then summed to obtain a total score (greater scores represent higher quality of life). FACT-Cog’s Cronbach’s α for internal consistency reliability was 0.88.

Psychosocial Health

Psychological distress was assessed by measuring perceived anxiety, depressive symptoms, and stress using the Patient-Reported Outcomes Measurement Information System (PROMIS) scales — Emotional Distress – Anxiety Short Form 8a (“PROMIS Anxiety;” Cronbach’s α = 0.93), Emotional Distress – Depression Short Form 8a (“PROMIS Depression;” Cronbach’s α = 0.90) and the Perceived Stress Scale (Cronbach’s α = 0.91).^35,36 Higher scores indicate higher symptom burden on all scales. Scale totals were transformed into z scores, and a composite “distress” score was computed as the mean of the three scores to reduce the number of statistical comparisons and account for the high intercorrelation among the measures (0.67 < r < 0.75, P < 0.001). Social dysfunction in the previous month was assessed using the Social Difficulties Inventory (SDI), which is recommended for use in cancer patients.^37-39 We used the SDI-16 total score in these analyses, that can range from 0-44 (Cronbach’s α = 0.91), with scores >10 indicating social dysfunction.³⁹

Data Analysis

Less than 5% of the data were missing across dependent and independent variables of interest. Visual inspection of missingness patterns via an upset plot suggested that the data were missing completely at random. Thus, case mean substitution was applied to the dependent and independent variables of interest before calculating total scores. For aim 1, we then used descriptive statistics to characterize the EDS scores and reasons for perceived discrimination before running the clustering analysis. The associations between EDS with quality of life and psychosocial health factors were assessed using Pearson or Spearman correlations based on normality assumptions.

To address aim 2, we ran a K-modes algorithm as proposed by Huang⁴⁰ to identify clusters based on the perceived reasons for experiencing discrimination. Hamming distances were used to calculate dissimilarity between categorical points and then cluster modes were updated based on the most frequent categorical values in each cluster.⁴¹ To identify the number of clusters, we examined the elbow graph, gap statistic, and silhouette width.^42,43 To assess cluster stability, we conducted multiple runs (n = 30) with random initializations (e.g., running the k-modes algorithm with different initial centroids) and assessed the Adjusted Rand Index (ARI), silhouette scores, and Jaccard Index between runs.^44,45 To test the generalizability of the model, we also calculated the proportion of consistent assignments and the ARI via bootstrapped resampling (n = 100), which consists of repeatedly sampling the dataset with replacements and running the clustering algorithm on each sample to assess cluster consistency.⁴⁶

Finally, for aim 3, differences in demographic characteristics and clinical factors were assessed between clusters using Chi-square, Fisher, analysis of variance (ANOVA), or Kruskal-Wallis tests as appropriate. Post hoc tests comparisons were examined using standardized residual analysis (significant if > |2.0|),⁴⁷ Bonferroni correction, or Dunn’s test adjusting for multiple comparison. We compared quality of life and psychosocial scores between clusters using analysis of covariance (ANCOVA) or ranked ANCOVA,⁴⁸ controlling for demographic and clinical factors that differed between clusters. Only factors that survived the post hoc comparison were retained as covariates. Post hoc tests using estimated marginal means with Bonferroni correction were conducted to explore significant ANCOVA results. All analyses were conducted in RStudio v.4.3.1. Nominal alpha was set at 0.05, but to control for multiple comparisons, we used the Bonferroni correction (0.05/7 = 0.007). The reporting of this study was conducted in accordance with the STROBE guidelines.⁴⁹

Results

Characterizing the Association Between Perceived Discrimination, Quality of Life, and Psychosocial Health

A total of 174 breast cancer survivors were included in the analysis. Figure 1 displays a bar plot with the frequency of perceived discrimination reasons before clustering. Overall, gender was the most common reason, with 59 participants reporting this. The mean score for EDS was 50.8 (SD = 162.7; Median = 7, Interquartile Range = 16). Correcting for multiple comparisons, greater perceived discrimination scores were significantly associated with lower cognitive (r = −0.403, P < 0.001), physical (r = −0.389, P < 0.001), social (r = −0.306, P < 0.001), emotional (r = −0.359, P < 0.001), and functional (r = −0.404, P < 0.001) quality of life, as well as higher levels of social dysfunction (r = 0.452, P < 0.001) and greater psychological distress composite scores (r = 0.393, P < 0.001).

Figure 1.

Bar Plot Displaying the Frequency of Discrimination Reasons Before Clustering. Note. The Y-Axis Represents the Frequencies and the X-Axis the Discrimination Reasons. The Numbers on Top of Each Bar are the Total of Participants Referring That Factor as One of the Reasons for Perceived Discrimination. Participants Could Select One or More Reasons. SES, Socioeconomic Status

Clustering Reasons for Perceived Discrimination

As depicted in Figure 2, the elbow graph suggested five and seven clusters, while the gap and silhouette graphs indicated six and nine clusters, respectively. Given our small sample size, we retained the six-cluster solution to maintain parsimony.

Figure 2.

Elbow, Gap, and Silhouette Graphs to Determine the Number of Clusters. Note. From Left to Right, Elbow, Gap, and Silhouette Graphs. Substantial Drops in the Elbow Statistic Suggest Significant Reduction in Within-Cluster Sum of Squares. For the Gap and Silhouette Graphs, Higher Points Indicate Better Cluster Separation and Cohesion, Respectively

Participants in cluster 1 were more likely to report gender, physical characteristics, and age as reasons for perceived discrimination. In cluster 2, gender alone emerged as a centroid (mode). In cluster 3, physical characteristics were the predominant reason for perceived discriminatory experiences. Cluster 5 participants referred gender, race/ethnicity, and age as the key factors contributing to their experiences of discrimination. For cluster 6, “other” factors emerged as the most common cause of discrimination. Participants in cluster 4 had lower counts for all reasons and none emerged as a mode, suggesting that they experienced discrimination less often. Table 1 shows for the percentage of perceived reasons for experiencing discrimination and centroids per cluster.

Table 1.

Centroids (Modes) of the Six-Cluster Solution

Cluster	Discrimination scores		Reasons for perceived discrimination
Cluster	Mean, SD	Median, IQR	Gender	Race/Ethnicity	Physical characteristics	Age	Physical disability	Socioeconomic status	Other
1 (n = 18)	145.1 ± 228.5	19.0, 97.1	94.4*	22.2	88.9*	83.3*	16.6	22.2	11.1
2 (n = 26)	48.4.1 ± 115.2	11.0, 13.5	100*	7.6	15.4	11.5	4.0	3.8	3.8
3 (n = 17)	86.0 ± 233.4	12.0, 35.5	29.4	35.3	100*	17.6	17.6	11.8	35.3
4 (n = 79)	23.1 ± 163.1	1.0, 2.7	0.0	7.6	0.0	8.9	1.3	1.3	0.0
5 (n = 13)	47.7 ± 65.1	17.0, 33.5	61.5*	76.9*	23.1	92.3*	0.0	8.3	0.0
6 (n = 21)	50.9 ± 77.8	18.0, 41.5	14.3	9.5	9.5	0.0	14.3	19.0	100*

Note. The table displays the percentage of participants reporting each reason for perceived discrimination per cluster. The star (*) indicates that the trait was one of the cluster modes. The “Gender” category includes gender and sexual orientation as reasons for perceived discrimination. The “Race/Ethnicity” category includes ancestry or national origins, race, religion, skin color, and tribe as reasons for perceived discrimination. The “Physical characteristics” category includes height, weight, and other physical traits as reasons for perceived discrimination. The “Socioeconomic status” category includes income and education. IQR, interquartile range; SD, standard deviation.

The mean ARI across random initializations was 0.527, suggesting moderate agreement between cluster assignments. The mean silhouette among random initializations also suggested moderate separation between clusters, with a value of 0.309. Lastly, the Jaccard Index across runs was 0.223, indicating that clusters tend to change somewhat between random initialized runs. The mean proportion of consistent assignments between bootstrapped resampled runs was 0.237, which is low. However, the mean ARI for the bootstrapped sampling runs was 0.506, indicating moderate agreement. Taken together, these findings suggest moderate stability and generalizability of the clustering solution.

Differences in Quality of Life and Psychosocial Health Scores Between Clusters

Among demographic characteristics, age (F = 15.285, P = 0.009) and financial toxicity (F = 13.175, P = 0.021) were significantly different between clusters (see Table 2). The Dunn’s post hoc tests indicated that participants in cluster 4 were significantly older than participants in cluster 2 (54.3 years vs 44.2 years; z = −3.737, Bonferroni corrected P = 0.001). None of the comparisons for financial toxicity survived the post hoc tests. Likewise, none of the clinical characteristics were significantly different between clusters. Thus, we controlled for age when comparing EDS, quality of life, and psychosocial scores between clusters.

Table 2.

Sociodemographic and Clinical Characteristics per Cluster

	Total sample (n = 174)	Cluster						Statistic	P-value
	Total sample (n = 174)	1 (n = 18)	2 (n = 26)	3 (n = 17)	4 (n = 79)	5 (n = 13)	6 (n = 21)	Statistic	P-value
Sociodemographic characteristics
Age (mean, SD)	50.9 ± 11.5	50.2 ± 9.8	44.2 ± 8.2	47.4 ± 10.6	54.3 ± 12.3	50.8 ± 9.2	50.4 ± 11.4	X² = 15.285	0.009
Education years (mean, SD)	16.8 ± 2.7	15.7 ± 2.0	17.5 ± 2.2	15.9 ± 2.3	17.1 ± 2.6	17.5 ± 2.9	15.6 ± 3.7	X² = 10.633	0.591
Racial minority
Yes	37 (21.3)	4 (22.2)	3 (11.5)	6 (35.3)	16 (20.3)	6 (46.2)	2 (9.5)	Fisher’s	0.097
No	134 (77.0)	14 (77.8)	23 (88.5)	11 (64.7)	61 (77.2)	7 (53.8)	18 (85.7)
Not reported	3 (1.7)	0 (0.0)	0 (0.0)	0 (0.0)	2 (2.5)	0 (0.0)	1 (4.8)
Ethnic minority
Yes	17 (9.8)	2 (11.1)	1 (3.8)	3 (17.6)	10 (12.7)	1 (7.7)	0 (0.0)	Fisher’s	0.351
No	153 (87.9)	16 (88.9)	25 (96.2)	14 (82.3)	66 (83.5)	12 (92.3)	20 (95.2)
Not reported	4 (2.3)	0 (0.0)	0 (0.0)	0 (0.0)	3 (3.8)	0 (0.0)	1 (4.8)
Employment status
Employed	103 (59.2)	10 (55.5)	20 (76.9)	9 (52.9)	45 (57.0)	9 (69.2)	10 (47.6)	X² = 5.627	0.344
Unemployed	71 (40.8)	8 (44.4)	6 (23.1)	8 (47.1)	34 (43.0)	4 (30.8)	11 (52.4)	X² = 5.627	0.344
Marital status
Partnered	121 (69.5)	14 (77.8)	21 (80.8)	10 (58.8)	57 (72.1)	6 (46.2)	13 (61.9)	Fisher’s	0.206
Unpartnered	49 (28.2)	3 (16.7)	5 (19.2)	7 (41.2)	20 (25.3)	6 (46.2)	8 (38.1)
Not reported	4 (2.3)	1 (5.55)	0 (0.0)	0 (0.0)	2 (2.5)	1 (7.7)	0 (0.0)
Income
≤ $50,000	32 (18.4)	2 (11.1)	4 (15.4)	6 (35.3)	13 (16.4)	2 (15.4)	5 (23.8)	Fisher’s	0.541
> $50,000	123 (70.7)	14 (77.8)	22 (84.6)	11 (64.7)	54 (68.3)	10 (76.9)	12 (57.1)
Not reported	19 (10.9)	2 (11.1)	0 (0.0)	0 (0.0)	12 (15.2)	1 (7.7)	4 (19.0)
Financial toxicity (mean, SD)	5.3 ± 3.6	6.1 ± 3.4	6.1 ± 3.4	7.0 ± 4.0	4.3 ± 3.4	5.4 ± 2.5	5.9 ± 4.6	X² = 13.175	0.021
Clinical characteristics
Years since diagnosis (mean, SD)	3.4 ± 2.3	3.1 ± 1.5	3.3 ± 2.7	3.0 ± 2.3	3.6 ± 2.5	2.9 ± 1.4	3.5 ± 2.2	X² = 2.132	0.830
Cancer type
Metastatic	52 (29.9)	5 (27.8)	8 (30.8)	6 (35.3)	23 (29.1)	3 (23.1)	7 (33.3)	Fisher’s	0.982
Non-metastatic	122 (70.1)	13 (72.2)	18 (69.2)	11 (64.7)	56 (70.9)	10 (76.9)	14 (66.7)	Fisher’s	0.982
Cancer stage
0 - I	50 (28.7)	8 (44.4)	7 (26.9)	3 (17.6)	23 (29.1)	5 (38.5)	4 (19.0)	Fisher’s	0.935
II - III	67 (38.5)	5 (27.8)	11 (42.3)	7 (41.2)	31 (39.2)	5 (38.5)	8 (38.1)
IV	52 (29.9)	5 (27.8)	8 (30.8)	6 (35.3)	23 (29.1)	3 (23.1)	7 (33.3)
Unsure or not reported	5 (2.9)	0 (0.0)	0 (0.0)	1 (5.9)	2 (2.5)	0 (0.0)	2 (9.5)
Current postmenopausal status
Yes	121 (69.5)	14 (77.8)	14 (53.8)	12 (70.6)	62 (78.5)	9 (69.2)	10 (47.6)	Fisher’s	0.288
No	35 (20.1)	3 (16.6)	8 (30.8)	3 (17.6)	12 (15.2)	4 (30.8)	5 (23.8)
Other, none, or not reported	18 (10.3)	1 (5.6)	4 (15.4)	2 (11.8)	5 (6.3)	0 (0.0)	6 (28.6)
Surgery
Yes	153 (87.9)	17 (94.4)	21 (80.8)	16 (94.1)	69 (87.3)	11 (84.6)	19 (90.5)	Fisher’s	0.780
No	21 (12.1)	1 (5.5)	5 (19.2)	1 (5.9)	10 (12.7)	2 (15.4)	2 (9.5)	Fisher’s	0.780
Chemotherapy
Yes	114 (65.5)	11 (61.1)	21 (80.8)	13 (76.5)	45 (57.0)	8 (61.5)	16 (76.2)	Fisher’s	0.194
No	60 (34.5)	7 (38.9)	5 (19.2)	4 (23.5)	34 (43.0)	5 (38.5)	5 (23.8)	Fisher’s	0.194
Radiotherapy
Yes	125 (71.8)	11 (61.1)	18 (69.2)	12 (70.6)	57 (72.1)	11 (84.6)	16 (76.2)	Fisher’s	0.809
No	49 (28.2)	7 (38.9)	8 (30.8)	5 (29.4)	22 (27.8)	2 (15.4)	5 (23.8)	Fisher’s	0.809
Hormone therapy
Yes	117 (67.2)	13 (72.2)	18 (69.2)	12 (70.6)	54 (68.3)	9 (69.2)	11 (52.4)	Fisher’s	0.795
No	57 (32.8)	5 (27.8)	8 (30.8)	5 (29.4)	25 (31.6)	4 (30.8)	10 (47.6)	Fisher’s	0.795
Immunotherapy
Yes	85 (48.9)	9 (50.0)	15 (57.7)	11 (64.7)	33 (41.8)	5 (38.5)	12 (57.1)	X² = 5.257	0.385
No	89 (51.1)	9 (50.0)	11 (42.3)	6 (35.3)	46 (58.2)	8 (61.5)	9 (42.9)	X² = 5.257	0.385
Comorbidities
No or one	64 (36.8)	3 (16.7)	12 (46.1)	3 (17.6)	33 (41.8)	4 (30.8)	9 (42.9)	Fisher’s	0.065
Two or more	63 (36.2)	10 (55.5)	7 (26.9)	9 (52.9)	26 (32.9)	6 (46.1)	5 (23.8)	Fisher’s	0.065

Note. All but one participant in cluster 3 identified as cisgender women. Continuous data was tested using Kruskal-Wallis’ rank sum test, and categorical data with Chi-square tests, unless otherwise specified. When comparing postmenopausal status at diagnosis, the “Other, none, or not reported” category was excluded.

The EDS scores were significantly different between clusters (F = 31.660, P < 0.001). Consistent with findings from the cluster analysis, the estimated marginal means post hoc tests indicated that participants in cluster 4 reported less frequent subjective discrimination acts than those in clusters 1 (t = 9.218, P < 0.001), 2 (t = 7.590, P < 0.001), 3 (t = 7.271, P < 0.001), 5 (t = −7.638, P < 0.001), and 6 (t = 7.974, P < 0.001). Differences in EDS scores between other clusters were not significant.

As shown in Table 3, the ANCOVA tests reported cluster differences in cognitive (F = 4.473, P < 0.001), physical (F = 4.050, P < 0.001), social (F = 2.930, P = 0.009), and functional (F = 3.746, P = 0.002) quality of life scores. Differences were also observed for the psychological distress composite (F = 7.291, P < 0.001) and social dysfunction (F = 4.127, P < 0.001). The estimated marginal means post hoc tests indicated that clusters 1 and 3 perceived more cognitive problems and social distress when compared to cluster 4. Physical quality of life scores were also significantly higher for participants in cluster 4 when compared to cluster 1. Last, cluster 3 had worse functional quality of life and greater psychological distress than cluster 4 (See superscripts in Table 3).

Table 3.

Functional Assessment of Cancer Therapy – General (FACT-G) and FACT-Cognitive Function-Quality of Life (FACT-Cog QOL) Scores per Cluster

	Cluster						Statistic	P-value
	1 (n = 18)	2 (n = 26)	3 (n = 17)	4 (n = 79)	5 (n = 13)	6 (n = 21)	Statistic	P-value
Cognitive QOL	7.8 ± 4.9^a	10.6 ± 4.8	7.4 ± 5.5^b	12.1 ± 3.7 ^ab	9.9 ± 4.7	10.3 ± 3.7	4.473	<0.001
Physical QOL	17.6 ± 5.4^a	20.9 ± 4.9	17.5 ± 6.4	22.6 ± 4.4^a	19.5 ± 6.8	21.5 ± 3.3	4.050	<0.001
Social QOL	17.1 ± 6.9	18.7 ± 4.9	17.4 ± 5.5	21.3 ± 5.0	18.5 ± 6.0	19.5 ± 6.2	2.930	0.009
Emotional QOL	16.9 ± 4.0	16.6 ± 3.7	16.4 ± 5.3	18.6 ± 4.1	16.5 ± 4.7	17.6 ± 3.2	3.115	0.006
Functional QOL	15.7 ± 5.7	18.2 ± 5.3	15.1 ± 6.2^a	20.0 ± 5.2^a	16.8 ± 4.7	18.9 ± 4.7	3.746	0.002
Social dysfunction	15.3 ± 10.2^a	12.0 ± 7.2	16.8 ± 9.5^b	7.4 ± 7.1 ^ab	11.6 ± 9.6	10.1 ± 6.1	7.291	<0.001
Psychological distress composite score	0.29 ± 0.9	0.11 ± 0.7	0.59 ± 0.8^a	−0.26 ± 0.8^a	0.28 ± 1.0	0.06 ± 0.7	4.127	<0.001

Note. Means and standard deviations were compared between each pair of clusters using estimated marginal means post hoc test with Bonferroni correction (alpha = 0.007). Significant differences between pairs were indicated using the same letter superscript in a given row. QOL, Quality of Life. The statistic (ranked analysis of covariance) and P-value in the right-far columns refer to the omnibus test comparing quality of life and psychosocial domains among clusters, while controlling for age. The psychological distress composite score includes measures of perceived stress, anxiety, and depression.

Discussion

Our results suggest that more frequent experiences of perceived discrimination are linked to poorer self-reported cognitive, emotional, and functional well-being, as well as lower psychosocial health, which is consistent with prior research on discrimination among people with cancer.^20,21,24,50 Participants were also categorized into six clusters based on reasons for perceived discrimination, and differences in health outcomes emerged between them. Specifically, those in cluster 1 (perceived discrimination due to gender, physical characteristics, and age) had lower cognitive and physical quality of life scores, and greater social dysfunction levels than participants from cluster 4 (few to no acts of perceived discrimination). Similarly, the ratings for cognitive and functional quality of life, and social and psychological distress, were worse for those in cluster 3 (perceived discrimination due to physical characteristics) when compared to cluster 4.

Gender emerged as the most frequent reason for perceived discrimination and was a mode in three out of six clusters. While most cancer research to date has examined gender disparities through the lens of biological sex differences, our study specifically examined how gender, as a perceived reason for discrimination, was associated to the health of cisgender women with breast cancer. The study results are in line with the broader body of literature suggesting that gender-based discrimination is both prevalent and associated with poor physical and mental health.^51-53 The study findings also expand the current literature by suggesting that reasons underlying perceived discrimination may differentially influence health among breast cancer survivors. Future research should continue to examine how diverse forms of discrimination uniquely shape women’s health and well-being.

As previously noted, participants in cluster 3, where physical appearance emerged as the primary reason for perceived discrimination, reported poorer health outcomes compared to those with few or no experiences of discrimination (cluster 4). Body changes associated with cancer and its therapies may potentially explain these differences. Prior research suggests that body changes resulting from cancer and its treatment are associated with lower psychological quality of life scores among women with breast cancer.^54-56 Physical alterations, such as scarring or hair loss, can also contribute to social isolation and withdrawal, as survivors may experience discomfort with their appearance or perceive rejection from others due to visible differences.⁵⁷ Furthermore, 12.1% of our participants reported experiencing weight-based discrimination. Weight gain is a common consequence of menopausal transitions induced by anti-estrogen therapy among breast cancer survivors.⁵⁸ Thus, we hypothesize that participants may have perceived discrimination due to treatment-related weight gain, potentially compounding the psychological burden—both from the perceived discriminatory acts themselves and the distress associated with weight changes.^59,60 Overall, the findings suggest that cancer patients may not only feel discriminated against due to changes in physical appearance, but that such perceptions also negatively influence their quality of life.

In a scoping review of the literature, Haase and colleagues⁶¹ synthesized the evidence examining ageism toward older adults in cancer settings. Although most of the literature consisted of opinion pieces, the review also identified empirical studies reporting mixed findings about the impact of ageism on cancer outcomes. However, most analyses used treatment disparities by age as a proxy for ageism. Few studies examined discrimination acts due to age, and none utilized validated and reliable measures to assess ageism. Our study indicates that 23% of the participants perceived their age as a reason for feeling discriminated against, and that being in cluster 1, where age emerged as a centroid, was associated with poorer cognitive and physical quality of life. Future studies directly measuring and addressing age-based discrimination in oncology research and practice are necessary.

Interestingly, health outcomes for participants in cluster 5—where “gender” and “race/ethnicity” emerged as cluster centroids—did not significantly differ from those of other clusters. This cluster also had a greater proportion of women from racial and ethnic minority groups, which are more likely to experience poorer health.^53,62-64 We attribute the discrepancy between our results and the existing literature to limited statistical power resulting from the small size of cluster 5, which included only 13 participants. The lack of significant differences may reflect Type II errors and limited variability in gender, race, and ethnicity rather than a true absence of disparities. Research studies with larger and more racially and ethnically diverse participants are necessary to confirm our findings.

Participants in Cluster 6 primarily selected “Other” as the reason for perceived discrimination. Although this is an original category in the EDS, we did not probe to clarify what additional factors may have been perceived as causes for discrimination. Emerging literature suggests that disease status can be a basis for discrimination.⁶⁵ Thus, we hypothesize that this “other” category may capture perceived experiences of discrimination related to cancer diagnosis and treatment that are “invisible”. Policies such as the Americans with Disabilities Act (ADA) exist to protect individuals with serious medical conditions from discrimination,⁶⁶ but such protections may be inconsistently implemented or insufficiently enforced in practice, especially when disabilities are not visible. Future research is necessary to examine how objective and subjective illness-related discrimination manifests across sociocultural and cancer care settings. Qualitative approaches or mixed-methods designs may be especially useful in identifying illness-specific discrimination across different phases of cancer care, including diagnosis, treatment, and survivorship.

Implications

Addressing perceived discrimination in cancer survivorship requires multifactorial approaches that target individual, interpersonal, clinical, and societal levels. Cancer care interventions and survivorship programs could explicitly address perceived gender-based discrimination as part of broader strategies to promote the health of women with breast cancer. For example, empowerment-focused interventions that build survivors’ confidence in navigating healthcare systems and making treatment decisions, as well as leadership and advocacy training opportunities for women in survivorship settings, can enhance survivors’ sense of empowerment in contexts where stereotypes may influence care and recovery.^67,68 Community-based and participatory approaches, in which survivors are engaged as partners in designing and leading support initiatives, may further enhance agency and resilience by centering survivors’ voices and lived experiences.⁶⁹

At the individual level, interventions such as cognitive therapy, psychosocial support, mindfulness, and positive psychology have demonstrated effectiveness in reducing appearance-related distress and fostering resiliency, and may be especially valuable for survivors who perceive physical appearance as a source of discrimination.^70-73 At the interpersonal level, interventions that strengthen patient–provider communication, including peer-counselor or decision support programs, have been shown to improve engagement and satisfaction among survivors, particularly those from marginalized groups.⁷⁴ Peer- and community-based support networks have also demonstrated benefits by enhancing coping, reducing loneliness, and fostering mutual psychological well-being.⁷⁵

Clinically, training providers to recognize, prevent, and respond to subtle forms of discrimination can contribute to more inclusive care environments that attend to both visible and less visible dimensions of survivors’ identities.^76,77 At the societal level, initiatives that challenge harmful norms and stereotypes, such as public health campaigns and media portrayals that normalize and affirm post-treatment body changes, may foster an environment that breast cancer survivors perceive as more inclusive.⁷⁸ Finally, future research should explore how tailored interventions across these levels can buffer the impact of objective and subjective discrimination on long-term health-related quality of life.

This study had several notable strengths. Valid and reliable measures were used, we included a multidimensional assessment of domains of quality of life, and controlled for demographic factors as appropriate, strengthening the internal validity of the design. Our sample of breast cancer survivors included those with advanced cancer (all cancer stages) and from a broad age range (22-88), enhancing the external validity of the findings. Additionally, to the best of our knowledge, this is the first study categorizing reasons for perceived experiences of discrimination and their association with quality of life and psychosocial health, which offers a foundation for additional research and for guiding intervention development that promotes the psychosocial health and quality of life of breast cancer survivors. Additionally, while no standardized criteria or formal power analysis guidelines exist for determining sample size in cluster analysis, our sample met commonly suggested thresholds reported in the literature (e.g., minimum of 50 participants⁷⁹ or adherence to the 2^d rule,⁸⁰ where d equals the number clustering variables—seven in our case, for a total of 128 observations required). We also strengthened our analysis by rerunning the clustering algorithm using multiple random initializations and bootstrapped resampling, which suggested that the clustering solution had adequate stability and generalizability.

Limitations

This study was a secondary analysis of cross-sectional data from a parent study on cognition, and although data missingness in our study was low and tests indicated that data were missing completely at random (MCAR), we addressed missingness using case-mean substitution to maximize sample retention. This approach may underestimate variability and attenuate associations compared to more robust methods such as multiple imputation. Thus, our findings should be interpreted with caution regarding the potential impact of the strategy used to handle data missingness.

We examined associations between perceived experiences of discrimination and psychosocial health and quality of life. Accordingly, inferences of causality cannot be made. The temporality of these relationships is also unclear; it is possible that lower quality of life and poorer psychosocial functioning may heighten perceptions of discrimination. In addition, our sample was comprised primarily by non-Hispanic white, cisgender, heterosexual, and highly educated breast cancer survivors who may have access to resources that can protect against discrimination. Although our statistical approach to compare clusters accounts for unequal sample sizes, it does not correct for sample size imbalance.⁸¹ Consequently, the overall small and uneven cluster sizes may have also reduced statistical power to detect significant differences. Experimental and longitudinal studies with larger samples are needed to clarify causal pathways and examine how reasons for perceived discrimination influence psychosocial health and quality of life over time. Future research should also include participants from diverse backgrounds, who might experience discrimination in a fashion that clusters differently.

Although the EDS is a valid and reliable instrument, it measures self-reported experiences of discrimination. In addition, we grouped reasons for perceived discrimination into categories (e.g., gender and sexual orientation; height, weight, and other physical characteristics) due to our sample size. This approach may have masked some of the variability between reasons for perceived discrimination in the clusters and limited our ability to identify unique effects of specific reasons for perceived discrimination.

Last, the EDS may not validly reflect the foci of discrimination experiences of breast cancer survivors. Cancer survivorship is associated with unique contexts of stigma (e.g., being blamed for one’s illness, facing negative assumptions about productivity and employability),^82-84 and these domains may not be adequately represented in a generic measure like the EDS. A cancer survivor–specific measure could allow for assessments that capture the social and medical factors experienced by people with cancer. Qualitative research methods may also offer deeper insights into the context and meaning of reported discrimination, particularly as it manifests within intersectional experiences. Such research could also provide information that could be used to develop a measure of perceived discrimination that is more sensitive to the experiences of cancer survivors.

Conclusion

Our findings highlight the need for addressing the influence of perceived discrimination among women with breast cancer. Future research should focus on developing and testing interventions designed to reduce discrimination within cancer care settings. Multidimensional approaches that simultaneously target the various perceived sources for discrimination may be particularly effective in improving breast cancer survivors’ quality of life. Survivorship care planning should also incorporate tailored psychosocial interventions, such as cognitive-behavioral therapy, mindfulness, and positive psychology, while also fostering empowerment, agency, and resilience through community-based and participatory initiatives. Clinically, training providers to recognize and respond to subtle forms of bias can further promote inclusive care environments. Together, these strategies underscore the need for comprehensive, multilevel interventions to mitigate the influence of perceived discrimination and enhance the long-term quality of life of breast cancer survivors.

Future research should extend these findings by applying qualitative methods to understand the specific types of discrimination experienced by breast cancer survivors, using larger and more diverse samples, and applying longitudinal designs to examine discrimination’s impact over time. Intersectionality-informed frameworks might help disentangle how multiple, co-occurring forms of discrimination affect survivorship outcomes in heterogeneous populations. Additionally, studies should explore the role of illness-related discrimination in cancer care settings, particularly among survivors with invisible treatment sequelae. Understanding these dynamics can inform the development of multilevel interventions and policies aimed at reducing discrimination-related health disparities and promoting equity in cancer survivorship.

Footnotes

Acknowledgements

We are deeply grateful to the breast cancer survivors who contributed their time and experiences to this study. We also recognize and appreciate the crucial support of the Breast Cancer Resource Center, breastcancer.org, Keep A Breast Foundation, Young Survivor Coalition, JoyBoots Survivors, and Breast Cancer Recovery in Action for their assistance with participant recruitment. We thank Heather Becker, PhD, from the University of Texas at Austin School of Nursing, for her valuable feedback during the development of this manuscript.

ORCID iDs

Oscar Y. Franco-Rocha

Ashley M. Henneghan

Ethical Considerations

This study was performed in line with the principles of the Declaration of Helsinki. The University of Texas at Austin Institutional Review Board reviewed and monitored all study related procedures (STUDY00002393).

Consent to Participate

Participants provided informed written consent.

Author Contributions

Conceptualization: OFR, AMH. Data curation: OFR, AMH. Formal analysis: OFR. Methodology: OFR, AMH. Project administration: AMH. Funding acquisition: AMH, RCM, KVD. Supervision: AMH, RCM, KVD. Validation: OFR, EF, MP, JS, SGM, AA, KVD, RCM, AMH. Visualization: OFR. Writing – original draft: OFR, EF, MP. Writing – review and editing: OFR, EF, MP, JS, SGM, AA, KVD, RCM, AMH.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by funding from the National Institutes of Health (R21NR020497; AMH, KVD, RCM) and the American Cancer Society (IRG-21-135-01-IRG). KVD is supported by grants from the NIH: K08CA241337 and R35CA283926. OFR is a MASCC Equity Fellowship.

Declaration of Conflicting Interests

The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: Dr. Smits has received grants from the National Institutes of Health, the Department of Defense, Cancer Prevention and Research Institute of Texas, and the Trauma Research and Combat Casualty Care Collaborative Prevention. He has received personal fees from Big Health, Boston University and Brown University for consulting, and from Elsevier and the American Psychological Association for editorial activities. Dr. Smits also has equity interest in Earkick and has received royalty payments from various publishers. The terms of these arrangements have been reviewed and approved by the University of Texas at Austin in accordance with its conflicts of interest policies. Dr. Moore has equity interest, is a consultant and receives compensation from NeuroUX. The terms of this arrangement have been reviewed and approved by UC San Diego in accordance with its conflict-of-interest policies.

Data Availability Statement

De-identified data will be made available upon reasonable requests to the corresponding author and with data use agreements in place*.

Appendix

References

American Cancer Society . Cancer Facts and Figures 2025. Atlanta, GA: American Cancer Society; 2025.

Cronin

Scott

Firth

, et al. Annual report to the nation on the status of cancer, part 1: national cancer statistics. Cancer. 2022;128:4251-4284. doi:10.1002/cncr.34479

Schilsky

Nass

Le Beau

Benz

EJJ

. Progress in cancer research, prevention, and care. N Engl J Med. 2020;383:897-900. doi:10.1056/NEJMp2007839

Rahool

Haider

Shahid

, et al. Medical and psychosocial challenges associated with breast cancer survivorship. Cureus. 2021;13:e13211. doi:10.7759/cureus.13211

King

Stafford

Butow

, et al. Psychosocial experiences of breast cancer survivors: a meta-review. J Cancer Surviv. 2024;18:84-123. doi:10.1007/s11764-023-01336-x

Marzorati

Voskanyan

Sala

, et al. Psychosocial factors associated with quality of life in cancer patients undergoing treatment: an umbrella review. Health Qual Life Outcome. 2025;23:31. doi:10.1186/s12955-025-02357-z

Zhu

Lian

Arndt

Thong

MSY

. Combined healthy lifestyle factors and psychosocial outcomes among cancer survivors: a systematic review and meta-analysis. J Cancer Surviv. 2024. doi:10.1007/s11764-024-01705-0

Pager

Shepherd

. The sociology of discrimination: racial discrimination in employment, housing, credit, and consumer markets. Annu Rev Sociol. 2008;34:181-209. doi:10.1146/annurev.soc.33.040406.131740

Yearby

. The social determinants of health, health disparities, and health justice. J Law Med Ethics. 2022;50:641-649. doi:10.1017/jme.2023.3

10.

Bertrand

Mullainathan

. Are emily and greg more employable than lakisha and jamal? A field experiment on labor market discrimination. Am Econ Rev. 2004;94:991-1013. doi:10.1257/0002828042002561

11.

Straiton

Aambø

Johansen

. Perceived discrimination, health and mental health among immigrants in Norway: the role of moderating factors. BMC Public Health. 2019;19:325. doi:10.1186/s12889-019-6649-9

12.

Pascoe

Smart Richman

. Perceived discrimination and health: a meta-analytic review. Psychol Bull. 2009;135:531-554. doi:10.1037/a0016059

13.

Zhou

Nutakor

Larnyo

Addai-Dansoh

Cui

Dey

NEY

. Perceived discrimination as a mediator between cultural identity and mental health symptoms among racial/ethnic minority adults in the United States: insights from the health information national trends survey 6. Front Psychol. 2024;15:1233919. doi:10.3389/fpsyg.2024.1233919

14.

Yolci

Schenk

Sonntag

P-T

Peppler

Schouler-Ocak

Schneider

. Observed and personally experienced discrimination: findings of a cross-sectional survey of physicians and nursing staff. Hum Resour Health. 2022;20:83. doi:10.1186/s12960-022-00779-0

15.

Clark

Cranston

Polin

, et al. Structural interventions that affect racial inequities and their impact on population health outcomes: a systematic review. BMC Public Health. 2022;22:2162. doi:10.1186/s12889-022-14603-w

16.

Liu

DeCuir

Kia

Peterson

Miller

Issaka

. Tools to measure the impact of structural racism and discrimination on gastrointestinal and hepatology disease outcomes: a scoping review. Clin Gastroenterol Hepatol. 2023;21:2759-2788.e6. doi:10.1016/j.cgh.2022.12.002

17.

Eldridge

Berrigan

. Structural racism and triple-negative breast cancer among black and white women in the United States. Health Equity. 2022;6:116-123. doi:10.1089/heq.2021.0041

18.

Tucker-Seeley

Abu-Khalaf

Bona

, et al. Social determinants of health and cancer care: an ASCO policy statement. JCO Oncol Pract. 2024;20:621-630. doi:10.1200/OP.23.00810

19.

Yedjou

Sims

Miele

, et al. Health and racial disparity in breast cancer. In: Ahmad

, ed. Breast Cancer Metastasis and Drug Resistance: Challenges and Progress. Cham, Switzerland: Springer International Publishing; 2019:31-49.

20.

Mandelblatt

Ruterbusch

Thompson

, et al. Association between major discrimination and deficit accumulation in African American cancer survivors: the Detroit research on cancer survivors study. Cancer. 2023;129:1557-1568. doi:10.1002/cncr.34673

21.

Shariff-Marco

Sangaramoorthy

Ellis

, et al. Impact of racial/ethnic discrimination on quality of life among breast cancer survivors. Am J Epidemiol. 2023;192:367-376. doi:10.1093/aje/kwac208

22.

Boehmer

Clark

Winter

Berklein

Ozonoff

. Sexual minority-specific experiences of colorectal cancer survivors. Health Psychol. 2022;41:884-892. doi:10.1037/hea0001229

23.

Ross

Rosser

BRS

Polter

, et al. Discrimination of sexual and gender minority patients in prostate cancer treatment: results from the restore-1 study. Stigma Health. 2023;8:85-92. doi:10.1037/sah0000356

24.

Boehmer

Ozonoff

Winter

, et al.

Anxiety and depression in colorectal cancer survivors: are there differences by sexual orientation?

Psychooncology. 2022;31:521-531. doi:10.1002/pon.5837

25.

Ussher

Allison

Perz

Power

Out with Cancer Study Team . LGBTQI cancer patients’ quality of life and distress: a comparison by gender, sexuality, age, cancer type and geographical remoteness. Front Oncol. 2022;12:873642. doi:10.3389/fonc.2022.873642

26.

Krieger

. Embodying inequality: a review of concepts, measures, and methods for studying health consequences of discrimination. Int J Health Serv. 1999;29:295-352. doi:10.2190/M11W-VWXE-KQM9-G97Q

27.

Arizpe

Ochoa-Dominguez

Navarro

, et al. Racial/ethnic disparities: discrimination’s impact on health-related quality of life-an all of us cancer survivors’ cross-sectional study. J Racial Ethn Health Dispar. 2025;12:1765-1776. doi:10.1007/s40615-024-02006-z

28.

Henneghan

Paolillo

Van Dyk

, et al. Feasibility and psychometric quality of smartphone administered cognitive ecological momentary assessments in women with metastatic breast cancer. Digit Health. 2025;11:20552076241310470. doi:10.1177/20552076241310474

29.

Henneghan

Paolillo

Van Dyk

, et al. Feasibility, reliability and validity of smartphone administered cognitive ecological momentary assessments in breast cancer survivors. Front Digit Heal. 2025;7:1543846. doi:10.3389/fdgth.2025.1543846

30.

Balsam

Molina

Blayney

Dillworth

Zimmerman

Kaysen

. Racial/ethnic differences in identity and mental health outcomes among young sexual minority women. Cult Divers Ethnic Minor Psychol. 2015;21:380-390. doi:10.1037/a0038680

31.

Krieger

Smith

Naishadham

Hartman

Barbeau

. Experiences of discrimination: validity and reliability of a self-report measure for population health research on racism and health. Soc Sci Med. 2005;61:1576-1596. doi:10.1016/j.socscimed.2005.03.006

32.

Michaels

Thomas

Reeves

, et al. Coding the everyday discrimination scale: implications for exposure assessment and associations with hypertension and depression among a cross section of mid-life African American women. J Epidemiol Community Health. 2019;73:577-584. doi:10.1136/jech-2018-211230

33.

Cella

Tulsky

Gray

, et al. The functional assessment of cancer therapy scale: development and validation of the general measure. J Clin Oncol. 1993;11:570-579. doi:10.1200/JCO.1993.11.3.570

34.

Wagner

Sweet

Butt

, et al. Measuring patient self-reported cognitive function: development of the functional assessment of cancer therapy-cognitive function instrument. J Support Oncol. 2009;7:W32-W39.

35.

Cohen

Kamarck

Mermelstein

. A global measure of perceived stress. J Health Soc Behav. 1983;24:385-396.

36.

Patient-Reported Outcomes Measurement Information System . Obtain & administer measures. HealthMeasures. https://www.healthmeasures.net/explore-measurement-systems/promis/obtain-administer-measures. Accessed March 12, 2025.

37.

Catt

Starkings

Shilling

Fallowfield

. Patient-reported outcome measures of the impact of cancer on patients’ everyday lives: a systematic review. J Cancer Surviv. 2017;11:211-232. doi:10.1007/s11764-016-0580-1

38.

Wright

Kiely

Johnston

Smith

Cull

Selby

. Development and evaluation of an instrument to assess social difficulties in routine oncology practice. Qual Life Res. 2005;14:373-386. doi:10.1007/s11136-004-5332-4

39.

Wright

Smith

Keding

Velikova

. The social difficulties inventory (SDI): development of subscales and scoring guidance for staff. Psychooncology. 2011;20:36-43. doi:10.1002/pon.1705

40.

Huang

. Extensions to the k-means algorithm for clustering large data sets with categorical values. Data Min Knowl Discov. 1998;2:283-304. doi:10.1023/A:1009769707641

41.

Huang

. A fast clustering algorithm to cluster very large categorical data sets in data mining. In: Proceedings of the SIGMOD Workshop on Research Issues on Data Mining and Knowledge Discovery. Vancouver, BC: University of British Columbia, Department of Computer Science; 1997:1-8.

42.

Arbelaitz

Gurrutxaga

Muguerza

Pérez

Perona

. An extensive comparative study of cluster validity indices. Pattern Recogn. 2013;46:243-256. doi:10.1016/j.patcog.2012.07.021

43.

Tibshirani

Walther

Hastie

. Estimating the number of clusters in a data set via the gap statistic. J R Stat Soc Ser B Stat Methodol. 2002;63:411-423. doi:10.1111/1467-9868.00293

44.

Kuncheva

Vetrov

. Evaluation of stability of k-means cluster ensembles with respect to random initialization. IEEE Trans Pattern Anal Mach Intell. 2006;28:1798-1808. doi:10.1109/TPAMI.2006.226

45.

Yang

. Temporal data clustering. In: Yang

, ed. Temporal Data Mining via Unsupervised Ensemble Learning. Cambridge, MA: Elsevier; 2017:19-34.

46.

Jain

Moreau

. Bootstrap technique in cluster analysis. Pattern Recogn. 1987;20:547-568. doi:10.1016/0031-3203(87)90081-1

47.

Sharpe

. Your chi-square test is statistically significant: now what? Pract assessment. Res Evaluaion. 2015;20:1-10. doi:10.7275/tbfa-x148

48.

Olejnik

Algina

. Parametric ANCOVA and the rank transform ANCOVA when the data are conditionally non-normal and heteroscedastic. J Educ Stat. 1984;9:129-149. doi:10.3102/10769986009002129

49.

von

Altman

Egger

, et al. Strengthening the reporting of observational studies in epidemiology (STROBE) statement: guidelines for reporting observational studies. BMJ. 2007;335:806-808. doi:10.1136/bmj.39335.541782.AD

50.

Franco-Rocha

Wheldon

Phillips

, et al. A cross-sectional analysis of psychosocial mediators between discrimination and health-related quality of life among sexual and gender minority cancer survivors. Psychooncology. 2025;34:e70154. doi:10.1002/pon.70154

51.

Vigod

Rochon

. The impact of gender discrimination on a woman’s mental health. eClinicalMedicine. 2020;20:100311. doi:10.1016/j.eclinm.2020.100311

52.

Homan

. Health consequences of structural sexism: conceptual foundations, empirical evidence and priorities for future research. Soc Sci Med. 2024;351:116379. doi:10.1016/j.socscimed.2023.116379

53.

Emmer

Dorn

Mata

. The immediate effect of discrimination on mental health: a meta-analytic review of the causal evidence. Psychol Bull. 2024;150:215-252. doi:10.1037/bul0000419

54.

Amini-Tehrani

Zamanian

Daryaafzoon

, et al. Body image, internalized stigma and enacted stigma predict psychological distress in women with breast cancer: a serial mediation model. J Adv Nurs. 2021;77:3412-3423. doi:10.1111/jan.14881

55.

Zhang

Pennell

Bernardo

, et al. Body image, physical activity and psychological health in older female cancer survivors. J Geriatr Oncol. 2021;12:1059-1067. doi:10.1016/j.jgo.2021.04.007

56.

Begovic-Juhant

Chmielewski

Iwuagwu

Chapman

. Impact of body image on depression and quality of life among women with breast cancer. J Psychosoc Oncol. 2012;30:446-460. doi:10.1080/07347332.2012.684856

57.

Rodrigues

ECG

Neris

Nascimento

de Oliveira-Cardoso

ÉA

Dos Santos

. Body image experience of women with breast cancer: a meta-synthesis. Scand J Caring Sci. 2023;37:20-36. doi:10.1111/scs.13102

58.

Buch-Larsen

Lund-Jacobsen

Andersson

Schwarz

. Weight change in post-menopausal women with breast cancer during chemotherapy-perspectives on nutrition, activity and bone metabolism: an interim analysis of a 5-year prospective cohort. Nutrients. 2021;13:2902. doi:10.3390/nu13082902

59.

Pila

Sabiston

Castonguay

Arbour-Nicitopoulos

Taylor

. Mental health consequences of weight cycling in the first-year post-treatment for breast cancer. Psychol Health. 2018;33:995-1013. doi:10.1080/08870446.2018.1453510

60.

Soares

Correia Silva

. Breast cancer: a review on quality of life, body image and environmental sustainability. World J Cancer Oncol Res. 2023;22:133-144. doi:10.31586/wjcor.2023.826

61.

Haase

Sattar

Pilleron

, et al. A scoping review of ageism towards older adults in cancer care. J Geriatr Oncol. 2023;14:101385. doi:10.1016/j.jgo.2022.09.014

62.

Zavala

Bracci

Carethers

, et al. Cancer health disparities in racial/ethnic minorities in the United States. Br J Cancer. 2021;124:315-332. doi:10.1038/s41416-020-01038-6

63.

Ginsburg

Vanderpuye

Beddoe

, et al. Women, power, and cancer: a lancet commission. Lancet. 2023;402:2113-2166. doi:10.1016/S0140-6736(23)01701-4

64.

Bertozzi

Londero

Diaz Nanez

, et al. Breast cancer care for the aging population: a focus on age-related disparities in breast cancer treatment. BMC Cancer. 2025;25:492. doi:10.1186/s12885-025-13893-8

65.

Caston

Waters

Williams

, et al. Patient-reported discrimination among limited-resourced cancer survivors: a brief report. J Psychosoc Oncol. 2023;41:630-641. doi:10.1080/07347332.2022.2154186

66.

US Equal Employment Opportunity Commission . Cancer in the workplace and the ADA. US equal employment opportunity commission. https://www.eeoc.gov/laws/guidance/cancer-workplace-and-ada. Accessed April 14, 2025.

67.

Ruiz-Romeo

Ciria-Suarez

Medina

, et al. Empowerment among breast cancer survivors using an online peer support community. Support Care Cancer. 2024;33:56. doi:10.1007/s00520-024-09119-5

68.

Avery

Thomas

Howell

Wilner

C-JD

. Empowering cancer survivors in managing their own health: a paradoxical dynamic process of taking and letting go of control. Qual Health Res. 2023;33:412-425. doi:10.1177/10497323231158629

69.

McNeill

IHC

Cho

Escoto

Harris

. Community outreach and engagement strategies to address breast cancer disparities. Curr Breast Cancer Rep. 2020;12:209-215. doi:10.1007/s12609-020-00374-z

70.

Sebri

Durosini

Triberti

Pravettoni

. The efficacy of psychological intervention on body image in breast cancer patients and survivors: a systematic-review and meta-analysis. Front Psychol. 2021;12:611954. doi:10.3389/fpsyg.2021.611954

71.

Huang

, et al. Effect of ehealth interventions on body image of patients with cancer: systematic review. J Med Internet Res. 2025;27:e55564. doi:10.2196/55564

72.

Brown-Iannuzzi

Adair

Payne

Richman

Fredrickson

. Discrimination hurts, but mindfulness may help: trait mindfulness moderates the relationship between perceived discrimination and depressive symptoms. Pers Indiv Differ. 2014;56:201-205. doi:10.1016/j.paid.2013.09.015

73.

Villalobos

Cabrera

. A constructive critique: advancing positive psychology with equity and justice. Int J Appl Posit Psychol. 2025;10:17. doi:10.1007/s41042-024-00209-7

74.

Sheppard

Wallington

Willey

, et al. A peer-led decision support intervention improves decision outcomes in black women with breast cancer. J Cancer Educ. 2013;28:262-269. doi:10.1007/s13187-013-0459-z

75.

Hemming

Duijts

SFA

Zomerdijk

, et al.

A systematic review of peer support interventions to improve psychosocial functioning among cancer survivors: can findings be translated to survivors with a rare cancer living rurally?

Orphanet J Rare Dis. 2024;19:473. doi:10.1186/s13023-024-03477-3

76.

Campbell

. Healthcare inequities and healthcare providers: we are part of the problem. Int J Equity Health. 2025;24:97. doi:10.1186/s12939-025-02464-9

77.

Nyblade

Stockton

Giger

, et al. Stigma in health facilities: why it matters and how we can change it. BMC Med. 2019;17:25. doi:10.1186/s12916-019-1256-2

78.

Aristokleous

Karakatsanis

Masannat

Kastora

. The role of social media in breast cancer care and survivorship: a narrative review. Breast Care. 2023;18:193-199. doi:10.1159/000531136

79.

Henry

Dymnicki

Mohatt

Allen

Kelly

. Clustering methods with qualitative data: a mixed-methods approach for prevention research with small samples. Prev Sci. 2015;16:1007-1016. doi:10.1007/s11121-015-0561-z

80.

Formann

. Die latent-class-analyse: Einführung in Theorie Und Anwendung [Latent Class Analysis: Introduction to Theory and Application]. Weinheim, Germany: Beltz; 1984.

81.

Conover

Iman

. Analysis of covariance using the rank transformation. Biometrics. 1982;38:715-724.

82.

Knapp

Marziliano

Moyer

. Identity threat and stigma in cancer patients. Health Psychol Open. 2014;1:2055102914552281. doi:10.1177/2055102914552281

83.

Rai

Syurina

Peters

RMH

Putri

Zweekhorst

MBM

. Non-communicable diseases-related stigma: a mixed-methods systematic review. Int J Environ Res Publ Health. 2020;17:6657. doi:10.3390/ijerph17186657

84.

Stergiou-Kita

Pritlove

Kirsh

. The “big C”-stigma, cancer, and workplace discrimination. J Cancer Surviv. 2016;10:1035-1050. doi:10.1007/s11764-016-0547-2

Reasons for Perceived Everyday Discrimination,Quality of Life,and Psychosocial Health of Breast Cancer Survivors: A Cross-Sectional Cluster Analysis

Abstract

Introduction

Methods

Results

Conclusion

Keywords

Introduction

Methods

Study Design

Instrumentation

Clinical and Demographic Characteristics

Everyday Discrimination

Quality of Life

Psychosocial Health

Data Analysis

Results

Characterizing the Association Between Perceived Discrimination, Quality of Life, and Psychosocial Health

Clustering Reasons for Perceived Discrimination

Differences in Quality of Life and Psychosocial Health Scores Between Clusters

Discussion

Implications

Limitations

Conclusion

Footnotes

Acknowledgements

ORCID iDs

Ethical Considerations

Consent to Participate

Author Contributions

Funding

Declaration of Conflicting Interests

Data Availability Statement

Appendix

References