The Communal Coping Model and Cancer Pain: The Roles of Catastrophizing and Attachment Style

The Journal of Pain, Vol 13, No 12 (December), 2012: pp 1258-1268 Available online at www.jpain.org and www.sciencedirect.com

The Communal Coping Model and Cancer Pain: The Roles of Catastrophizing and Attachment Style Lynn R. Gauthier,*,y Gary Rodin,y,z,x Camilla Zimmermann,z,x David Warr,x,{ S. Lawrence Librach,x,jj Malcolm Moore,x,{ Frances A. Shepherd,x,{ and Lucia Gagliese*,y,z,x *School of Kinesiology and Health Science, York University, Toronto, Ontario, Canada. y Support Systems and Outcomes, University Health Network, Toronto, Ontario, Canada. z Department of Psychosocial Oncology and Palliative Care, Princess Margaret Hospital, & Ontario Cancer Institute, University Health Network, Toronto, Ontario, Canada. x Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada. { Department of Medical Oncology, Princess Margaret Hospital, University Health Network, Toronto, Ontario, Canada. k Mount Sinai Hospital, Toronto, Ontario, Canada.

Abstract: Pain is among the most common symptoms of cancer, with impacts on multiple domains of well-being. Biopsychosocial factors play an important role in adjustment to cancer pain. The Communal Coping Model (CCM), which may elucidate the social context of cancer pain, suggests that people catastrophize to convey distress and elicit support. Attachment style, one’s ability to elicit and respond to available support, may be an important factor, but this has not been tested in people with cancer pain. This study examined pain catastrophizing, attachment style and relational context in relation to perceived solicitous, distracting, and punishing responses of significant others to pain in 191 patients with advanced cancer. Consistent with the CCM, higher pain catastrophizing was related to more frequent solicitous and distracting responses. Pain catastrophizing, attachment anxiety, and significant other type interacted in relation to punishing responses. Higher pain catastrophizing was related to less frequent punishing responses only in anxiously attached patients who identified their spouse/partner as their significant other. These results provide support for the CCM of cancer pain, and contribute to refinement of the model. Future research that includes patients and their caregivers is required to further explicate the social context of cancer pain. Perspective: This article investigates the Communal Coping Model in people with cancer pain. In partial support of the model, we found that pain catastrophizing was related to more frequent solicitous and distracting responses but less frequent punishing responses only in anxiously attached patients who identified their spouse/partner as their significant other. ª 2012 by the American Pain Society Key words: Cancer pain, pain catastrophizing, attachment style, Communal Coping Model, biopsychosocial.

M

ost people with cancer experience pain9 that impacts on multiple domains of function.18 Biopsychosocial factors are important in adjustment to

Received April 5, 2012; Revised September 17, 2012; Accepted October 3, 2012. This project was supported by the Canadian Institutes of Health Research (CIHR MOP-62866) and the Canada Foundation for Innovation to L.G. L.R.G. is a Research Student of the Canadian Cancer Society through an award from the National Cancer Institute of Canada. The authors have no conflicts of interest to declare. Presented, in part, at the 13th World Congress on Pain, Montreal, Canada, 2010. Address reprint requests to Lucia Gagliese, School of Kinesiology and Health Sciences, York University, 4700 Keele St, Toronto, Ontario, Canada, M3J 1P3. E-mail: [email protected] 1526-5900/$36.00 ª 2012 by the American Pain Society http://dx.doi.org/10.1016/j.jpain.2012.10.001

1258

cancer pain.39 Elucidating their interrelationships may improve pain management and well-being.23,39 Some social factors, such as perceived helpful (solicitous/distracting) and negative (punishing) responses from significant others (SO),43 are associated with greater pain, disability, depression, and pain behaviors in patients with chronic nonmalignant pain.5,12,14,22,42,59,63 There is far less investigation in people with cancer. In 1 study, although perceived solicitous, distracting and punishing responses were associated with pain behaviors; only punishing responses were related to pain and depression.3 There may be a unique social context of cancer pain71 that precludes generalizations from people with noncancer pain.

Gauthier et al The Communal Coping Model (CCM) of catastrophizing68 may elucidate perceived support and may clarify the social context of cancer pain. Pain catastrophizing, characterized by hypervigilance to pain and excessive focus on its negative implications,67,68 is associated with greater pain and impaired functioning and wellbeing.4,57,76 The CCM suggests that people catastrophize to convey distress and elicit support.68 Therefore, catastrophizing may have an interpersonal goal of eliciting helpful or better perceived support.5,69,70 Although empirical evidence for the CCM is emerging, the data remain unclear. In people with chronic nonmalignant pain, catastrophizing is related to more frequent solicitous11,13,58 and punishing responses5,11 in some studies, but not others.5,15 In people with cancer pain, catastrophizing is related to greater tangible, but not emotional, support.40 Although difficult to integrate, these findings suggest an interpersonal role of catastrophizing in certain types of support. Relationship to the caregiver may modify these associations,29 but other factors remain unexamined. Attachment style may be important but remains uninvestigated in the CCM. The attachment system develops in infancy based on caregiver interactions and remains stable throughout the lifespan.6,24,55 Two dimensions have been described. High attachment avoidance is characterized by a deactivating style toward threats, minimization of their negative implications, and discomfort with dependency. High attachment anxiety is characterized by a hyperactivating style toward threats, magnification of their negative implications, and fear of rejection and abandonment.8,55,60 Securely attached people are low on both dimensions.55 Attachment style may be important to adjustment to chronic nonmalignant pain.50,54,60 Insecure attachment is associated with greater pain and distress.17,20,47,53 Attachment anxiety is also related to greater catastrophizing in pain-free people and those with nonmalignant pain.48,49,51,52 The relationship between attachment avoidance and catastrophizing is equivocal.48,49,51,52 This remains uninvestigated in people with cancer. Published studies examining attachment style and perceived support in people with nonmalignant and cancer pain are unavailable. In studies of patients with advanced cancer that did not consider pain, insecure attachment is associated with lower perceived social support.35,62 Therefore, attachment style may also be important in patients’ perceptions of SO’ responses to their cancer pain. Since attachment style may operate on catastrophizing and perceived support, incorporating it into the CCM may improve our understanding of individual differences in this model. Given evidence that relational context moderates the relationship between catastrophizing and nonmalignant pain,29 understanding its impact may clarify how the CCM operates in different supportive circumstances in people with cancer pain. Our aim was to refine the CCM of cancer pain by examining relationships between pain catastrophizing and perceived SO’ responses to pain and

The Journal of Pain

1259

the moderating roles of attachment style and relational context. We hypothesized that: 1) catastrophizing would be positively related to attachment anxiety but not avoidance; 2) attachment avoidance and anxiety would be negatively related to solicitous and distracting responses and positively related to punishing responses; and 3) catastrophizing would be positively related to solicitous and distracting responses and negatively related to punishing responses. Attachment style and relational context would moderate these relationships.

Methods Participants Outpatients attending Palliative Care, Pain, Gastrointestinal, Lung, Breast, Gynaecology, and Genitourinary clinics at Princess Margaret Hospital (PMH), a comprehensive cancer center in Toronto, Ontario, Canada, were recruited between May 2006 and December 2010 for a larger study of the impact of cancer pain (we have previously published a separate analysis with a smaller subset of this sample.27) Patients receiving home palliative care through the Temmy Latner Centre for Palliative Care (TLCPC) also were recruited between August 2009 and December 2010. Patients were eligible if they were $18 years old, had advanced cancer, cancer-related pain based on patient report to healthcare provider, and sufficient English fluency to provide informed consent and complete questionnaires. Patients with documented cognitive impairment identified by the physician, medical chart, or cognitive screen38 were not eligible. Ethics approval was obtained from the Research Ethics Boards of the University Health Network, Mount Sinai Hospital and York University.

Procedures Clinic staff at PMH identified eligible patients and determined their desire to be approached. A research assistant (RA) approached patients during their outpatient appointment, explained the study and obtained written informed consent. Eligible patients receiving homecare through the TLCPC were identified by their physicians. These participants were telephoned, and the study was explained to them. The RA visited interested patients in their homes to explain the study and obtain informed consent. Following consent, the RA administered a short cognitive screen, the Short Orientation Memory Concentration Test (SOMC38). Participants scoring <20 were withdrawn. Demographic and clinical information, including prescribed analgesics and nonpharmacologic treatments for pain, was collected from remaining participants. They were given a questionnaire package which they could complete with the help of the RA or on their own at home and return using the provided postage-paid, self-addressed envelope. The RA telephoned those who had not returned their questionnaire packages after 2 weeks. Reasons for participant withdrawal were recorded.

1260

The Communal Coping Model and Cancer Pain

The Journal of Pain

Measures Cognitive Screen The Short Orientation-Memory-Concentration Test (SOMC38) is a valid and reliable 6-item measure of cognitive impairment, memory and orientation to time, person and place.73 It has been used as a cognitive screen in other studies of patients with advanced cancer.62

Pain Pain severity was measured with the Brief Pain Inventory (BPI19), which includes 11-point numeric rating scales (NRS) anchored with the descriptors ‘‘no pain’’ and ‘‘pain as bad as you can imagine.’’ A pain severity score is calculated from the mean of 4 items assessing worst, least, average, and current pain. The validity and reliability of the BPI have been documented extensively.28,61 Cronbach’s alpha in this study was .89. Pain relief from medications or treatments for pain was assessed with a scale ranging from 0 to 100%. To assess adequacy of analgesia, the Pain Management Index (PMI) was calculated by subtracting the participant’s BPI worst pain score from the score assigned to highest level of analgesic prescribed,7,18 according to the World Health Organization’s (WHO) Analgesic Ladder.74

Physical Functioning Comorbidities were measured with the Charlson Comorbidity Index (CCI16) which is a well-established, reliable, and valid measure of 19 co-occurring conditions.21 Higher scores indicate greater comorbidities. Functional status was measured with the Karnofsky Performance Status Scale (KPS37) which is an observer-rated measure ranging from 100 (Normal activity, no complaints, no evidence of disease) to 0 (dead). It is a valid and reliable measure of functional status in people with advanced cancer.75 This was completed by the RA. There were no differences in KPS scores across different RAs (76.3 6 13.0 versus 79.4 6 11.3 versus 82.5 6 8.5; F(2,188) = 2.15, P = .12), suggesting acceptable cross-rater consistency.

Pain Catastrophizing The total score of the Pain Catastrophizing Scale (PCS67) was used to measure rumination about pain, magnification of its negative implications, and a helpless attitude towards one’s ability to manage pain. Higher scores reflect greater pain catastrophizing. The PCS has been validated for patients with nonmalignant pain67 and has been used to measure pain catastrophizing in cancer patients.4 Cronbach’s alpha in this study was .95.

Attachment Style Attachment style was measured with the Experiences in Close Relationships Inventory (ECR8). Two subscales are calculated: The avoidance subscale measures discomfort with closeness, and the anxiety subscale measures fear of rejection and abandonment. Higher scores indicate greater attachment insecurity. The ECR has been used to measure attachment style in patients with cancer or chronic nonmalignant pain with good reliability and

8,48,62

validity. Cronbach’s alpha for the avoidance and anxiety subscales was .89 and .90, respectively.

SO Responses to Pain The Multidimensional Pain Inventory (MPI43) Caregiver Responses Scale instructs participants to identify a SO ‘‘as the person to whom [the participant] feels closest. This includes anyone that [the participant] relates to on a regular or infrequent basis.’’ It also asks whether they currently live with this person. Participants are then instructed to respond to questions about their perception of the frequency of their SO’ solicitous, distracting, and punishing responses to their pain-related communications. Higher scores reflect more frequently perceived responses from SO. The MPI is valid and reliable and has been used extensively among patients with nonmalignant pain.43,72 Cronbach’s alpha for the solicitous, distracting, and punishing responses subscales was .76, .76, and .75, respectively.

Data Analyses Scales with #20% of items with missing responses were imputed with the mean of the participant’s response for that scale.36 The major results did not change between analyses with the imputed and the nonimputed data; therefore analyses with imputed data are presented. Descriptive statistics were used to characterize participants on demographic, clinical, and study variables. Where skewness values exceeded 1, scales were transformed to normalize their distributions. Bivariate analyses were conducted between the main independent (pain catastrophizing and attachment avoidance and anxiety) and outcome variables (perceived solicitous, distracting, and punishing responses). Correlations between demographic and clinical variables and the main independent and outcome variables were calculated to determine variables to adjust for each model. Where a candidate correlate was associated with at least 1 of the independent variables and the relevant outcome variable at P #.05, it was considered for inclusion in the model. Multivariate regression models investigated the relationships of MPI SO type, pain catastrophizing and attachment style on the SO responses outcome variables. Both ECR subscales were included in each regression model based on prior evidence.48 To assess interactions, variables were centered and product terms were created. Variables were entered in blocks, with the centered independent variables entered first, followed by the product terms.2,25,32,33 Lower order 2-way interactions were included in models where 3-way interactions were tested. Significant interactions were investigated by plotting simple slopes for a given variable at high (1 standard deviation (SD) above the mean) and low (1 SD below the mean) values of the other variable.2,25,32,33 All data were analyzed using SPSS v.19.

Results Sample characteristics Between May 2006 and December 2010, 398 patients were approached. Two-hundred and ninety-five consented

Gauthier et al and 103 (25.9%) refused to participate. Reasons for refusal to participate were lack of interest (n = 45, 43.7%), lack of pain despite reporting pain to healthcare providers (n = 18, 17.5%), illness/fatigue (n = 13, 12.7%), lack of time (n = 12, 11.7%), patient-reported language, comprehension, or memory issues (n = 5, 4.9%), and inability to answer emotional questions (n = 1, 1%). Six people (5.8%) did not give a reason for refusing and in 3 cases (2.9%) the family caregiver refused access to the patient. There were no sex differences between those who consented to participate and those who refused, although patients receiving home palliative care were more likely than those attending outpatient clinics to refuse to participate (85.7% versus 22.7%; c21 = 41.08, P # .0001). Of those who agreed to participate, 2 participants scored <20 on the SOMC and were withdrawn. One hundred and two (34.6%) did not return the questionnaire package. Reasons for failing to return the questionnaire package were illness due to disease progression (n = 31, 30.4%), death (n = 28, 27.5%), and lack of interest in continuing to participate (n = 12, 11.8%). Thirty-one participants (30.4%) did not return the questionnaire package for unknown reasons. Compared to those who did not return the questionnaire package, participants who returned the questionnaire package (n = 191) were younger (56.8 6 11.7 years old versus 60.5 6 13.8 years old; t293 = 2.40, P = .02), more likely to be Caucasian (80.1 versus 57.7%; c21 = 16.85, P = .0001), and report English as their primary language (85.3 versus 74.5%; c21 = 5.58, P = .02). More single, married/partnered, and separated/divorced participants returned the questionnaire package than widowed participants (78.3, 66.1, and 64.5% versus 40%, respectively; c23 = 9.15, P = .03), and more participants with lung/thoracic, breast, gynaecologic/genitourinary, and gastrointestinal primary tumors than participants with other cancers returned the questionnaire package (76.2, 73.9, 73.1, and 60.5% versus 52.6%, respectively; c24 = 10.34, P = .04). Those who returned the questionnaire package had higher SOMC (25.97 6 2.20 versus 25.07 6 2.38; t288 = 3.22, P = .001) and KPS scores (80.20 6 10.71 versus 77.07 6 10.62; t288 = 2.37, P = .02) than those who did not return the questionnaire package. Characteristics of participants who returned the questionnaire package and are included in this analysis are listed in Table 1. The average age was 56.8 6 11.7, and 55% were female. Median disease and cancer pain duration was 24 months and 10.5 months, respectively. The average functional status score (KPS) was 80.2 6 10.7 and ranged from 90 to 50. Sixty-six percent of participants had a CCI score of 0. Eighty-eight percent of participants had adequate analgesia according to the PMI (PMI scores $0), and 77.5% reported $50% pain relief (range 0–100%). Despite this, 60.4% reported moderate-tosevere BPI worst pain ($5 on NRS). The MPI SO variable was collapsed to form 2 groups; those who identified their spouse/partner (65.4%) and those who identified another person (29.8%) as their SO. Those who identified their spouse/partner as their

The Journal of Pain

1261

Participant Demographic, Disease and Treatment-Related Characteristics (N = 191)

Table 1.

MEAN 6 SD/N (%) Age Female Ethnicity Caucasian Non-Caucasian Primary language English Non-English Missing Highest level of education completed Elementary or high school College or undergraduate Graduate or professional degree Marital status Single Married or partnered Separated or divorced Widowed Parental status No children Live with child/ren Live separately from child/ren Identify significant other Spouse/partner/companion Housemate/roommate, friend, parent/child/ other relative, other Missing Live with significant other Yes No Missing Primary tumor type Gastrointestinal Gynecologic/genitourinary Breast Lung Other Recruitment site Palliative care clinic Pain clinic Temmy Latner Centre Other solid tumor clinics Disease duration Pain duration Highest WHO class of prescribed analgesic None NSAID/acetaminophen/ASA/adjuvant Weak opioid Strong opioid Undergoing chemotherapy/radiation/hormone therapy Nonpharmacologic treatment for pain* Experience with chronic nonmalignant pain

56.8 6 11.7 105 (55) 153 (80.1) 38 (19.9) 163 (85.3) 27 (14.1) 1 (.5) 65 (34) 98 (51.3) 28 (14.7) 36 (18.8) 127 (66.5) 20 (10.5) 8 (4.2) 49 (25.7) 66 (34.6) 76 (39.8) 125 (65.4) 57 (29.8) 9 (4.7) 139 (72.8) 43 (22.5) 9 (4.7) 46 (24.1) 43 (22.5) 33 (17.3) 32 (16.8) 37 (19.4) 148 (77.5) 20 (10.5) 2 (1) 21 (11) 39.1 6 42.6 18.7 6 28.2 2 (1) 14 (7.3) 58 (30.4) 117 (61.3) 109 (57.1) 110 (57.6) 72 (37.7)

Abbreviations: NSAID, non-steroidal anti-inflammatory; ASA, acetylsalicylic acid. *One or more of transcutaneous electrical nerve stimulation (TENS), biofeedback, acupuncture, massage, exercise, stretching/yoga, breathing exercises, heat/cold application, meditation, visualization, distraction, rest, reiki, reflexology, psychotherapy, support group, naturopathic and herbal preparations, physiotherapy, art therapy, breathing exercises, qi gong, magnetic resonator, chiropractor, biotherapy, prayer.

.03 .12 .14 .19* .06 .29** .23** .10 .05 .07 .23** .11 .02 .18* .52** .09 .05 .33** .03 .13 .03

*Two-tailed P # .05. yNumber of participants with scorable data. **Two-tailed P # .001.

0–9.75 >1–240 0–5.25 .33–6 0–6 0–51 1–6.17 1–5.94

182 190 174 176 166 184 180 180

MPI PUNISHING RESPONSES PAIN DURATION Ny

3.79 6 2.05 18.70 6 28.21 1.06 6 1.18 4.52 6 1.21 2.76 6 1.47 20.50 6 12.34 3.07 6 .96 2.56 6 1.09 BPI severity Pain duration (months) MPI punishing responses MPI solicitous responses MPI distracting responses PCS Total ECR Avoidance ECR Anxiety

Descriptive statistics and Pearson correlation coefficients for study measures are listed in Table 2. We hypothesized that catastrophizing would be positively related to attachment anxiety but not avoidance. Partially consistent with this hypothesis, pain catastrophizing was positively correlated with attachment anxiety (r = .44; P # .001); however, surprisingly, it was also positively correlated with attachment avoidance (r = .29; P # .001). Table 3 displays the results of the multivariate regression analyses with uncentered independent variables. We hypothesized that attachment avoidance and anxiety would be negatively related to solicitous and distracting responses and positively related to punishing responses. Partially consistent with this hypothesis, attachment avoidance was negatively correlated with perceived solicitous responses (Table 2: r = .19, P #.05) and this relationship was retained in the multivariate regression analysis (Table 3: B = .228, P #.05). While it was not correlated with distracting responses in the bivariate analysis, in the multivariate regression analysis,

ACTUAL RANGE

Testing Hypotheses 1 and 2

Table 2.

The MPI punishing responses subscale was positively skewed. A logarithmic transformation normalized the distribution. Correlations and simple regression analyses were conducted with the untransformed and transformed scales (log[MPI punishing 1 1]). The main findings did not change; thus the results for the nontransformed scale are reported for the sake of clarity. Marital status was significantly related to perceived solicitous responses and attachment avoidance (P # .02); however, it was excluded as a covariate in the analyses of the full sample (n = 191) due to colinearity with the MPI SO variable. No other demographic or clinical variable was related to both the relevant outcome variable and at least 1 other independent variable. BPI pain severity and pain duration were also unrelated to all of the outcome variables; therefore, they were excluded from the multivariate models.

Descriptive Statistics and Pearson’s Correlation Coefficients for Study Measures

Data Cleaning and Model Building for Correlational and Multivariate Regression Analyses

MPI SOLICITOUS RESPONSES

MPI DISTRACTING RESPONSES

PCS TOTAL

ECR AVOIDANCE

ECR ANXIETY

SO were older (58.3 6 10.9 years old versus 54.0 6 13.7 years old, P = .01), more likely to live with their SO (86.9 versus 13.1%, P = .0001) and have children (73.4 versus 20.9%, P = .0001) than those who identified another person as their SO (n = 57). Men were more likely than women to identify their spouse/partner than another person as their SO (76.7 versus 56.2%, P = .001). There were no other demographic or clinical differences between the groups. Six participants who reported being married or partnered nonetheless identified another person as their SO. Five of these individuals identified a parent, child, or other relative and 1 identified another person as their SO on the MPI. There were no significant differences between the groups on any of the demographic, clinical or other variables included in this study.

.05 .05 .28** .12 .01 .44** .38**

The Communal Coping Model and Cancer Pain

The Journal of Pain

MEAN 6 SD

1262

Gauthier et al

The Journal of Pain

1263

Multivariate Linear Regression Models: Relationships of Pain Catastrophizing and Attachment Style to Perceived Significant Other Responses to Pain

Table 3.

MPI PERCEIVED PUNISHING RESPONSES B (SE) Step 1 MPI significant othery Step 2 PCS total ECR avoidance ECR anxiety

.159 (.183)

b .086

D R2 .004

.008 (.008) .090 .110 (.094) .094 .297 (.093) .273** .086** F4, 168 = 4.14, P = .003 R2 = .090

MPI PERCEIVED SOLICITOUS RESPONSES b

B (SE) .554 (.195)

.208**

MPI PERCEIVED DISTRACTING RESPONSES

D R2

B (SE)

.055**

.225 (.229)

.020 (.008) .198* .228 (.100) .182* .103 (.100) .087 .058* F4, 167 = 5.32, P = .0001 R2 = .113

b

D R2

.072

.001

.039 (.010) .332** .285 (.117) .194* .140 (.118) .101 .111** F4, 165 = 5.23, P = .001 R2 = .112

Abbreviation: SE, Standard Error of Beta. * P # .05. y0 = nonspouse/partner, 1 = spouse/partner. **P # .01.

We hypothesized that catastrophizing would be positively related to solicitous and distracting responses and negatively related to punishing responses and that attachment style and relational context would moderate these relationships. Pain catastrophizing was unrelated to solicitous responses in the bivariate analysis, but in the multivariate regression analysis, consistent with our hypothesis, it was positively related to solicitous responses (Table 3: B = .020, P # .05). Also consistent with our hypothesis, it was positively related to distracting responses in both the bivariate (Table 2: r = .23, P # .001) and multivariate regression analysis (Table 3: B = .039, P # .01). Counter to our hypothesis about the moderating roles of attachment style and relational context, pain catastrophizing did not interact with attachment anxiety or MPI SO type in relation to perceived solicitous and distracting responses. Pain catastrophizing was unrelated to perceived punishing responses in the bivariate analysis. However, there was a 3-way interaction between pain catastrophizing, MPI SO type, and attachment anxiety (B = .030 (SE = .014), b = .232, P = .04). Post hoc probing of simple slopes revealed that the relationship between pain catastrophizing and perceived punishing responses was conditional on attachment anxiety among those who identified their spouse/partner as their SO (Fig 1A). Pain catastrophizing was negatively associated with per-

MPI Significant Other - spouse/partner MPI Perceived Punishing Respon nses Scale

Testing Hypothesis 3

ceived punishing responses in those with high attachment anxiety but not in those with low attachment anxiety. Although there appeared to be an opposite

2.5 low ECR Anxiety

2

high ECR Anxiety

1.5 1 0.5 0 low PCS (-1SD)

high PCS (+1SD)

Pain Catastrophizing Scale

MPI Perceived Punishing Punis Responses nses S Scale

attachment avoidance was also negatively related to distracting responses (Table 3: B = .285, P # .05). However, counter to our expectation, attachment avoidance was unrelated to punishing responses in the bivariate and multivariate regression analyses. Also counter to our expectations, attachment anxiety was unrelated to solicitous and distracting responses in both the bivariate and multivariate regression analyses. Consistent with our hypothesis about the relationship between attachment anxiety and punishing responses, they were positively related in both the bivariate (Table 2: r = .28, P # .001) and multivariate regression analyses (Table 3: B = .297, P # .01).

2.5

MPI Significant Other - non-spouse/partner low ECR Anxiety

2

high ECR Anxiety

1.5 1 0.5 0 low PCS (-1SD)

high PCS (+1SD)

Pain Catastrophizing Scale

Figure 1. Three-way interaction between MPI significant other type, pain catastrophizing, and attachment anxiety in relation to perceived punishing responses. (A) Predicted values of perceived punishing responses in those who identified a spouse/ partner as their significant other. Simple slopes analysis revealed the slope for high attachment anxiety (11SD) was significant (B = .029 (SE = .013), b = .286, t = 2.15, P = .03) while the slope for low attachment anxiety (1SD) did not reach significance (B = .006 (SE = .014), b = .064, t = .46, P = .65). (B) Predicted values of perceived punishing responses in those who identified a nonspouse/partner as their significant other. Simple slopes analysis revealed the slopes for high and low attachment anxiety were nonsignificant (high attachment anxiety: B = .019 (SE = .015), b = .234, t = 1.22, P = .229; low attachment anxiety: B = .011 (SE = .016), b = .135, t = .68, P = .50).

1264

The Journal of Pain

pattern of findings among those who identified a nonspouse/partner as their SO (Fig 1B), post hoc probing of the simple slopes did not reach significance. Pain catastrophizing, attachment avoidance, and MPI SO type did not interact in relation to perceived solicitous, distracting or punishing responses.

Discussion This is the first study to investigate the roles of pain catastrophizing and attachment style in the CCM in patients with cancer-related pain. Consistent with the CCM, greater catastrophizing was related to more frequently perceived distracting and solicitous responses when we adjusted for attachment style, suggesting that attachment style may have contributed to previous inconsistent findings.5,11,13,15,58 Interestingly, there was not a direct relationship between pain catastrophizing and punishing responses. Instead, this relationship was moderated by attachment anxiety and relationship to the SO. These findings suggest that, in addition to pain catastrophizing, attachment style and relational context are important factors in the CCM of cancer pain. The CCM was only partially supported when we considered patients’ perceptions of punishing responses. If catastrophizing serves to communicate pain and to elicit desired support from SOs, it might be expected to be related to more frequent solicitous responses5 and thus less frequent punishing responses. Although some studies of people with chronic nonmalignant pain have demonstrated that catastrophizing is related to more frequent punishing responses,5,11 others have not found a direct relationship between catastrophizing and punishing responses.15 Furthermore, patients with cancer pain may perceive more frequent solicitous responses and less frequent punishing responses from their significant others compared to people with chronic nonmalignant pain.71 This unique social context of cancer pain makes it difficult to generalize findings from people with chronic nonmalignant pain to those with cancer pain. Therefore, this hypothesis is an extension of the CCM specifically to people with cancer pain. In this study of people with cancer pain, instead of a direct relationship between catastrophizing and punishing responses, we found that it was moderated by SO type and attachment anxiety. In partial support of our third hypothesis, higher pain catastrophizing was associated with less frequent punishing responses, only among anxiously attached patients who identified their spouse/partner as their SO. Those who catastrophize may have a support- and caretaking-demanding interpersonal style.45 This is consistent with the dependency associated with attachment anxiety31,55,60,64 and provides further evidence of similarity between the 2 constructs.49 Perhaps among anxiously attached married/partnered individuals, who may exaggerate threats and their negative implications,8,55,60 catastrophizing serves to alleviate the fear of cancer pain by mobilizing support. An alternate hypothesis might be that catastrophizing would be related to more frequent punishing responses from

The Communal Coping Model and Cancer Pain SOs, but that this relationship is tempered in those with high attachment anxiety who place importance on preserving relationships. However, the positive association between attachment anxiety and punishing responses and the lack of association between catastrophizing and punishing responses in the simple regression analysis do not support this hypothesis. Interestingly, in people with low attachment anxiety, perceptions of punishing responses did not differ across levels of pain catastrophizing, regardless of relational context. It is possible that in those who do not hyperactivate in response to stressors, pain catastrophizing is not the main mechanism deployed to elicit support. Although this is the first study to demonstrate this interaction in relation to punishing responses, other studies have found that catastrophizing is more strongly related to pain among those who live with their caregiver compared to those who do not,29 and that reactions to different types of support depend on attachment style.66 Taken together, these findings suggest that future studies of the CCM should consider attachment style and relational context. Also, consistent with prior research,26 we found that women were less likely than men to identify a spouse/partner as their SO. Given this gender difference, and the importance of relational context, future in-depth examinations of the CCM in different supportive contexts are necessary. While we found that attachment avoidance was related to less frequent solicitous and distracting responses, it played a smaller role than attachment anxiety and did not interact with pain catastrophizing in relation to SO responses. Those with high attachment avoidance may not perceive frequent negative or positive support as they adopt a deactivating style toward stressor management, doubt their partner’s capacity to provide support, and prefer to deal with stressors on their own.8,55,60 If individuals with high attachment avoidance are reluctant to seek support,8,55,60 based on the CCM, we might expect a negative relationship between attachment avoidance and pain catastrophizing. Instead, we found a positive relationship, raising the question of the function of pain catastrophizing in those with high attachment avoidance. A number of explanations can be considered. We found a moderate positive correlation between attachment avoidance and anxiety. Therefore, it may be possible that because of the shared features of attachment anxiety and pain catastrophizing, this relationship reflects shared variance between the 2 attachment dimensions.49 It may also be possible that the goals of catastrophizing differ across attachment style dimensions. Among those with high attachment anxiety, the goals may be to communicate pain and distress, whereas among those with high attachment avoidance the goals remain to be determined and may be better explained by other models.68 Future studies are needed to understand catastrophizing in this attachment context. In the context of chronic nonmalignant pain, solicitous responses to pain behaviors may reinforce pain cata-

Gauthier et al 68

strophizing. Therefore, the use of catastrophizing to elicit support from SOs may be maladaptive, leading to heightened pain, functional limitations, and psychological distress.68 However, among women with pain due to advanced breast cancer, more frequently perceived solicitous responses from partners were not associated with pain severity or depressive symptoms.3 It will be important for future studies to determine whether the use of catastrophizing to elicit solicitous responses from SOs leads to negative outcomes in people with cancer pain. The following limitations should be considered. First, the cross-sectional design precludes statements regarding the directionality of relationships. Longitudinal studies are required to address this. Second, most participants were receiving specialized symptom management, and those who returned the questionnaire package had better functional status than those who did not return the questionnaire package. Therefore, these results may be generalizable only to patients with higher functional status who are receiving specialized symptom management. The KPS was completed by 1 rater, as is typical1,10,37,56; therefore we are not able to calculate interrater reliability. However, given that the KPS has demonstrated good-to-excellent interrater reliability in previous studies,75,77 we are confident that the KPS is a reliable observer-rated measure of functional status. Finally, we only had self-report data from patients. As such, we do not know if the behavioral expression of pain catastrophizing in these attachment and relational contexts is associated with perceptions of supportive responses or the actual provision of support from SOs. Future studies testing the CCM that include both members of the couple are required. In addition to considering patient and caregiver pain catastrophizing,46 these studies should also assess the behavioral expression of catastrophizing, the match between desired and given support and the attachment style of both members of the couple, given the impact this may have on mate selection and maintenance of relationships.34 Consistent with a prior study,29 this study demonstrates that supportive context is important to consider. While spouses may be among the most common of informal caregivers,30,44 other caregiving relationships are also prevalent. Therefore, it is essential that future studies examine the CCM in nonpartnered as well as partnered caregiving relationships. These studies are necessary to determine the success of catastrophizing as an interpersonal pain-coping strategy and its effects

References 1. Abernethy AP, Shelby-James T, Fazekas BS, Woods D, Currow DC: The Australia-modified Karnofsky Performance Status (AKPS) scale: A revised scale for contemporary palliative care clinical practice [ISRCTN81117481]. BMC Palliat Care 4:7, 2005 2. Aiken LS, West SG: Multiple regression: Testing and interpreting interactions. Newbury Park, CA, Sage Publications, 1991

The Journal of Pain

1265

on patients’ and caregivers’ well-being. This may be especially salient in anxiously attached patients who catastrophize about pain. We demonstrated that pain catastrophizing is related to less frequent punishing responses in anxiously attached patients who identify their spouse/partner as their SO. However, because of the hypervigilance and reassurance-seeking associated with anxious attachment,65 and in light of evidence that higher patient pain catastrophizing is associated with higher caregiver stress and compromised wellbeing,40,41 this combination may be associated with negative caregiver outcomes. Studies addressing this could contribute to the development of patient and caregiver interventions.40,41 The present study provides valuable evidence refining the CCM and clarifying the social context of cancer pain. Consistent with the CCM, we have demonstrated that together with attachment style, pain catastrophizing does play an interpersonal role in perceptions of more frequent solicitous and distracting support. However, while we found support for the CCM in relation to punishing responses in anxiously attached patients who identified a spouse/partner as their SO, it was less clear in those who were securely attached and in anxiously attached patients who identified a nonspousal SO. This study also has important clinical implications. An understanding of patients’ attachment style can provide insight into perceived support in cancer patients who catastrophize about pain. The results can also help us to understand who may be at risk for negative or maladaptive supportive responses to pain that may lead to negative outcomes, including increased pain and disability, that have been associated with this type of response in people with chronic nonmalignant pain.5,12,14,22,42,59,63 These data may guide interventions tailored to patients’ supportive needs that target pain catastrophizing and maximize quality of life.

Acknowledgments We are grateful to the staff in the outpatient clinics at Princess Margaret Hospital, the staff of the Temmy Latner Centre for Palliative Care, and we thank Laura Katz and Kim Thao Tran for help with recruitment. We are also grateful to the members of the Cancer Pain Research Unit and the Research Development Seminar for comments on earlier drafts of the manuscript. Most importantly, we are grateful to the participants who contributed their time and efforts to make this study possible.

3. Badr H, Milbury K: Associations between depression, pain behaviors, and partner responses to pain in metastatic breast cancer. Pain 152:2596-2604, 2011 4. Bishop S, Warr D: Coping, catastrophizing and chronic pain in breast cancer. J Behav Med 26:265-281, 2003 5. Boothby JL, Thorn BE, Overduin LY, Ward LC: Catastrophizing and perceived partner responses to pain. Pain 109: 500-506, 2004

1266

The Journal of Pain

6. Bowlby J: Attachment and loss: Attachment. New York,NY, Basic Books, 1982 7. Breitbart W, Rosenfeld BD, Passik SD, McDonald MV, Thaler H, Portenoy RK: The undertreatment of pain in ambulatory AIDS patients. Pain 65:243-249, 1996 8. Brennan KA, Clark CL, Shaver PR: Self-report measurement of adult attachment: An integrative overview, in Simpson JA, Rholes WS (eds): Attachment Theory and Close Relationships. New York, NY, The Guilford Press, 1998, pp 46-76 9. Bruera E, Kim HN: Cancer pain. JAMA 290:2476-2479, 2003 10. Buccheri G, Ferrigno D, Tamburini M: Karnofsky and ECOG performance status scoring in lung cancer: A prospective, longitudinal study of 536 patients from a single institution. Eur J Cancer 32A 7:1135-1141, 1996 11. Buenaver LF, Edwards RR, Haythornthwaite JA: Pain-related catastrophizing and perceived social responses: Interrelationships in the context of chronic pain. Pain 127: 234-242, 2007 12. Campbell P, Jordan KP, Dunn KM: The role of relationship quality and perceived partner responses with pain and disability in those with back pain. Pain Med 13: 204-214, 2012 13. Cano A: Pain catastrophizing and social support in married individuals with chronic pain: The moderating role of pain duration. Pain 110:656-664, 2004 14. Cano A, Johansen AB, Franz A: Multilevel analysis of couple congruence on pain, interference, and disability. Pain 118:369-379, 2005

The Communal Coping Model and Cancer Pain 24. Fraley CR, Shaver PR: Adult romantic attachment: Theoretical developments, emerging controversies, and unanswered questions. Rev Gen Psychol 4: 132-154, 2000 25. Frazier PA, Tix AP, Barron KE: Testing moderator and mediator effects in counseling psychology research. J Couns Psychol 51:115-134, 2004 26. Fuhrer R, Stansfeld SA: How gender affects patterns of social relations and their impact on health: A comparison of one or multiple sources of support from ‘‘close persons’’. Soc Sci Med 54:811-825, 2002 27. Gauthier LR, Rodin G, Zimmermann C, Warr D, Moore M, Shepherd F, Gagliese L: Acceptance of pain: A study in patients with advanced cancer. Pain 143: 147-154, 2009 28. Ger LP, Ho ST, Sun WZ, Wang MS, Cleeland CS: Validation of the Brief Pain Inventory in a Taiwanese population. J Pain Symptom Manage 18:316-322, 1999 29. Giardino ND, Jensen MP, Turner JA, Ehde DM, Cardenas DD: Social environment moderates the association between catastrophizing and pain among persons with a spinal cord injury.[see comment]. Pain 106:19-25, 2003 30. Giovannetti ER, Wolff JL: Cross-survey differences in national estimates of numbers of caregivers of disabled older adults. Milbank Q 88:310-349, 2010 31. Hazan C, Shaver P: Romantic love conceptualized as an attachment process. J Pers Soc Psychol 52:511-524, 1987

15. Cano A, Leong L, Heller JB, Lutz JR: Perceived entitlement to pain-related support and pain catastrophizing: Associations with perceived and observed support. Pain 147: 249-254, 2009

32. Holmbeck GN: Toward terminological, conceptual, and statistical clarity in the study of mediators and moderators: Examples from the child-clinical and pediatric psychology literatures. J Consult Clin Psychol 65:599-610, 1997

16. Charlson ME, Pompei P, Ales KL, MacKenzie CR: A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation. J Chronic Dis 40: 373-383, 1987

33. Holmbeck GN: Post-hoc probing of significant moderational and mediational effects in studies of pediatric populations. J Pediatr Psychol 27:87-96, 2002

17. Ciechanowski P, Sullivan M, Jensen M, Romano J, Summers H: The relationship of attachment style to depression, catastrophizing and health care utilization in patients with chronic pain. Pain 104:627-637, 2003 18. Cleeland CS, Gonin R, Hatfield AK, Edmonson JH, Blum RH, Stewart JA, Pandya KJ: Pain and its treatment in outpatients with metastatic cancer. N Engl J Med 330: 592-596, 1994 19. Cleeland CS, Ryan KM: Pain assessment: Global use of the Brief Pain Inventory. Ann Acad Med Singapore 23: 129-138, 1994 20. Davies KA, Macfarlane GJ, McBeth J, Morriss R, Dickens C: Insecure attachment style is associated with chronic widespread pain. Pain 143:200-205, 2009 21. Extermann M: Measurement and impact of comorbidity in older cancer patients. Crit Rev Oncol Hematol 35:181-200, 2000 22. Fillingim RB, Doleys DM, Edwards RR, Lowery D: Spousal responses are differentially associated with clinical variables in women and men with chronic pain. Clin J Pain 19:217-224, 2003 23. Foley KM: The treatment of pain in the patient with cancer. CA Cancer J Clin 36:194-215, 1986

34. Holmes BM, Johnson KR: Adult attachment and romantic partner preference: A review. J Soc Pers Relat 26:833-852, 2009 35. Hunter MJ, Davis PJ, Tunstall JR: The influence of attachment and emotional support in end-stage cancer. Psychooncology 15:431-444, 2006 36. Jensen MP, Chang HY, Lai YH, Syrjala KL, Fann JR, Gralow JR: Pain in long-term breast cancer survivors: Frequency, severity, and impact. Pain Medicine 11:1099-1106, 2010 37. Karnofsky DA, Burchenal JH: The clinical evaluation of chemotherapeutic agents in cancer, in Macleod CM (ed): Evaluation of Chemotherapeutic Agents. New York, NY, Columbia University Press, 1949, pp 191-205 38. Katzman R, Brown T, Fuld P, Peck A, Schechter R, Schimmel H: Validation of a Short Orientation-MemoryConcentration Test of cognitive impairment. Am J Psychiatry 140:734-739, 1983 39. Keefe FJ, Abernethy AP, Campbell L: Psychological approaches to understanding and treating disease-related pain. Annu Rev Psychol 56:601-630, 2005 40. Keefe FJ, Lipkus I, Lefebvre JC, Hurwitz H, Clipp E, Smith J, Porter L: The social context of gastrointestinal cancer pain: A preliminary study examining the relation of

Gauthier et al patient pain catastrophizing to patient perceptions of social support and caregiver stress and negative responses. Pain 103:151-156, 2003 41. Keefe FJ, Porter LS, Labban J: Emotion regulation processes in disease-related pain: A couples-based perspective, in Snyder DK, Simpson J, Hughes JN (eds): Emotion Regulation in Couples and Families: Pathways to Dysfunction and Health. Washington, DC, American Psychological Association, 2006, pp 207-229 42. Kerns RD, Haythornthwaite J, Southwick S, Giller EL Jr: The role of marital interaction in chronic pain and depressive symptom severity. J Psychosom Res 34:401-408, 1990 43. Kerns RD, Turk DC, Rudy TE: The West Haven-Yale Multidimensional Pain Inventory (WHYMPI). Pain 23:345-356, 1985 44. Kurtz ME, Given B, Kurtz JC, Given CW: The interaction of age, symptoms, and survival status on physical and mental health of patients with cancer and their families. Cancer 74(Suppl 7):2071-2078, 1994 45. Lackner JM, Gurtman MB: Pain catastrophizing and interpersonal problems: A circumplex analysis of the communal coping model. Pain 110:597-604, 2004 46. Leonard MT, Cano A: Pain affects spouses too: Personal experience with pain and catastrophizing as correlates of spouse distress. Pain 126:139-146, 2006 47. McWilliams AA, Cox BJ, Enns MW: Impact of adult attachment styles on pain and disability associated with arthritis in a nationally representative sample. Clin J Pain 16:360, 2000 48. McWilliams LA, Asmundson GJ: The relationship of adult attachment dimensions to pain-related fear, hypervigilance, and catastrophizing. Pain 127:27-34, 2007 49. McWilliams LA, Holmberg D: Adult attachment and pain catastrophizing for self and significant other. Pain 149: 278-283, 2010 50. Meredith P, Ownsworth T, Strong J: A review of the evidence linking adult attachment theory and chronic pain: Presenting a conceptual model. Clin Psychol Rev 28: 407-429, 2008 51. Meredith PJ, Strong J, Feeney JA: Evidence of a relationship between adult attachment variables and appraisals of chronic pain. Pain Res Manag 10:191-200, 2005 52. Meredith PJ, Strong J, Feeney JA: The relationship of adult attachment to emotion, catastrophizing, control, threshold and tolerance, in experimentally-induced pain. Pain 120:44-52, 2006 53. Meredith PJ, Strong J, Feeney JA: Adult attachment variables predict depression before and after treatment for chronic pain. Eur J Pain 11:164-170, 2007 54. Mikulincer M, Florian V: The relationship between adult attachment styles and emotional and cognitive reactions to stressful events, in Simpson JA, Rholes WS (eds): Attachment Theory and Close Relationships. New York, NY, The Guilford Press, 1998, pp 142-165 55. Mikulincer M, Shaver PR: The attachment behavioral system in adulthood: Activation, psychodynamics, and interpersonal processes, in Zanna MP (ed): Advances in Experimental Social Psychology. Waltham, MA, Academic Press, 2003, pp 53-152

The Journal of Pain

1267

56. Mor V, Laliberte L, Morris JN, Wiemann M: The Karnofsky Performance Status Scale. An examination of its reliability and validity in a research setting. Cancer 53:2002-2007, 1984 57. Novy D, Berry MP, Palmer JL, Mensing C, Willey J, Bruera E: Somatic symptoms in patients with chronic noncancer-related and cancer-related pain. J Pain Symptom Manage 29:603-612, 2005 58. Pence L, Cano A, Thorn B, Ward LC: Perceived spouse responses to pain: The level of agreement in couple dyads and the role of catastrophizing, marital satisfaction, and depression. J Behav Med 29:511-522, 2006 59. Pence LB, Thorn BE, Jensen MP, Romano JM: Examination of perceived spouse responses to patient well and pain behavior in patients with headache. Clin J Pain 24: 654-661, 2008 60. Porter LS, Davis D, Keefe FJ: Attachment and pain: Recent findings and future directions. Pain 128:195-198, 2007 61. Radbruch L, Loick G, Kiencke P, Lindena G, Sabatowski R, Grond S, Lehmann KA, Cleeland CS: Validation of the German version of the Brief Pain Inventory. J Pain Symptom Manage 18:180-187, 1999 62. Rodin G, Walsh A, Zimmermann C, Gagliese L, Jones J, Shepherd FA, Moore M, Braun M, Donner A, Mikulincer M: The contribution of attachment security and social support to depressive symptoms in patients with metastatic cancer. Psychooncology 16:1080-1091, 2007 63. Romano JM, Turner JA, Jensen MP, Friedman LS: Chronic pain patient-spouse behavioral interactions predict patient disability. Pain 63:353-360, 1995 64. Selcuk E, Zayas V, Hazan C: Beyond satisfaction: The role of attachment in marital functioning. J Fam Theory Rev 2: 258-279, 2010 65. Shaver PR, Schachner DA, Mikulincer M: Attachment style, excessive reassurance seeking, relationship processes, and depression. Pers Soc Psychol Bull 31:343-359, 2005 66. Simpson JA, Winterheld HA, Rholes WS, Orina MM: Working models of attachment and reactions to different forms of caregiving from romantic partners. J Pers Soc Psychol 93:466-477, 2007 67. Sullivan MJL, Bishop SR, Pivik J: The Pain Catastrophizing Scale: Development and validation. Psychol Assess 7: 524-532, 1995 68. Sullivan M, Thorn B, Haythornthwaite J, Keefe F, Martin M, Bradley L, Lefebvre J: Theoretical perspectives on the relation between catastrophizing and pain. Clin J Pain 17:52-64, 2001 69. Sullivan MJ, Martel MO, Tripp D, Savard A, Crombez G: The relation between catastrophizing and the communication of pain experience. Pain 122:282-288, 2006 70. Thorn BE, Ward LC, Sullivan MJ, Boothby JL: Communal coping model of catastrophizing: Conceptual model building. Pain 106:1-2, 2003 71. Turk DC, Sist TC, Okifuji A, Miner MF, Florio G, Harrison P, Massey J, Lema ML, Zevon MA: Adaptation to metastatic cancer pain, regional/local cancer pain and non-cancer pain: Role of psychological and behavioral factors. Pain 74: 247-256, 1998

1268

The Journal of Pain

72. Turk DC, Rudy TE: Toward an empirically derived taxonomy of chronic pain patients: Integration of psychological assessment data. J Consult Clin Psychol 56:233-238, 1988 73. Wade DT, Vergis E: The Short Orientation-MemoryConcentration Test: A study of its reliability and validity. Clin Rehabil 13:164-170, 1999 74. World Health Organization. Cancer pain relief, Geneva, Switzerland, World Health Organization, 1986

The Communal Coping Model and Cancer Pain 75. Yates JW, Chalmer B, McKegney FP: Evaluation of patients with advanced cancer using the Karnofsky Performance Status. Cancer 45:2220-2224, 1980 76. Zaza C, Baine N: Cancer pain and psychosocial factors: A critical review of the literature. J Pain Symptom Manage 24: 526-542, 2002 77. Zimmermann C, Burman D, Bandukwala S, Seccareccia D, Kaya E, Bryson J, Rodin G, Lo C: Nurse and physician inter-rater agreement of three performance status measures in palliative care outpatients. Support Care Cancer 18:609-616, 2010