|Home | About | Journals | Submit | Contact Us | Français|
We investigated the association between psychological distress and indices of social integration and communal coping among sisters from hereditary breast and ovarian cancer (HBOC) families.
Sixty-five sisters from 31 HBOC families completed the Brief Symptom Inventory-18 and the Colored Eco-Genetic Relationship Map, which identified members of participants’ social support networks. Hierarchical linear models were used for all analyses to account for the clustering of sisters within families.
Intra-family correlation coefficients suggested that sisters shared perceptions of breast cancer risk and worry, but not ovarian cancer risk and worry. Further, sisters demonstrated shared levels of anxiety and somatization, but not depressive symptoms. Communal coping indices quantifying shared support resources were negatively related to anxiety and somatization. The number of persons with whom cancer risk information was shared exhibited a positive trend with somatization. Social integration, as measured by the size of participants’ emotional support network, was negatively associated with anxiety. Lower depression scores were observed among participants with more persons playing multiple support roles and fewer persons providing tangible assistance.
Understanding how support relationships impact well-being among persons adjusting to HBOC risk, and the particular role of family in that process, will facilitate developing appropriate management approaches to help cancer-prone families adjust to their cancer risk. Published in 2008 by John Wiley & Sons Ltd.
The practice of medicine will become increasingly familial, rather than individual, as additional gene variations associated with specific diseases are identified [1,2]. Therefore, knowledge of family and other pertinent social relationships should play a wider role in the genetic counseling process . For women from families genetically predisposed to breast and ovarian cancer, the risk of developing cancer involves many factors and affects multiple people’s lives. Thus, we might expect successful coping to take on a more communal orientation. This was seen on the Facing Our Risk of Cancer Empowered (FORCE) website for women who carried BRCA1/2 mutations. These women, many of whom expressed dissatisfaction with the support that they received from their physicians, used the FORCE message board as a source of information and a means for communal coping .
Coping is a dynamic process involving initial appraisal of an event (e.g. learning about family history of breast–ovarian cancer or BRCA1/2 carrier status through genetic testing) as threatening; experiencing cognitive, physiological, and emotional effects of stress; and taking action to cope with the threat using available resources [5,6]. The use of social support is one specific coping strategy that has been studied extensively [7,8]. Although definitions of social support vary across studies, most define it as interactions with family and friends who provide informational, tangible, and/or emotional assistance [9–11]. The concept of social support encompasses the existence and quantity of social relationships, the functional content of support relationships, and the structure of interactions (e.g. multiplexity, reciprocity, or shared supports) among support network members [12–14].
The literature has focused primarily on the psychosocial impact of genetic testing and the role of friends and family in the support process [15,16]. Motives for communicating about genetic risk include both the desire for information exchange and receipt of emotional support . Several studies have indicated that receipt of support from one’s partner and family of origin is associated with reduced psychological distress [18,19]. Although these studies examined using social support as a coping mechanism, they did not investigate the size, function, or structure of the social support network of women at risk of carrying a deleterious BRCA1/2 mutation. The current paper addresses this gap in the literature by examining the role of support network structures that represent communal coping strategies among sisters from hereditary breast and ovarian cancer (HBOC) families.
Communal coping is a conceptual model that incorporates the conjoining, interactive support processes that families use to cope with a shared threat [20–22]. Some individuals are primarily social in their response to problems, defining themselves in terms of their interconnectedness and relationships with their family and friends [13,23]. When faced with a shared health threat (e.g. perceived hereditary disease risk), family members may cooperate via reciprocal exchange of support or the use of shared support resources [22,24].
Individual coping focused on social integration differs from communal coping relative to the appraisal of the health threat and the actions engaged in to alleviate the threat . Individual coping may focus on social integration when the health threat is perceived as an individual, rather than a joint, problem and support is sought from others to alleviate stress; or the health threat is appraised as shared, but the responsibility to adapt rests with the individual. In contrast, communal coping involves both shared appraisal of the threat and shared actions to alleviate the threat. This process can prompt the use of shared resources and reciprocal exchange of support to confront the stressor .
A communal orientation to coping may be particularly relevant for sisters in HBOC families. Women are more likely than men to see themselves as interconnected with others and to see their actions as socially interwoven . Women seem to benefit more than men from relationships with same-sex relatives [12,25]. Women with sisters report significantly more contact, affection, and perceived and actual support from siblings than do men or women without sisters . Within HBOC families, women are more likely to discuss their risk with sisters than with brothers [27,28]; most sisters share close relationships and communicate openly about their risk . Sisters derive support from the process of sharing their BRCA1/2 genetic test results with each other , and may benefit more from collectively coping with risk because they share the risk of being a mutation carrier and developing breast or ovarian cancer.
The communal coping literature is composed of theoretical [21,22], qualitative [20,29], and quantitative studies that focus on the individual . To our knowledge, no one has quantitatively analyzed support processes among individuals who are coping collectively, primarily due to methodological challenges. The interdependent nature of the communal coping model requires methods that can quantitatively characterize the connections among those coping together. The social network approach provides a conceptual framework for quantifying interpersonal relationship patterns and their effect on behavior .
This report uses social network methodologies to capture social integration, multiplexity, reciprocity and shared resource constructs, and to examine their relationship to sisters’ psychological adaptation to HBOC risk. We aim to investigate whether sisters from HBOC families share appraisals of cancer threat (as measured by perceived risk and worry regarding breast and ovarian cancer) and examine whether/how utilization of social integration and/or communal coping among HBOC sisters enhances psychological adaptation.
The present sample comprises 65 sisters from 311 families with known deleterious BRCA1/2 mutations selected from 186 participants in an IRB-approved National Cancer Institute (NCI) protocol that is investigating new breast cancer screening methods. The families were primarily non-Hispanic, Caucasian (N=30), with one Hispanic/Latino, Caucasian family.2 Eight families had deleterious BRCA2 mutations and 23 had BRCA1 mutations. Sample characteristics are presented in Table 1.
The Breast Imaging Study is a 4-year, prospective cohort study of women from families with known BRCA1/2 mutations. Eligible women were between ages 25 and 56, and had a known deleterious BRCA1/2 mutation, or were first- or second-degree relatives of BRCA1/2 mutation carriers or relatives of individuals with BRCA-associated cancers in mutation-positive families. Female participants were recruited between 2001 and 2007 from families participating in a long-term prospective study of HBOC, self-referrals in response to media advertising in the Washington, DC area, or referrals from physicians or genetic counselors. All participants had received prior genetic education and counseling, and most had undergone clinical genetic testing, often many years prior to participating in the Breast Imaging Study. A clinician at the Warren Magnusen Clinical Center of the National Institutes of Health obtained informed consent and conducted a medical history and physical examination. During this visit, participants completed questionnaires regarding the psychosocial impact of being a member of a family at high genetic risk of cancer, as well as the Colored Eco–Genetic Relationship Map (CEGRM).
The CEGRM (Figure 1) is a visual research tool to assess social interactions and support exchanges between family members and friends of participants . An investigator (J.P. or L.H.) administered a 20–30 min semi-structured interview using a genetic pedigree as a template [32,33]. The CEGRM operationalized social support resources provided by family and friends to participants, who indicated support exchanges regarding information, tangible aid, and emotional support by placing colored stickers onto the pedigree next to the relevant individuals. CEGRMs were scanned into digital images; social exchange data were coded into databases by two of three investigators (N. K., A. L., L. H.) for subsequent social network analyses with inter-rater reliability exceeding 0.96 for all support functions. Investigators’ hand-written comments on the CEGRM scripts were used to clarify discrepancies in coding.
Demographic data included age, partnered status, race/ethnicity, education, whether mutation status was known at assessment, previous cancer diagnosis, and the number of first- and second-degree relatives with breast/ovarian cancer.
Psychological distress was measured using the Brief Symptom Inventory-18 (BSI-18; ) administered at the baseline visit immediately preceding any evaluations. A standardized, validated instrument, it assesses the respondent’s overall current psychological distress levels. Three six-item subscales measured somatization, depression, and anxiety symptoms, each with reliability indices ranging from 0.74 to 0.84 . Normative T-scores derived from general population norms were computed for each BSI-18 subscale.
Appraisal of the cancer threat was assessed using two constructs: perceived cancer risk and cancer worry. Perceived risk of developing cancer reflects comparative and absolute assessments of cancer lifetime risk [35,36]. On the basis of Shiloh and Ilan , we constructed a perceived risk index by summing standardized comparative and absolute risk assessments. PRIs were constructed for breast and ovarian cancer separately with reliability scores of 0.74 and 0.84 for breast and ovarian cancer, respectively. Cancer worry was measured using the Lerman Cancer Worry Scale, a three-item scale that quantifies concerns about developing cancer and the impact of cancer worry on mood and daily functioning . It has high internal consistency (α=0.86). Questions were asked separately for breast and ovarian cancer, with mean scores computed for each; higher scores indicated higher cancer worry levels.
Social integration and communal coping characteristics were derived from the CEGRM process described above. We considered three functional types of social exchanges: information exchange, tangible assistance, and emotional support. Additionally, multiplex relations (the provision of multiple support roles by individual network members ) were examined, hypothesizing that such persons might be considered as providing more intimate, closer support . In this analysis, multiplexity represents those persons who provide information, tangible assistance, and emotional support; it is the intersection of the three support functions (Figure 2(a)). All indices described below were constructed for informational, tangible, emotional support, and multiplex relations.
Social integration was measured using the size of participants’ social support networks (i.e. the number of persons providing each support type).
Reciprocity is a family-level variable. For the 25 families with two participating sisters, these indices indicated whether both sisters selected each other as providing a given support type. In six families with three or more sisters participating, these indices represent the proportion of participating sisters that reciprocate a given support relation.
Shared supports is a family-level variable indicating the number of persons providing support to two or more participating sisters. The shared support indices exclude participating sisters from the computation to reduce collinearity with the reciprocity indices.
Descriptive statistics were used to describe participants’ psychosocial features and their social support networks’ structural characteristics. All 65 sisters considered within these analyses had complete data; thus, no responses were lost due to missing data within the presented analyses.
HLM 6.02 , a statistical program for multilevel modeling, was used to fit all statistical models. HLM accounts for the clustered structure of the data, with sisters nested within their families of origin. All fitted models were random-intercept models because the small sample size made it difficult to estimate random-slope models. To assess each variable’s statistical importance, a Wald statistic was computed based on the robust standard errors . Standardized coefficients  were estimated to permit comparisons of effects for each subscale analysis.
Inferential analyses focused on two questions. First, we examined whether sisters shared appraisals regarding their cancer risk. These analyses examined the intraclass correlation coefficients (ICCs) for risk perceptions and cancer worry regarding breast and ovarian cancer. Large ICCs indicated significant agreement among family members regarding their risk and worry. We also evaluated whether knowledge of being mutation-negative was related to participants’ appraisals of risk and worry.
Next, we examined associations between social integration and communal coping with psychological distress. Given the small sample size, we used strategies that minimized the number of parameters being fit while allowing us to examine the joint contributions of social integration and communal coping variables. Initial analyses examined potential covariates. Candidate covariates included age, partnership status, knowledge of mutation-negative status at assessment, personal cancer history, family history of breast–ovarian cancer, and whether the participant was from a newly recruited family rather than NIH’s historical cohort. Our final analysis was limited to those demonstrating a bivariate association with at least one distress measure. A block modeling approach, in which social integration and communal coping variables were entered into the model in blocks, evaluated which measures exhibited trends with psychological distress, controlling for important covariates. The final set of models investigated the importance of those social support and communal coping indices that exhibited trends within the block modeling analysis.
Regarding breast cancer, there was significant intra-family correlation for perceived risk (ICC=0.30; χ2(30) =58.97, p=0.001) and cancer worry (ICC=0.23; χ2(30) =49.60, p=0.01) suggesting that sisters had similar perceptions of breast cancer risk and worry. However, there was no significant within-family correlation for perceived risk of developing ovarian cancer (ICC=0.02; χ2(30) =40.09, p = 0.10) or ovarian cancer worry (ICC<0.01; χ2(30) = 26.98, p>0.50).
We assessed whether there were significant differences in the appraisals of participants who knew that they were mutation-negative at the time of assessment versus other participants (e.g. known mutation carriers, those tested after study enrollment, or who declined testing). Analyses controlled for personal history of cancer and whether the kindred was newly recruited or a family from NIH’s historical cohort. Participants who knew that they were mutation-negative at assessment appropriately perceived their risk of developing breast and ovarian cancer as significantly lower then those who were known carriers or unaware of their mutation status (t(60) = −6.36, p<0.001; t(60) = −2.36, p = 0.02, respectively). However, there were no significant differences in breast or ovarian cancer worry between those who knew that they were mutation-negative at assessment and the other participants (t(62) = −1.52, p = 0.13; t(60) = −1.01, p = 0.32, respectively).
On average, study participants did not appear to be psychologically distressed when compared with general population norms (Table 2). Participants discussed their cancer risk with more members of their support network than the number from whom they received tangible assistance and/or emotional support. Sister participants reciprocated information exchange more than other support relations, had many shared information partners, but had far fewer shared supports providing tangible assistance and/or emotional support.
There was significant within-family correlation for anxiety (ICC = 0.32; χ2(29) = 60.94, p = 0.001) and somatization (ICC = 0.20; χ2(29) = 46.15, p = 0.02), but not for depressive symptoms (ICC = 0.02; χ2(29) = 29.38, p = 0.45).
We conducted a preliminary analysis of bivariate associations with potential covariates discussed previously to identify variables for inclusion in the primary analyses. Knowledge of mutation-negative status prior to assessment, participant cancer history, partnership status, and newly recruited status were important correlates of at least one BSI-18 subscale, and were included in subsequent analyses.
The social integration and communal coping indices were entered into the multilevel model in blocks consisting of the information, tangible assistance, emotional support, and multiplex relations (Table 3). Those variables demonstrating an association within the block models were entered into a final model to assess joint contributions across the social integration and communal coping indices (Table 4). For anxiety, the final model indicated that neither the size of the emotional support network nor the number of shared emotional supports among participating sisters was significant when considered jointly. However, when considered separately, both number of network members providing emotional support and the number of shared emotional supports among participating sisters were negatively correlated with anxiety scores. The final model for somatization yielded a positive trend for the number of shared information partners among participating sisters, and a significant negative association for the number of shared emotional supports. Finally, for depression scores, there was a positive association for the size of participants’ tangible assistance network and a negative association for the number of persons playing multiple support roles for the participant.
Understanding the complex interplay between the social context that surrounds family members and their psychological functioning has been relatively understudied, despite its potential importance in providing care to at-risk families. In part, this may stem from difficult methodological issues in defining and assessing social context. Yet, it is important for researchers to understand how the familial support context is related to the psychological functioning of women having genetic testing for BRCA1/2. Toward this end, we investigated whether sisters from HBOC families share appraisals of risk and worry regarding breast and ovarian cancer; and examined whether psychological adaptation is improved by the use of social integration and cooperative support resources among sisters from HBOC families.
Significant within-family correlation was found for breast cancer risk and worry, suggesting that sisters perceived a shared threat for breast cancer. The finding that those women who knew they were mutation-negative prior to assessment viewed themselves as being at significantly reduced risk of both breast and ovarian cancer versus carriers and mutation-unknown participants is consistent with the literature . However, these women also reported levels of cancer worry that were similar to carriers and mutation-unknown participants. Although previous studies suggest that mutation-negative women experience a decrease in cancer worry following genetic testing , the current results suggest that it may be difficult for mutation-negative women to transition from a lifetime of worrying about breast/ovarian cancer risk based on their strong family history to believing themselves at general population risk .
The significant intra-family correlation of anxiety and somatization scores indicates that sisters share levels of distress. Interpersonal attachments expose individuals to stress experienced by others in their network through the processes of contagion [44–46], in which interconnected individuals indirectly transfer their stress or successful adaptation to stress to another person . Communal aspects of coping were associated with anxiety and somatization in our data, suggesting that the overlap in support resources among sisters may be a mechanism through which sisters develop shared levels of anxiety and somatization. Although a causal relationship cannot be determined within the current study design, future research will aim to understand whether shared support resources do lead to a contagion process within families. Interestingly, depression did not exhibit this clustering pattern, suggesting that this mood disorder is not related to sisters’ interactions. That communal coping indices were not associated with depressive symptoms further suggests that depression may represent a psychological construct better suited to individual rather than communal approaches to coping.
Larger numbers of shared emotional supports were inversely associated with anxiety and somatization scores. Considering this effect with the clustering of anxiety and somatization levels, sisters who share a larger number of emotional supports have lower and similar levels of anxiety and somatization, whereas sisters with fewer shared emotional supports have higher and similar levels of anxiety and somatization. Additionally, access to larger numbers of support resources was negatively associated with anxiety. The potential impact of emotional support on anxiety and somatization and the lack of an association with tangible assistance indicate the value of intervention strategies that encourage developing coordinated emotional support resources within the family to reduce anxiety among sisters adjusting to their HBOC risk. This may be particularly relevant in families where all sisters or only a single sister carries the deleterious mutation .
Larger numbers of shared informational supports correlated positively with somatization scores, suggesting that communal coping might undermine coping effectiveness by exacerbating somatization symptoms . These results highlight the potential for negative adaptation when sisters communicate about their HBOC risk jointly with many family members. Being socially connected is highly beneficial to the coping process; however, this can incur a cost if the stress spills over onto others [20,49,50]. Alternatively, sisters with higher levels of anxiety may feel a greater need to discuss their risk of HBOC with many family members. Our results suggest a possible need for the provision of distress-reducing resources for sisters with a highly clustered risk information exchange system within the family.
We were surprised that reciprocity in support among sisters was not significantly associated with reduced distress. Some research suggests that actual support is not as meaningful as perceived support availability in encouraging successful coping [51,52]. In this report, we examined the reciprocity of support based on the overlap between participating sisters’ responses on their CEGRM assessment. We did not assess participants’ perceptions of whether support was reciprocated, a concept related to mutuality . This finding represents a measurement limitation that presents an avenue for further inquiry. The Support Bank Theory , in which reciprocity is construed as a set of payments and withdrawals that evolve over time, a process not tractable in cross-sectional data, represents an alternate explanation.
Several important limitations of this study warrant consideration. First, our analysis was based on cross-sectional data from a relatively small sample of sisters from HBOC families. Although the sister dyad design, use of CEGRM data, and social network methods are unique, the lack of prospectively collected information for this initial analysis limits inferences regarding the directionality of the observed associations. Thus, we cannot determine whether social integration and communal coping reduce distress or whether the psychological context of distress places constraints on families’ social environments. Future research using a longitudinal design will permit disentangling the directionality of these effects, and provide guidance on how to integrate these findings into genetic education and counseling efforts tailored to the needs and social context of the family.
In addition, there are several measurement concerns with regard to the social integration and communal coping indices. It has been argued that the strength (or quality) and duration of the relationship are important characteristics to consider . Unfortunately, the current CEGRM implementation does not capture these factors. With regard to the communal coping indices, not all sisters within each family participated in the current study, limiting our ability to measure reciprocity in support comprehensively. However, non-participating sisters were included in the shared support index as well as the quantity of available supports; therefore, their role in the coping process was not completely ignored. Finally, we were unable to consider shared support from non-family members because there was insufficient information about the identities of participants’ non-familial relationships to link information across sisters. There almost certainly are shared supports among friends or coworkers that were not detected in these analyses.
Despite these limitations, this study represents an important early step in examining a communal framework for understanding coping with hereditary disease. Our findings suggest that the quantity, function, and communal aspects of social exchange are differentially correlated with levels of anxiety, somatization, and depression. Further, coordinated support resources within the family may be indicative of either adaptive processes, as in shared emotional supports, or maladaptive when considering the exchange of hereditary risk information. Families are complex social systems that can facilitate or impede the coping process; the data presented herein suggest there may be real value in elucidating the social context in which risk communications and support processes occur. These results have important implications to health-care providers as they facilitate adaptation for at-risk sisters and their families. The insight gained from prospective, follow-up studies hold genuine promise for improving clinical management of the complex psychological and behavioral issues occurring not only in families where sisters are at risk for developing breast/ovarian cancer but also for families at increased risk for other inherited cancers as well.
We are very grateful to the families, and the sisters in particular, who participated in NCI protocol 01-C-0009: Breast Imaging Screening Studies in Women at High Genetic Risk of Breast Cancer: Annual Follow-up Study and NCI protocol 78-C-0039: Clinical, Laboratory, and Epidemiologic Characterization of Individuals and Families at High Risk of Cancer. We would like to thank Sato Ashida, Andy Baxevanis, Alan Guttmacher, Kim Kaphingst, Aunchalee Loscalzo, Colleen McBride, Suzanne O’Neill, and Bruce Simons-Morton for their comments on an earlier draft of the article; and Ron Kase, Ann Carr, Kathy Nichols, Nicole Dupre, and Usha Singh of Westat for their help with data management. This research was supported, in part, by funding from the Intramural Research Program of the National Human Genome Research Institute, by funding from the Intramural Research Program of the National Cancer Institute to NCI’s Clinical Genetics Branch, and by support services contracts NO2-CP-11019 and NO2-CP-65504 with Westat.
†This article is a US Government work and is in the public domain in the USA.
1There were two sets of sisters who are cousins from one family. Each sister set is considered as a separate nuclear family in these analyses.
2Two sisters from one family provided differing race and ethnicity information: one indicated that she was Hispanic and provided no race information; the other indicated that she was White/non-Hispanic.
3Seventy sisters provided CEGRM assessments that were used to construct the communal coping indices; however, the 65 sisters who provided BSI-18 measurements are the focus of the primary analyses.