Abstract
This cross-sectional study investigated attachment style, coping strategies, social support, and posttraumatic growth (PTG) in 54 cancer survivors. Secure attachment was significantly associated with active coping, positive reframing, and religion, and these were all associated with PTG. Insecure types of attachment and social support variables were unrelated to PTG. Regression analysis suggests that positive reframing and religion as coping strategies may mediate the relationship between secure attachment and PTG.
The Diagnostic and Statistical Manual of Mental Disorders-IV (American Psychiatric Association, 2004) now includes life-threatening illness as a possible trigger for posttraumatic stress disorder (PTSD). Much of the research on the impact of cancer has focused on the negative effects. However, researchers have begun investigating the positive effects associated with being diagnosed with cancer, providing a more comprehensive understanding of the impacts of cancer. Tedeschi and Calhoun (1996) originally described posttraumatic growth (PTG) as how “survivors of traumatic events perceive personal benefits, including changes in perceptions of self, relationships with others, and philosophy of life, accruing from their attempts to cope with trauma and its aftermath” (p. 458). Trauma associated with a life-threatening illness such as cancer can be a trigger for PTG, but although specific correlates have been identified, pathways to PTG remain unclear.
Research has identified approach-oriented, active, and social support coping strategies as antecedents to posttraumatic growth (PTG) in cancer survivors (Bellizzi & Blank, 2006; Sears, Stanton, & Danoff-Burg, 2003; Swickert & Hittner, 2009; Thornton & Perez, 2006; Widows, Jacobsen, Booth-Jones, & Fields, 2005). Conversely, availability of support from spouses, family members, and friends has also been found to be related to higher scores of PTG in survivors of cancer (Cordova, Cunningham, Carlson, & Andrykowski, 2001; Leung, Gravely-Witte, Macpherson, Irvine, Stewart, & Grace, 2010; Thornton & Perez, 2006; Weiss, 2004).
These correlates with PTG have also been found to be related to attachment. For example, Ognibene and Collins (1998) observed that securely attached adults tend to seek more social support as a way of coping following a stressful event; whereas, insecurely attached adults either tend to not seek out support or engage in more avoidant coping strategies. To this end, secure attachment to others in adulthood likely coincides with a sense of trust and confidence in oneself and others which may promote more frequent use of positive coping strategies and emotional expressiveness in times of distress. More specifically, research has found that individuals’ attachment styles predict coping strategies used when managing a stressful event (Mikulincer & Florian, 1995). Although it has not been examined in this context, attachment styles may also be predictors of PTG through a common association with coping and support factors.
Research has demonstrated that attachment styles predict psychological adjustment for individuals with chronic illnesses (Hamama-Raz & Solomon, 2006; Turner-Cobb et al., 2002). Although PTG is not conceptualized as an adjustment to a traumatic event but rather as a response to a traumatic event, there appears to be a theoretical relationship between these two constructs, and this relationship may be mediated by coping and social support. Because attachment styles have been found to directly influence individuals’ coping strategies and the availability/utilization of social support, they may indirectly influence growth following a traumatic event such as cancer via pathways that include the use of active and positive coping strategies and social support. However, no published research has yet examined this theoretical relationship between attachment styles and PTG following a diagnosis of cancer.
The aim of this study is to investigate the relationship between attachment styles and PTG and to determine if social support and/or coping strategies act as mediators between these variables. The paucity of research on attachment styles and PTG has fostered the present exploratory study as the next step in understanding the pathways to PTG. We resolved to address two research questions: 1) Is there a relationship between attachment styles and PTG? 2) Does active or positive coping and/or social support mediate the relationship between attachment styles and PTG in cancer survivors?
Method
Participants
Participants were recruited through advertisements placed on Internet websites (craigslist.com, facebook.com) and sent through a University email listserv and an email registry of cancer survivors at a local hospital. Additionally, a snowball sampling strategy was used by inviting cancer survivors to refer other cancer survivors whom they knew. Eligibility was defined as adult survivors of cancer (individuals diagnosed at the age of 18 or later) currently in remission without any prior relapses who have undergone and completed treatment. Finally, participants must not have reported any current clinical psychiatric diagnosis and must have reported being in otherwise good general health. Fifty-seven individuals expressed interest in participating, but two did not complete one or more of the measures, and one person agreed to be in the study but did not return the survey packet. The final sample consisted of 54 participants.
Procedures
All procedures and survey measures were approved by the appropriate Institutional Review Board. Participants were sent a consent form and a questionnaire packet. All survey data were collected via self-report paper questionnaires, which took approximately 20 minutes to complete. Data collected included demographic information, disease related information, and four self-report measures evaluating posttraumatic growth, attachment style, coping strategies, and social support. Measures were counterbalanced to control for order effects.
Measures
Demographic information
Age, gender, ethnicity, marital status, level of education, number of children, and number of children living at home were collected.
Disease-related information
Age at diagnosis, type of cancer, time since diagnosis, type and duration of treatment were collected. Additionally, as a potential control variable, participants responded to a question written for this study: “Is there any history of cancer in your family? Please explain.” Participants who responded yes to this item and identified a parent, grandparent, or sibling were categorized as having a history of cancer in the family. Finally, participants reported their general physical health in response to “Please indicate in general how your physical health is today,” which was scored on a Likert scale from 1 (very poor) to 5 (excellent).
Posttraumatic growth
The Posttraumatic Growth Inventory (PTGI; Tedeschi & Calhoun, 1996) measures reports of positive changes experienced by individuals following a traumatic event. This scale consists of 21 items representing five subscales: Relating to Others, New Possibilities, Personal Strength/Growth, Spirituality, and Appreciation for Life. Items are scored on a six-point Likert scale ranging from 0 to 5 with higher scores indicating more change. The overall Cronbach’s alpha in this study was .92. All five subscales were found to be significantly correlated with each other (r’s range from .39 to .63), and strongly correlated with the overall PTGI score (r’s range from .74 to .86). As a result and because the aims of this study focus on general growth following cancer diagnosis, the overall PTG scores were used.
Attachment styles
The Measure of Attachment Qualities (MAQ; Carver, 1997a) consists of 14 items measuring four aspects of attachment: avoidance, ambivalence-worry, ambivalence-merger, and security. Item responses were general ratings of attachment behaviors with others and were not specific to the participant’s cancer diagnosis. Items are scored on a four-point Likert scale from 1 to 4 with higher scores indicating greater adherence to that particular attachment style. Carver (1997a) reports good convergent validity with other measures of attachment and test-retest reliabilities and adequate internal consistencies. In this study, the Cronbach’s alphas for the four subscales ranged from .64 to .72.
Coping
Coping strategies were assessed using the Brief COPE Inventory (Carver, 1997b) which consists of 14 subscales, each with two items. Participants responded to items regarding how they have coped with issues in their lives associated with their cancer since their diagnosis. Because of the specificity of coping associated with their cancer, this index attempts to measure coping strategies as, at least in part, situational and not necessarily dispositional. Participants responded to each item on a 4-point Likert scale ranging from 1 to 4 with higher scores indicating greater use of that particular coping strategy. In this study, Cronbach’s alpha for the entire scale was .80, and the subscales ranged from .50-.90.
Social support
The Medical Outcomes Study (MOS; Sherbourne & Stewart, 1991) Social Support Survey is a 19 item measure of functional social support. This measure differs conceptually from use of support as a coping strategy in that it analyzes the availability of four different types of support (emotional/informational, tangible, affectionate, and positive social interaction). It was designed for use with populations of patients with chronic health conditions and responses are scored on a Likert scale from 1 (none of the time) to 5 (all of the time). The overall Cronbach’s alpha was 0.96. A principal components analysis was conducted on the 19 items, and all items loaded strongly on one component (.598 - .880) which explained 61.2% of the variance in the scale. Additionally, a correlation analysis of all subscales and the total MOS score revealed that all correlations were significant at the p < .001 level, suggesting a one-dimensional measure of support. As a result, the overall MOS score was used in analyses.
Data analysis
Descriptive statistics, correlations, analyses of variance, and hierarchical regression models were employed to test relationships among the primary variables. Correlation coefficients were calculated among all independent variables to test for multicollinearity. Correlation and analysis of variance procedures were conducted to determine associations of demographic and disease related variables with the dependent variable for inclusion as covariates in the final regression models.
Results
The mean age of the sample was 52.8 (SD = 10.5), and the mean age at diagnosis was 48.3 (SD = 10.9). The mean time since diagnosis was 4.5 years (SD = 2.8). Frequencies for other demographic and disease-related variables are presented in Table 1.
Sample Demographic and Disease-Related Characteristics (n = 54)
Analysis of attachment style, coping and social support with PTG
The four attachment subscales, the coping subscales, and the MOS were subjected to correlational analysis with the total PTGI scores. Of the four attachment subscales, only secure attachment was significantly correlated with PTGI (r = .37, p = .006). Active coping, (r = .35, p = .010), positive reframing (r = .50, p < .001), and religion (r = .48, p < .001) were significantly correlated with the PTGI scores. Of note, even though one subscale of the PTGI measures religious and spiritual change, religion as a coping strategy was still significantly correlated with PTGI when the religion/spirituality subscale was removed. Additionally, secure attachment was significantly correlated with the aforementioned three coping variables (r = .47, p < .001; r = .31, p = .022; r = .45, p = .001 respectively). The MOS (r = .22, p = ns.) was not significantly related to PTGI scores.
Based on these findings, the only regression model conducted included secure attachment as the independent variable, active coping, positive reframing, and religion as mediator variables, and PTGI as the dependent variable.
Regression analyses and test for mediation
Regression analysis was conducted controlling for demographic on an a priori ground (age, gender, marital status, and education) and disease-related variables (months since diagnosis and type of cancer) in the first step of the model which accounted for 13.7% of the variance in PTGI scores (see Table 2). Secure attachment was entered in step two, and the coping variables entered in step three. The model as a whole accounted for 46.1% of the variance in PTGI scores. Secure attachment accounted for an additional 8.3% of the variance after controlling for the demographics and disease related variables, and this increase was statistically significant (F = 4.801, p = .034). The coping variables accounted for an additional 24.1% of the variance (F = 6.269, p = .001) when added to the model, and positive reframing (β = .31, p = .045) and religion (β = .46, p = .006) were significant contributors to the model, but active coping was not. With the inclusion of these coping variables, the standardized beta coefficient of secure attachment was reduced from .31 to .02 and was no longer statistically significant suggesting mediation.
Regression Analysis: Demographics, Disease Related Variables, Secure Attachment, and Constructive Coping on PTGI
Notes: n = 54; * p < .05; ** p < .01
Dummy coding: Gender 0 = male, 1 = female, Marital status 0 = single, separated, divorced, or widowed, 1 = married or living with partner, Education 0 = less than a four-year degree, 1 = four-year college degree or greater, Cancer type 0 = other, 1 = breast cancer
Discussion
The aim of this study was to examine the relationship between attachment styles and PTG and to determine if social support and/or coping strategies act as mediators between these variables. The results reveal noteworthy findings through exploratory regression analyses.
This research adds to the existing body of knowledge on PTG in cancer survivors by investigating the relationship of attachment styles and PTG in cancer survivors, an area that has not been reported in professional journals. Active coping, positive reframing, and religion were related to secure attachment, and all of these factors were related to PTG. Interestingly, measures of social support (availability and use of) and insecure attachment were not related to PTG. In line with prior research (Bellizzi & Blank, 2006; Sears et al., 2003; Thornton & Perez, 2006; Widows et al., 2005), individuals who employed active coping, planning, positive reframing, and acceptance techniques in dealing with their diagnosis of cancer reported greater PTG. This demonstrates that coping strategies that are functional, positive, or constructive may foster PTG in cancer survivors.
However, contrary to findings from prior research (Cordova et al., 2001; Thornton & Perez, 2006; Weiss, 2004), perceived social support and use of social support as a coping mechanism were not found to be associated with PTG. It may be that for some individuals, use of such support resources may indicate other barriers (e.g., depression, physical ailments and financial burdens) which may be unrelated to or even impede the development of PTG, but more research is needed to better understand these contradictory findings.
Secure attachment was found to be positively associated with PTG. Although the correlation between avoidant attachment and PTG was negative, the relationship approached but did not reach statistical significance. Of note, these findings support the finding from a study that evaluated the relationship between attachment and PTG in political prisoners in which a positive correlation between secure attachment and PTG was found (Salo, Qouta, & Punamäki, 2005). The findings by Salo and colleagues, along with the findings from this study, suggest that the relationship between attachment and PTG may not be trauma specific.
Pathway and mediation analyses
The focus of the exploratory regression model was to test the relationship of secure attachment style and coping variables with PTG. After controlling for demographic and disease-related variables, secure attachment was found to be a significant predictor of PTG but was no longer significant when coping variables were added to the model. This suggests that there is shared variance between secure attachment and the coping variables (specifically positive reframing and religion).
Positive reframing and religion were unique contributors to PTG over and above other variables in the regression models suggesting two potential pathways to PTG. One speculative explanation for this finding is that individuals who are secure in their relationships may be more likely to feel comfort that others can be counted on, which in turn may allow them the confidence to focus on their own needs. In doing so, these individuals may experience fewer external relationship stressors and be in position to better reframe the meaning of the trauma and recognize personal developments related to PTG. In addition to this positive reframing approach, use of religion as a way to cope can be regarded as a means for making sense and meaning of a life-threatening trauma. Park (2009) notes that a strong belief system can function as a meaning making pathway used to positively interpret a traumatic experience
Limitations
It needs to be recognized that the sample size used in this study was relatively small (n = 54) and was not a probability sample. The lack of diversity in this sample is demonstrated by the participants being predominantly older, Caucasian, female, married, and with a high level of education. In addition, the majority of survivors in this sample were breast and prostate cancer survivors. It is further possible that these participants were willing to volunteer for this study because they had previously recognized positive outcomes attributed to their cancer diagnosis. Conversely, individuals who have experienced more negative outcomes related to their experiences with cancer may have not been willing to volunteer for research studies such as this one. This would result in positively skewing the levels of PTG reported in this study. As a result, generalizations of these results cannot be made. Additionally, because of the small sample size, it is possible that some of the non-significant findings, particularly the negative relationship between avoidant attachment and PTGI, may become significant with a larger sample that allows for more power.
Aside from the small sample size, it is important to note that even though retrospective questioning of some survey items was employed, this research is cross-sectional and unable to identify causal associations. Future research employing a longitudinal prospective design may be the best attempt to overcome the limitations of recall bias and causal ordering. In addition, qualitative interviews are likely to provide a more rich understanding of how attachment and coping interact and provide a pathway to PTG.
Conclusion
These findings suggest that individuals who demonstrate secure attachment with significant others in their lives may be better equipped to cope with and manage the diagnosis of and survivorship issues related to cancer. Indeed, attachment style, particularly secure attachment, can be a predictor of posttraumatic growth, but there may be a mechanism to growth at work. Securely attached individuals may be more likely to use active and positive coping strategies, which appears to be one pathway to achieve and recognize growth compared to individuals with less secure attachment styles. However, coping may foster PTG through other pathways, just as securely attached individuals may recognize PTG through means other than active and positive coping.
The influence of social support is less clearly defined and requires more research. Securely attached individuals are more likely to seek and use support as a coping strategy compared to avoidant attached individuals; yet, in this study, avoidant attachment and social support variables were found to be unrelated to PTG. Individuals classified as having a high level of avoidant attachment were found to use support coping less often and have a smaller degree of perceived support available to them compared to securely attached adult cancer survivors, and avoidant attachment was found to be negatively associated with growth, although statistically nonsignificant. These findings warrant continued research in this area.