Differences in Coping with Breast Cancer Between Lesbian and Heterosexual Women: A Life Course Perspective

Abstract

Background: We tested a theoretical framework to explain differences in coping responses to breast cancer between lesbian and heterosexual women.

Materials and Methods: Breast cancer survivors were recruited through cancer registries and community-based sampling. Cross-sectional telephone surveys were completed among self-identified lesbian (n = 330) and heterosexual (n = 595) women who were diagnosed with breast cancer. Five subscales from the Mini-Mental Adjustment to Cancer (Mini-MAC) Scale were used to measure coping with breast cancer among women post-treatment. Mediation analysis was used to examine the explanatory power of life course factors (e.g., parenting and education) in explicating the association between sexual identity and coping responses.

Results: Lesbian women had lower mean scores on the anxious preoccupation and cognitive avoidance subscales (p < 0.05). These differences were moderated by age at diagnosis, with differences in anxious preoccupation and cognitive avoidance greater among women diagnosed with breast cancer before 45 years of age. Having children mediated the association between lesbian identity and anxious preoccupation, but only among women diagnosed at younger ages. College education mediated the association between lesbian identity and cognitive avoidance among women diagnosed at older ages.

Conclusions: Despite previous evidence of suboptimal cancer care and gaps in supportive services, lesbian women with breast cancer demonstrate adaptive coping. This study calls for an increased focus on life course factors, both in the empirical and theoretical literature, which may partially explain some of this resiliency. Identifying mechanisms that lead to active coping can inform supportive care for both lesbian and heterosexual women.

Introduction

There are over 3.1 million breast cancer survivors in the United States. Based on current population data, we can estimate that at least 6.8% (or 210,800) of these survivors identify as lesbian/gay or bisexual, which we define as sexual minorities.¹ Very little systematic research has been published on the experiences and psychosocial outcomes of sexual minority cancer survivors.² Existing evidence suggests that sexual minority cancer survivors are less satisfied with the cancer care they received compared to their heterosexual counterparts.³ There are also gaps in supportive services tailored for sexual minority cancer survivors and evidence of discriminatory attitudes and practices directed to this stigmatized population in healthcare environments.²

Paradoxically, research focusing on quality of life outcomes among breast cancer survivors post-treatment find few differences between sexual minority and heterosexual women.⁴ Current theoretical frameworks, like the minority stress model, suggest that sexual minority cancer survivors would suffer additional burdens because of their stigmatized identity. This theoretical prediction belies current evidence—at least regarding breast cancer survivors. Arena et al.⁵ hypothesized that sexual minorities may develop mastery of cognitive processes to deal with distress since the development of a sexual minority identity is typified by stress, emotional turmoil, and risk of social rejection.^6,7 Indirect evidence for this hypothesis is found in studies of coping among cancer survivors. Although limited in scope, these studies reported that lesbian-identified women engaged in more active and adaptive coping strategies, including less cognitive avoidance and less fatalism than did heterosexual women.^5,8 These findings were partially replicated in another study, in which sexual minority women reported less cognitive avoidance, less anxious preoccupation, and less helplessness–hopelessness cognitions than heterosexual women.⁴

Overall these studies suggest that sexual minority women—by using more active and less anxiety focused coping strategies—exhibit adaptive coping in the face of a distressing health threat. However, not all the evidence supports this claim. Lesbian-identified women scored lower on fighting spirit—an active coping strategy associated with better mental adjustment to cancer—compared to their heterosexual counterparts.⁸ Differences in coping outcomes may partially result from differences in age among the study samples.^4,5,8 Existing evidence suggests that women diagnosed with breast cancer at younger ages exhibit increased levels of distress.^9,10 Age at diagnosis may serve as a proxy for life course milestones relevant to long-term psychosocial outcomes for cancer survivors. We propose that life course factors drive the observed differences in coping between sexual minority, particularly lesbian-identified women, and heterosexual women.

A life course approach focuses on the timing and importance of events (e.g., having children) and social roles (e.g., parent) in human development; thus, our conceptual framework is informed by Role Theory and Social Stress Theory.^11–13 Role theory postulates that differences in competing expectations surrounding roles (e.g., parental role vs. sick role) can be the source of significant stress.¹¹ Role conflict is evident in research on parenting among cancer patients and survivors, in which the demands of the illness not only add to the normal stressors of parenting but also create additional distress when the cancer patient or survivor is unable to fulfil previous parental expectations.^14–16 These additional stressors may result in differential coping strategies. Sexual minorities have historically been excluded from social institutions that establish norms relating to marriage, parenting, and family. Previous research suggests that there is much more variability in family structures and social support networks among sexual minority compared to heterosexual women.^8,17,18 This current body of research lacks a cohesive theoretical framework that seeks to systematically identify explanatory mechanisms, such as life course factors, in coping responses based on sexual identity. Previous research indicates that lesbian women are less likely to have children¹⁹; thus, differences in coping between lesbian and heterosexual women may be explained by differences in parental roles.

A related, but distinct proposition is that differences in socioeconomic status (SES) may account for differences in coping between lesbian and heterosexual women. Social stress theory postulates that individuals with lower SES lack external resources necessary to manage stress associated with adverse life events.¹² In response to stressful life events like cancer, individuals with lower SES may rely on cognitive and behavioral avoidance strategies compared to individuals with higher SES.²⁰ For example, higher education was negatively associated with cognitive avoidant coping among cancer patients.²⁰ Previous research found that lesbian-identified women are more likely to graduate from college than heterosexual women.^21–23 Thus, systematic differences in educational attainment may explain differences in coping between lesbian and heterosexual women.

Based on these theoretical approaches, we propose that observed differences in coping with breast cancer—in which lesbian-identified women respond with more active adaptive coping strategies—result, in part, from differences in two life course factors: parenthood and higher education. In this study, we tested the following hypotheses: the associations between sexual identity and coping will be mediated by (H1) parental responsibility (i.e., having underage children in the household will result in more anxious and avoidant coping strategies) and (H2) higher education attainment (i.e., college education will result in less anxious and more active coping strategies). In addition, we propose that the explanatory power of these life course factors will be greater among women impacted by cancer at younger ages. Thus, we hypothesize (H3) that life course factors (i.e., parental responsibility and college education) will exhibit a greater impact on the coping responses among women diagnosed at younger ages.

Materials and Methods

We utilized a merged dataset from two parent breast cancer studies to increase the size and diversity of the sample, thereby allowing us to conduct detailed analyses by age and sexual identity. Both parent studies recruited sexual minority and heterosexual breast cancer survivors and collected survey data through telephone. Study staff made up to 10 attempts to contact survivors for initial screening and then administered the survey to eligible survivors. One parent study utilized the Massachusetts Cancer Registry, identifying women with primary diagnoses of DCIS (ductal carcinoma in situ) or I–III breast cancer between the years 2000 and 2005. Keeping the inclusion criteria identical, the registry data were further supplemented using community-based methods (e.g., targeted community outreach and respondent referrals), which provided an overall sample of 438 breast cancer survivors in this parent study.^4,24,25

The second parent study relied exclusively on targeted recruitment from the community through the Dr. Susan Love Army of Women (AOW), an organization whose mission is to incre ase recruitment for breast cancer studies (www.armyofwomen.org). Using the AOW's e-list, the second parent study recruited sexual minority women with breast cancer of any stage and subsequently heterosexual women who resembled the sexual minority women with respect to age and breast cancer history. This second parent study was conducted from October 2011 to June 2012 and yielded a sample of 535 breast cancer survivors. No women participated in both studies and by combining the two parent studies, the sample for this study increased to 330 self-identified lesbian, 48 bisexual, and 595 heterosexual women with a breast cancer diagnosis. Bisexual-identified women were removed for this analysis due to the small subsample size, leaving a final analytic sample of 925.

The institutional review boards of Boston University and the Massachusetts Department of Public Health approved all aspects of this study.

Measures

Coping with breast cancer

The abbreviated version of the Mental Adjustment to Cancer scale (i.e., the Mini-Mental Adjustment to Cancer [Mini-MAC]) was used to measure the following five dimensions of coping: (1) Helplessness–Hopelessness; (2) Anxious Preoccupation; (3) Cognitive Avoidance; (4) Fighting Spirit; and (5) Fatalism.²⁶ The helplessness–hopelessness (e.g., “I feel like giving up”), anxious preoccupation (e.g., “I worry about the cancer returning or getting worse”), and cognitive avoidance (e.g., “I deliberately push all thoughts of cancer out of my mind”) subscales are thought to represent maladaptive coping, as they are positively correlated with increased anxiety and poorer mental adjustment.⁴ The fatalism (e.g., “I've had a good life what's left is a bonus”) and fighting spirit (e.g., “I am determined to beat this disease”) subscales measure adaptive coping cognitions related to a positive reflection on one's life and a sense of optimism.²⁷ Reliability was previously demonstrated as acceptable, with subscale alpha coefficients ranging from 0.62 (Fatalism) to 0.87 (Helplessness–Hopelessness).²⁶ The fighting spirit and helplessness–hopelessness subscales were negatively correlated and combined into single fighting spirit subscale by taking the differences between the two scores (i.e., Fighting Spirit–Helplessness/Hopelessness) as previously suggested in the literature.²⁸

Mediators

Two life course factors were considered potential mediators. These included the following: having a college education (vs. no college education) and having children younger than 18 years of age living in the household.

Moderator

Age at breast cancer diagnosis was dichotomized as first diagnosis at 44 years of age or younger, or at 45 years of age or older. This allowed us to examine the effect of breast cancer timing on the associations between life course factors and coping. Consistent with the literature, 45 was used as the cut point to differentiate breast cancer diagnoses of older women (defined as 45 and older) from younger women with “early-onset breast cancer,” who are likely premenopausal, within child bearing years, and for whom role conflicts related to parenting and education are most likely.²⁹

Sociodemographics and clinical variables

Sociodemographic characteristics analyzed included age, race, employment status, poverty status, and health insurance. Women were also asked to self-report cancer disease stage at the time of interview and type of treatments received, including surgery, chemotherapy, radiation, and/or hormone therapy (e.g., tamoxifen or aromatase inhibitor therapy). Only women who had concluded active treatment were recruited for the parent study; however, a small percentage of women in the second study were undergoing active invasive treatment (e.g., radiation or chemotherapy at the time of survey interview). Thus, active treatment was examined as a control variable. The number of comorbidities was assessed using a previously developed instrument.³⁰

Statistical analysis

Differences in sociodemographic and clinical characteristics by sexual identity (i.e., lesbian vs. heterosexual) were assessed using chi-square test of independence or analysis of variance, where appropriate. Bivariate differences in outcomes between lesbian and heterosexual women were examined for the total sample, as well as stratified by age at breast cancer diagnosis. We used the inferential approach outlined by MacKinnon et al. to estimate the indirect effects of sexual identity on coping mediated by the life course variables.^31,32 This approach does not assume normality of data and employs a bootstrapping method to estimate standard errors and the statistical significance of the assocations.³¹ The coping outcomes (e.g., anxious preoccupation) were simultaneously regressed on sexual identity (main predictor) and the life course variables (mediators). In the same model, the association of sexual identity (main predictor) and the life course variables (mediators) was estimated. This was done separately for each outcome and stratified by age at cancer diagnosis (<45 years old vs. ≥45 years old). Probit regression coefficients were used to estimate the associations between sexual identity and the binary mediators (e.g., college educated). Ordinary least-squares regression coefficients were used to estimate associations between a predictor and the continuous coping outcomes (e.g., anxious preoccupation). Two-sided p-values ≤0.05 were considered statistically significant. p-Values <0.10 were considered significant when identifying control variables for multivariate analysis.

Results

Clinical and sociodemographic characteristics are reported in Table 1. A larger proportion of lesbian (31.9%) compared to heterosexual women (21.6%) were diagnosed with breast cancer before 45 years of age (p < 0.01). There were some differences by stage of cancer, with a larger proportion of lesbian women (23.0% vs. 17.1%) diagnosed with noninvasive breast cancer (p = 0.08). In addition, lesbian women were less likely to receive radiation therapy (60.9% vs. 69.9%; p < 0.01) than heterosexual women, and were more likely to have mastectomy without reconstructive surgery (33.6% vs. 17.4%; p < 0.01). There were no differences in the percentage of lesbian and heterosexual women who were undergoing active cancer treatment.

Table 1.

Clinical and Demographic Characteristics of the Sample

	Total (N = 925), n (%)	Lesbian (n = 330), n (%)	Heterosexual (n = 595), n (%)	p
Clinical variables
Age at diagnosis				<0.01
<45 Years old	232 (25.2)	104 (31.9)	128 (21.6)
≥45 Years old	688 (74.8)	466 (78.5)	222 (68.1)
Years since diagnosis, mean (SD)	5.1 (4.5)	5.3 (3.9)	4.9 (4.8)	0.31
Disease stage				0.08
DCIS	178 (19.2)	76 (23.0)	102 (17.1)
Stage 1	323 (34.9)	107 (32.4)	216 (36.3)
Stage 2	270 (29.2)	100 (30.3)	170 (28.6)
Stage 3	69 (7.5)	24 (7.3)	45 (7.6)
Stage 4	62 (6.7)	14 (4.2)	48 (8.1)
Unknown	23 (2.5)	9 (2.7)	14 (2.4)
Surgery received				<0.01
Lumpectomy	453 (49.0)	151 (45.8)	302 (50.8)
Mastectomy only	214 (23.1)	111 (33.6)	103 (17.3)
Mastectomy without reconstruction	239 (25.8)	62 (18.8)	177 (29.8)
Neither lumpectomy nor mastectomy	19 (2.1)	6 (1.8)	13 (2.2)
Adjuvant therapies
Radiation therapy	617 (66.7)	201 (60.9)	416 (69.9)	<0.01
Chemotherapy	491 (53.1)	181 (54.9)	310 (52.1)	0.42
Hormone therapy	629 (68.1)	217 (65.8)	412 (69.4)	0.26
Receiving active treatment
Current radiation therapy	23 (2.4)	6 (1.8)	15 (2.5)	0.49
Current chemotherapy	56 (5.9)	21 (6.4)	33 (5.6)	0.61
No. of comorbidities				0.72
0	97 (10.6)	31 (9.5)	66 (11.2)
1	161 (17.6)	62 (19.1)	99 (16.8)
2	213 (23.3)	73 (22.5)	140 23.8)
3 or more	443 (48.5)	159 (48.9)	284 (48.2)
Sociodemographics
Age, mean years (SD)	56.8	54.5 (8.3)	58.1 (10.4)	<0.01
White race	824 (89.1)	292 (88.5)	532 (89.4)	0.67
Employed	581 (62.8)	231 (70.0)	350 (58.8)	<0.01
Below poverty level	331 (35.8)	95 (28.8)	236 (39.7)	<0.01
Had health insurance	916 (99.0)	326 (98.8)	590 (99.2)	0.73
Partnered	731 (79.1)	262 (79.6)	469 (78.8)	0.77
Family/life coursea
Children under 18 in household	188 (20.3)	49 (14.9)	139 (23.4)	<0.01
College degree	668 (72.5)	259 (78.5)	409 (69.1)	<0.01

^aLife course factors were assessed at the time of interview.

DCIS, ductal carcinoma in situ; SD, standard deviation.

In terms of life course factors, a smaller proportion of lesbian (14.9%) compared to heterosexual (23.5%) women had children younger than 18 years of age living in the household (p < 0.01). Conversely, a larger proportion of lesbian (78.5%) compared to heterosexual (69.1%) women completed a college degree (p < 0.01). Among the overall sample, a larger proportion of women diagnosed younger than 45 years of age (44.8%) compared to women diagnosed at 45 years of age or older (12.2%) had children living in the household at the time of the survey (p < 0.01). In addition, a larger proportion of women diagnosed in the younger age group had a college degree (77.6%) compared to women diagnosed at 45 years of age or older (70.7%; p < 0.05).

There were bivariate differences between lesbian and heterosexual women on two of the four coping variables (Table 2). Lesbian women had lower mean scores on the anxious preoccupation and cognitive avoidance subscales (p < 0.05). These differences were moderated by age at diagnosis, with differences in anxious preoccupation and cognitive avoidance greater among women diagnosed with breast cancer before 45 years of age.

Table 2.

Mean (Standard Deviation) Scores on the Mini-Mental Adjustment to Cancer Scale by Sexual Orientation and Stratified by Age at Breast Cancer Diagnosis

	Full sample			<45 Years old at diagnosis			≥45 Years old at diagnosis
	Lesbian	Heterosexual	p	Lesbian	Heterosexual	p	Lesbian	Heterosexual	p
Anxious preoccupation	17.5 (4.5)	18.2 (4.6)	0.03	17.7 (4.7)	20.0 (4.6)	<0.01	17.4 (4.4)	17.8 (4.4)	0.30
Cognitive avoidance	9.1 (2.5)	9.7 (2.4)	<0.01	8.8 (2.5)	10.0 (2.1)	<0.01	9.3 (2.5)	9.7 (2.5)	0.05
Fatalism	15.3 (2.5)	15.5 (2.6)	0.12	15.1 (2.2)	15.1 (2.5)	0.98	15.3 (2.6)	15.7 (2.6)	0.11
Fighting spirit	29.6 (4.7)	29.6 (4.6)	0.97	29.7 (4.2)	29.9 (4.0)	0.79	29.6 (4.8)	29.5 (4.8)	0.82

Control variables

As shown in Table 1, several demographic (employment status and income) and clinical factors (stage of cancer and types of treatments received) differed between lesbian and heterosexual women. Of these factors, stage and treatment types were also significantly correlated with anxious preoccupation; however, stage and treatment were also highly correlated with each other. To avoid issues with multicollinearity, only stage of cancer was included as a control variable. Similarly, income and education were significantly associated with cognitive avoidance and sexual identity; however, income and college education (one of the mediators) were strongly correlated; so only education was included in multivariate analyses.

Anxious preoccupation: <45 years of age at diagnosis

Results from the multivariate analyses are reported in Figure 1. Lesbian identity was negatively associated with anxious preoccupation among women diagnosed before 45 years of age after adjusting for children, college education, and controlling for stage at diagnosis (Fig. 1A). Women with minor children and who were diagnosed before 45 years of age had higher scores on anxious preoccupation after adjusting for all other variables. There was a significant indirect effect of lesbian identity on anxious preoccupation partially mediated by having underage children.

FIG. 1.

Path models stratified by age at breast cancer diagnosis. (A, B) Anxious preoccupation regressed on children, college, and sexual identity controlling for stage at diagnosis. (C, D) Cognitive avoidance regressed on children, college, and sexual identity. No control variables were added for (C) or (D). Results are regression coefficients with standard errors in parentheses. Solid lines indicate direct effects and dotted lines indicate indirect effects. *p ≤ 0.05.

Anxious preoccupation: ≥45 years of age at diagnosis

Lesbian identity was not associated with anxious preoccupation among women diagnosed at or after 45 years of age and no indirect effects were identified (Fig. 1B). College education did not mediate the association between lesbian identity and anxious preoccupation in either age group.

Cognitive avoidance: <45 years of age at diagnosis

Lesbian identity was negatively associated with cognitive avoidance among women diagnosed before 45 years of age after adjusting for children and college education (Fig. 1C). This association was not mediated by children or college as indicated by lack of significant indirect effects. Women with minor children who were diagnosed before 45 years of age had higher scores on cognitive avoidance after adjusting for all other variables.

Cognitive avoidance: ≥45 years of age at diagnosis

Lesbian identity was not associated with cognitive avoidance among women diagnosed at or after 45 years of age after adjusting for children and college (Fig. 1D); however, there was a significant indirect effect of lesbian identity on cognitive avoidance fully mediated by college education. Unlike women diagnosed before 45 years of age, women with college education who were diagnosed at or after 45 years of age had lower scores on cognitive avoidance after adjusting for all other variables.

Discussion

In this study, we identified differences in coping with breast cancer between lesbian and heterosexual women. Some of our findings replicated results from previous studies. For example, as was previously found, lesbian women in this study also reported less avoidant coping strategies than heterosexual women.^4,5,8 However, in contrast to previous reports that found no differences in anxiety-focused coping between lesbian and heterosexual women, we found that lesbian women reported lower levels of anxious preoccupation than their heterosexual counterparts.^4,8 No other differences in coping (e.g., Fatalism and Fighting Spirit) were identified.

Our findings add important nuances to this growing body of research. The differences in coping that we identified were largely driven by the variation among women diagnosed at younger versus older ages (H3). Thus, the differences in coping between lesbian and heterosexual women were more pronounced among women diagnosed with breast cancer at younger ages. The inclusion of life course factors helped to contextualize some of these differences.

Our hypotheses that parental responsibility would explain the differences in coping found between lesbian and heterosexual women were partially supported. Having children mediated the association between lesbian identity and anxious preoccupation (H1), but only among women diagnosed at younger ages (H3). This suggests that differences in anxiety-focused coping between lesbian and heterosexual women are, at least in part, driven by conflicts in parental roles. Heterosexual women diagnosed with breast cancer at a young age are more likely to have parental responsibilities, which explains their higher level of anxiety-focused coping, whereas lesbian women in this situation are less likely to have children and thus have lower anxiety-focused coping. While our mediated models provide some empirical support for these claims, a significant inverse association still existed between lesbian identity and anxious preoccupation among young women with breast cancer after adjusting for the effect of children. Thus, differences in parental responsibility did not account for all the variation in this finding. There are likely other explanatory factors, like the conditioned resiliency hypothesis,⁵ which account for the additional variation.

The patterns of associations were different when examining differences in cognitive avoidance. In the multivariate models, differences in cognitive avoidance between lesbian and heterosexual women were only found in women diagnosed with breast cancer at younger ages, and these differences reflect less avoidance among lesbian women. However, these differences were not mediated by life course factors in this younger age group. Conversely, college education mediated the association between lesbian identity and cognitive avoidance among women diagnosed at older ages (H2). The statistically significant bivariate association between lesbian identity and lower levels of cognitive avoidance was completely attenuated in the multivariable model with the addition of educational attainment. Older lesbian women, that is those diagnosed with breast cancer at 45 years of age or later, were more likely to have completed college and in turn had lower levels of avoidant coping. In contrast, lesbian and heterosexual women diagnosed with breast cancer at younger ages had similar educational levels. This finding may reflect the impact of cancer on educational attainment, where early-onset breast cancer prevents or delays college. An alternative explanation is that there may be a cohort difference in college completion between older and younger lesbian women. Differences in the findings from previous studies on older and younger cohorts of sexual minority women, in which sexual minority identity was positively associated with higher educational achievement in the older cohort,³³ but negatively associated with educational achievement in the younger cohort,³⁴ provide preliminary support for this latter explanation.

These results have several implications for future theoretical and empirical work that can further our understating of how and why sexual orientation matters in cancer survivorship research and practice. Others hypothesized that sexual minorities develop mastery of cognitive processes to deal with distress and are therefore better able to cope with a life-threatening illness like breast cancer.⁵ An alternative hypothesis, which is partially supported by our findings, is that sexual minorities systematically differ from heterosexuals in life course factors that are proximately related to distress and subsequent coping responses. Therefore, it is not minority stress related to sexual orientation that results in differences in coping, but the indirect effects of sexual orientation on life course trajectories. Further research is needed to identify the reasons why younger lesbian women, when faced with a serious health threat, engage in more active coping then their heterosexual counterparts.

The findings of this study are limited by several methodological constraints. Most importantly, the study design was cross-sectional and thus we are unable to make causal claims. Our samples were not randomly selected and the probability of inclusion is unknown. Thus, the external validity of these findings is unknown. Overall, this sample of women was characterized by adaptive coping strategies; thus, limited variability in the coping outcomes may explain some of the null findings. In addition, the demographic characteristics of the study sample are not representative of the U.S. population. For example, the majority of participants were white, thus findings from this study might not explain coping responses to breast cancer among ethnoracial minority women.

While limited by these methodological constraints, this study adds to the growing literature in survivorship research focused on sexual minority populations. In this study, we replicated some of the findings from previous research that reflects the resiliency of sexual minority women. Despite evidence of suboptimal cancer care and gaps in supportive services, lesbian women with breast cancer demonstrate adaptive coping.^2,3 This study calls for an increased focus on life course factors, both in the empirical and theoretical literature, which may partially explain some of this resiliency. Identifying mechanisms that lead to active coping can inform supportive care for both lesbian and heterosexual women.

Author Disclosure Statement

No competing financial interests exist.