Partner-assisted emotional disclosure for patients with gastrointestinal cancer

Results from a randomized controlled trial

Authors


  • Presented at the Fourth Biennial Cancer Survivorship Research Conference entitled “Cancer Survivorship Research: Mapping the New Challenges,” Atlanta, Georgia, June 18-20, 2008.

Abstract

BACKGROUND:

For patients with cancer who are married or in an intimate relationship, their relationships with their partners play a critical role in their adaptation to illness. However, cancer patients and their partners often have difficulty in talking with each other about their cancer-related concerns. Difficulties in communication ultimately may compromise both the patient-partner relationship and the patient's psychological adjustment. The current study tested the efficacy of a novel partner-assisted emotional disclosure intervention in a sample of patients with gastrointestinal (GI) cancer.

METHODS:

One hundred thirty patients with GI cancer and their partners were assigned randomly to receive 4 sessions of either partner-assisted emotional disclosure or a couples cancer education/support intervention. Patients and partners completed measures of relationship quality, intimacy with their partner, and psychological distress before randomization and at the end of the intervention sessions. Data were analyzed using multilevel modeling.

RESULTS:

Compared with an education/support condition, the partner-assisted emotional disclosure condition led to improvements in relationship quality and intimacy for couples in which the patient initially reported higher levels of holding back from discussing cancer-related concerns.

CONCLUSIONS:

Partner-assisted emotional disclosure is a novel intervention that builds on both the private emotional disclosure and the cognitive-behavioral marital literature. The results of this study suggested that this intervention may be beneficial for couples in which the patient tends to hold back from discussing concerns. The authors concluded that future research on methods of enhancing the effects of partner-assisted emotional disclosure is warranted. Cancer 2009;115(18 suppl):4326–38. © 2009 American Cancer Society.

For patients with cancer who are married or in an intimate relationship, the relationship with their partner plays a critical role in adapting to their illness.1-3 Both patients and partners often perceive that social support and family functioning diminish in the year after a cancer diagnosis,3-5 suggesting that some couples may be at risk of relationship distress as a result of the burdens imposed by the cancer experience. One factor that may led to deterioration in the patient-partner relationship is the challenge of communicating effectively about cancer-related concerns. Married cancer patients tend to rate their spouses as their most important confidant6, 7; however, cancer patients often feel constrained in talking about their concerns with their partners, and partners often withdraw or distance themselves from the patient's emotional distress.2, 6, 8-10 These avoidant patterns are present even in relationships that are satisfying.11

Patients' inability to talk openly with their partners about their cancer-related concerns ultimately may compromise the quality of the patient-partner relationship as well as the patient's psychological adjustment. Emotional disclosure—defined as the expression of information that is personal, private, and emotional in nature12—is a central component of the emotional support (eg, love, concern, and understanding) that partners provide to each other.13, 14 The construct of emotional disclosure has been operationalized as both 1) the degree to which individuals express their thoughts and feelings and 2) the degree to which they hold back from doing so. Although expression and holding back may be perceived as opposites, in fact, their negative correlation is only moderate (eg, −.3215 to −.3816), suggesting that they may represent orthogonal constructs. Holding back may reflect active inhibition of expression rather than a general tendency toward inexpression. Prior studies suggest that high levels of holding back, more than low levels of expression, are associated significantly with poorer relationship functioning and increased psychological distress.16, 17 To date, however, all of the studies in this area have been correlational; thus, it is not possible to draw causal conclusions regarding the effects of disclosure or holding back.

Several studies, including several conducted with cancer patients,17-20 have demonstrated that expressive writing protocols can produce improvements in psychological and physical well being (eg, improved affect, decreased psychological distress, and self-reported symptoms).21, 22 In these protocols, participants typically write for 15 to 20 minutes over several sessions about their deepest thoughts and feelings regarding stressful experiences. In this paradigm, disclosure is private and anonymous; participants are alone when they disclose, and their disclosures are not shared with anyone other than the researchers. When applied to patients with cancer, this intervention led to improved sleep and vigor in patients with metastatic renal cell cancer17 and to significant reductions in physical symptoms and medical appointments among patients with breast cancer.18

In a separate tradition, increasing emotional disclosure between spouses is often a focus of cognitive-behavioral couple therapy.23, 24 A large body of empirical literature demonstrates that, in a variety of settings, cognitive-behavioral interventions for couples can significantly improve both individual adjustment and relationship functioning.25 Cognitive-behavioral couple interventions focus specifically on learning relationship skills, such as communication. Increases in emotional disclosure that occur during the course of cognitive-behavioral couples therapy have been linked to enhanced intimacy and marital satisfaction.26 However, to our knowledge, the potential benefits of facilitating disclosure regarding cancer-related issues between cancer patients and their partners have not been addressed.

There are several reasons why increasing emotional disclosure between cancer patients and their partners is likely to lead to benefits for couples' relationships. First, as noted above, increasing emotional disclosure may lead to increases in intimacy and relationship satisfaction.27 Second, helping couples to discuss their cancer-related concerns may help partners have a better understanding of patients' needs and to provide more effective support. In addition, it is possible that patients may derive the same individual psychological benefits from disclosure to their partner as they do from private emotional disclosure protocols. Finally, partners may benefit from hearing the patients talk openly about their concerns. Often, a partner senses when the patient has concerns and is not expressing them, and this can lead to increased distress for the partner; when the patient shares his/her concerns, the partner may experience this as a relief.

The purpose of this study was to evaluate the efficacy of a new partner-assisted emotional disclosure protocol for patients with gastrointestinal (GI) cancer. GI cancers make a particularly good model in which to study the effects of this protocol for several reasons. First, they are common: Colorectal cancers are the third leading site of cancer occurrence and cancer mortality for both men and women.28 Second, they often are discovered at advanced stages, when the prognosis for survival is poor.28, 29 Although it has been demonstrated that treatments improve survival, tumor control, and quality of life because they relieve tumor-related symptoms, these treatments often are associated with significant toxicities. Patients report multiple disease-related and treatment-related side effects, including fatigue, pain, difficulty eating, weight loss, problems with bowel function, and sexual problems.29-32 Many patients also experience significant emotional distress, including disturbances in body image, anxiety and worry, depression, and fears of disease progression and death.3, 29-32 Thus, the diagnosis and treatment of GI cancer poses numerous physical and psychological challenges for patients as well as their loved ones.

The partner-assisted emotional disclosure intervention was designed to systematically train patients and partners in strategies to facilitate the patients' disclosure and give the patient the opportunity to talk with their partner about their cancer-related concerns. We hypothesized that the intervention would lead to improvements in relationship quality, intimacy, and psychological distress for patients and their partners relative to an attention control condition. In addition, we conducted exploratory analyses examining whether intervention effects differed by sex or by the degree to which patients reported holding back from disclosing cancer-related concerns to their spouse at baseline.

MATERIALS AND METHODS

Participants

Eligible participants were 1) individuals diagnosed with stage II through IV GI cancer who had life expectancy ≥6 months and 2) their spouses or intimate partners (grading determined according to the TNM staging system). Patients were recruited from the GI oncology clinics at Duke University and University of North Carolina Hospitals. This study was approved by the Institutional Review Boards at Duke University Medical Center and the University of North Carolina School of Medicine, and informed consent was obtained from all participants.

In total, 603 patients were screened for the study, deemed eligible, and approached regarding participation (for a diagram of the Consolidated Standards of Reporting Trials, see Fig. 1). Of these patients, 150 (25%) expressed willingness to participate, and their partners also agreed to take part. The most common reasons for refusal were that the patient lived at a distance from the medical center and was not willing to travel for the sessions (n = 112), the patient and/or partner were not interested (n = 95) or believed that the study would be too time-consuming (n = 24), and that the patient was too ill to participate (n = 18).

Figure 1.

This is the Consolidated Standards of Reporting Trials (CONSORT) flowchart. PAED indicates partner-assisted emotional disclosure; EDU, education; Tx, treatment.

Of the 150 patients who consented to participate, 20 withdrew before randomization. Reasons for withdrawal included death/declining health (n = 9), lack of time (n = 6), distance (n = 3), and lost to contact (n = 2). This resulted in 130 couples who were included in the current analyses.

Procedures

After completing informed consent, participants in the study were administered pretreatment measures and then were assigned randomly to 1 of the 2 conditions: partner-assisted emotional disclosure or education/support. The treatment conditions are described below. Randomization was stratified by patient sex and cancer stage (stages II and III vs stage IV). After completing the intervention, participants in both conditions completed post-treatment measures. All evaluations were completed by telephone interviews, which were conducted by a research assistant who was blind to the participant's treatment condition. Each participant (patient and partner) was reimbursed $40 ($20 for each of the 2 evaluations).

Measures

The following measures were completed by both patients and partners.

Relationship quality

Relationship quality was assessed using the Quality of Marriage Index (QMI).33 The QMI is a 6-item inventory that assesses relationship satisfaction using broadly worded, global items. Participants indicate whether they agree with each item on a 0-10 scale from 0 (very strongly disagree) to 10 (very strongly agree;). Sample items include “My relationship with my partner makes me happy” and “I really feel like part of a team with my partner.” The QMI has demonstrated good reliability and validity and correlates highly with longer, well validated measures of marital adjustment such as the Dyadic Adjustment Scale.33, 34 Cronbach alpha in the current study was .92 for patients and .89 for partners.

Intimacy

Intimacy with the partner was assessed using the Miller Social Intimacy Scale (MSIS).35 The MSIS is a 17-item self-report inventory that assesses the degree of intimacy, closeness, and trust that an individual feels in a relationship toward his or her partner. Sample items include “How close do you feel to your partner most of the time?” and “When you have leisure time, how often do you choose to spend it with your partner alone?” There is good evidence that the measure discriminates close friends from casual friends and happily married couples from distressed couples, and the MSIS can detect clinically meaningful change after psychological treatment.35, 36 Cronbach alpha in the current study was .92 for patients and .88 for partners.

Psychological distress

Psychological distress was assessed using the Profile of Mood States-Short Form (POMS-SF).37 This measure asks participants to describe their mood over the past week by rating each of 30 adjectives on a scale from 0 (very much unlike this) to 3 (very much like this). There are 6 subscales (anxiety, depression, anger, vigor, fatigue, and confusion) and a Total Mood Disturbance (TMD) scale. The POMS has good temporal stability, excellent internal consistency, and good convergent validity with other measures of psychological distress. The POMS is used often in cancer research studies and has been deemed sensitive to improvements after psychological interventions in cancer patients.38, 39 In the current study, Cronbach alphas for the subscales ranged from .74 to .93 for both patients and partners. Cronbach alpha for the TMD was .88 for patients and .90 for partners.

Holding back

Holding back from disclosure of cancer-related thoughts and feelings was assessed using a modified version of a measure developed by Pistrang and Barker.15 The measure consists of 10 items that assess the extent to which patients talk about their cancer-related feelings and concerns to their spouse and how much they hold back from doing so on a scale from 0 (not at all) to 5 (a lot). Only the holding back scale was used in the current study. In previous studies of patients with GI cancer16 and breast cancer,7, 15 high levels of holding back were associated significantly with increased psychological distress and poorer relationship functioning. Cronbach alpha in the current study was .88.

Partner-Assisted Emotional Disclosure Intervention

The partner-assisted emotional disclosure (emotional disclosure) protocol was designed to train couples systematically in skills intended to help patients disclose their feelings and concerns related to the cancer experience. Couples attended 4 face-to-face sessions with a master's level therapist (social worker or psychologist). The first session lasted approximately 75 minutes and Sessions 2 through 4 lasted 45 minutes each. Although the intention was to schedule sessions weekly, couples were given up to 8 weeks to complete the 4 sessions to accommodate delays because of the patient's medical condition and/or to coordinate sessions with other appointments at the medical center. In the first session, the therapist provided a rationale for the intervention and trained the couple in skills designed to help the patient disclose his/her concerns related to the cancer experience. Patients' guidelines for disclosure included 1) thinking about an experience related to having cancer that caused strong emotions; 2) telling their partner about the experience in as much detail as possible, including both the events and feelings related to the experience; and 3) pausing periodically to give their partner the opportunity to respond (eg, talking in paragraphs). The partners' role was that of supporter. Their guidelines included 1) trying to put themselves in the patient's place and understand what the experience was like for him/her; 2) avoiding problem-solving, reassurance, or advice giving; and 3) reflective listening (eg, summarizing what the patient said). Training included both didactic and experiential components and was summarized in handouts that were given to the couples. The therapist helped the couple identify partner responses that the patient found helpful and unhelpful and assisted patients in generating a list of cancer-related topics to focus on in their disclosure sessions. Partners were instructed to keep the focus of the conversation primarily on the patient's experience and to disclose their own thoughts and feelings only as necessary to facilitate the patient's disclosure. (This continued primary focus on the patient's thoughts and feelings distinguishes the intervention from other couple-based interventions, such as cognitive-behavioral couple therapy, which places a bilateral focus on both individuals' thoughts and feelings shared reciprocally during a conversation24).

In the subsequent 3 sessions, the therapist initially presented a brief review of the strategies taught in the first session. Then, the patient was provided with an opportunity to talk about his/her feelings and concerns about the cancer experience. The couple was instructed that their goal was for the patient to focus on the expression of his/her feelings and concerns related to the cancer experience, and the goal for the partner was to use the skills learned in the initial session to facilitate disclosure as well as to communicate acceptance and understanding. During the disclosure, the therapist intervened only as necessary to 1) discourage negative interactions, 2) stimulate disclosure using the concern list generated by the patient in the first session, and 3) ensure that the couple's discussion focused primarily on the patient's experience rather than the partner's. Couples were encouraged to continue these conversations on their own outside of sessions; however, they were not given specific home practice assignments.

Cancer Education/Support Couple Condition

To hold constant the focus on the couple and cancer during intervention, couples in the cancer education/support condition also attended 4 face-to-face sessions. The therapists and the scheduling of the sessions were the same as those for the partner-assisted emotional disclosure intervention. The education/support sessions centered on presenting information relevant to living with cancer and were supplemented with handouts. The sessions focused on the following topics: orientation to the cancer center and the treatment team; suggestions for communicating with healthcare providers; resources for health information, psychosocial support, and financial concerns; evaluating health information on the internet; the impact of cancer on different domains of quality of life; and suggestions for maintaining quality of life. Patients and partners in the education/support condition did not receive any training in communication skills, and patients were not encouraged to discuss their thoughts or feelings related to the cancer experience with their partner.

Uniformity of Treatment

Several steps were taken to ensure that the treatment protocols were uniform and that the therapists followed the treatment protocols in a uniform manner. These included therapist training, use of detailed treatment outlines, audio taping of sessions and therapist supervision, and assessments of treatment adherence and therapist competence.

Data Analyses

Multilevel modeling for dyadic data40, 41 was used to test group differences over time and to determine whether these differences varied for patients versus partners. Multilevel modeling has several advantages over traditional analytic techniques, such as repeated-measures analysis of variance, in that it allows for unbalanced data in terms of the number and timing of data points, accommodates both time-varying and time-invariant covariates, and is able to account for the influence of data from 1 member of a couple on data from the other member of that couple.41

Age, time since diagnosis, and cancer stage were centered and then entered first into the model as covariates. Main effects included in the model were time (before vs after the intervention), treatment condition (partner-assisted emotional disclosure vs education/support), and social role (patient vs partner). Interaction effects that were included in the model were time × treatment and time × treatment × role. A significant intervention effect is indicated by a significant 2-way interaction between treatment (partner-assisted emotional disclosure vs education/support) and time. A significant 3-way interaction (time × treatment × role) indicates that the treatment effects were different for patients versus partners.

To explore the potential moderating effects of sex and patient holding back assessed at baseline, additional models were run that included the main effects and interaction effects of 1) sex and 2) patient holding back. A significant moderation effect was indicated by a significant 3-way interaction between the moderator (sex or holding back), treatment, and time or by a significant 4-way interaction (moderator × treatment × time × role).

Data were analyzed first by intent to treat (eg, including all randomized couples; n = 130). Of these, 108 couples completed post-treatment assessments (Fig. 1). Reasons for dropout included death/declining health (n = 10), not enough time/conflicting demands (n = 8), and loss to contact/not returning to medical center (n = 4). A second set of analyses was conducted that included only those couples who attended at least 1 treatment session (partner-assisted emotional disclosure, n = 53; education/support, n = 59). The number of sessions completed by couples in each condition is shown in Figure 1. The reasons for not completing sessions were similar to those for dropping out.

The distribution of relationship quality was negatively skewed. Thus, this variable was log transformed, and analyses were performed with the log-transformed values.

Participant Characteristics

Participant characteristics are presented in Table 1. Seventy-one percent of the patients were men. Participants were predominantly Caucasian and well educated. The majority of patients (64.6%) had Stage IV disease. On average, couples had been married for 23.4 years.

Table 1. Participant Characteristics
CharacteristicPatients, N=130Partners, N=130
  1. SD indicates standard deviation; IQR, interquartile range.

Mean±SD age, y59.4±1259.3±12.3
Sex, % men7129
Race, %  
 Caucasian84.582
 African American11.611.2
 Other/unknown3.96.4
Education, %  
 <12 y14.28.2
 High school graduate30.932.3
 Some college29.630.2
 College graduate/postgraduate25.329.3
Median time since diagnosis (IQR), d207.5 (668) 
Cancer site, %  
 Colorectal42 
 Pancreatic15 
 Esophageal11 
 Other32 
TNM Cancer stage, %  
 II12.3 (n=16) 
 III23.1 (n=30) 
 IV64.6 (n=84) 
Cancer treatments at baseline, %
 Chemotherapy60.8 (n=79) 
 Radiation8.5 (n=11) 

RESULTS

Intent to Treat

Results from the of mixed model dyadic analyses that included all participants indicated that there was a significant time × treatment effect for relationship quality (β = −0.07; standard error [SE], 0.03; P = .02). The time × treatment effect was not significant for intimacy or mood disturbance. The 3-way interactions between time, treatment, and social role (patient vs partner) also were not significant, indicating that the treatment effect did not differ for patients versus partners for any of the outcome variables. There were no significant main effects of time, indicating that these variables did not change over time in general. There also were no significant main effects of social role, indicating that levels of relationship quality, intimacy, and mood disturbance did not differ significantly between patients and partners.

The significant time × treatment interaction for relationship quality was graphed according to the strategies recommended by Preacher et al (Fig. 2).42 Patients and spouses in the partner-assisted emotional disclosure condition reported increases in relationship quality from before treatment to after treatment relative to patients and spouses in the education/support condition.

Figure 2.

This chart illustrates the estimated trajectories of change in patients' and partners' scores (logged values) on the Quality of Marriage Index (QMI) by treatment condition (intent-to-treat analysis). ED indicates emotional disclosure; Educ, education.

Models that included sex as a potential moderator indicated that there were no significant interactions between sex, treatment condition, and time. However, analyses that included patient holding back indicated significant 3-way interactions (holding back × treatment condition × time) for relationship quality (B = 0.10; SE, 0.03 [P < .0001]) and intimacy (B = 0.56; SE, 0.28 [P = .02]). Figures 3 and 4 illustrate the finding that positive effects of the partner-assisted emotional disclosure intervention on relationship quality and intimacy occurred only when patients reported high levels of holding back from talking about cancer-related concerns to their spouse at baseline.

Figure 3.

These charts illustrate the estimated trajectories of change in patients' and partners' scores (logged values) on the Quality of Marriage Index (QMI) by treatment condition and patient baseline holding back (intent-to-treat analysis). ED indicates emotional disclosure; Educ, education.

Figure 4.

These charts illustrate the estimated trajectories of change in patients' and partners' scores on the Miller Social Intimacy Scale by treatment condition and patient baseline holding back (intent-to-treat analysis). ED indicates emotional disclosure.

Treatment Completers

The analyses described above were repeated; and, this time, only those couples who had completed at least 1 session of the intervention were included (N = 53 in partner-assisted emotional disclosure; N = 59 in education/support). The time × treatment interaction remained significant for relationship quality (B = −0.08; SE, 0.04 [P = .02]). In addition, there was a significant 3-way interaction between time, treatment, and social role (patient vs partner) for intimacy (B = −0.60; SE, 0.30 [P = .05]). There were no main or interaction effects for mood disturbance.

The significant time × treatment × role interaction for intimacy was graphed according to the strategies recommended by Preacher et al.42 Figure 5 illustrates the positive effect of partner-assisted emotional disclosure on intimacy, which was observed among patients but not partners. However, in models that contained patient baseline holding back, the 3-way interaction between holding back, treatment condition, and time continued to be significant for both relationship quality and intimacy; the 4-way interaction (holding back × treatment condition × time × role) was not significant. This suggests that both patients and partners experienced improvements in relationship quality and intimacy after the partner-assisted emotional disclosure intervention if the patient initially reported high levels of holding back. Models that included sex as a potential moderator indicated that there were no significant interactions between sex, treatment condition, and time.

Figure 5.

These charts illustrate the estimated trajectories of change in scores on the Miller Social Intimacy Scale by social role (patient vs partner) and treatment condition (treatment completer analysis). ED indicates emotional disclosure; Educ, education.

DISCUSSION

Findings from this study suggest that a brief, couple-based intervention that specifically targeted communication of cancer-related thoughts and feelings in patients with GI cancer led to improvements in relationship quality and intimacy for couples in which the patient initially reported higher levels of holding back from discussing cancer-related concerns. This finding is noteworthy in light of the brevity of the intervention, the length of marriage among couples in the study (mean, 24 years), and the severity of the patient's illness (66% of the patients were diagnosed with Stage IV cancer). Improving the quality of the couple's relationship is particularly important in this population because of their shortened life expectancy and the vital role that spouses play in patients' care at the end of life. This study had several methodological strengths, including comparing the partner-assisted emotional disclosure intervention with a couples cancer education/support condition that controlled for time and attention given to both patients and partners and the use of sophisticated data-analytic techniques (eg, multilevel models) that are uniquely suited to analyzing data from couple-based interventions.40, 41

“Holding back” likely reflects patients' active inhibition of disclosing their cancer-related concerns with their partners rather than a general tendency toward inexpression. Thus, patients who report high levels of holding back may want to discuss these concerns but may be reluctant to do so. It is noteworthy that, on average, patient's levels of holding back, as assessed at baseline, were quite low (mean, 0.82; standard error, 0.99; range, 0-4). Thus, even patients who were considered “high” in holding back did not indicate that they held back from discussing their cancer-related concerns with their spouse to a large degree. These levels of holding back are similar to what we observed in a previous study with the same patient population.16 However, even this relatively modest degree of holding back was associated with significantly poorer individual and relationship functioning.16 Results from the current study suggest that helping patients overcome their reluctance to express their concerns may lead to significant improvements in the quality of their marital relationships and in the amount of intimacy they share with their spouse. This is consistent with research indicating that disclosure of personal, vulnerable information is a central component of intimacy13, 14 that, in turn, is linked closely to relationship satisfaction.27

Contrary to our hypotheses and to the literature on private emotional disclosure, the intervention did not lead to improvements in psychological distress among patients or partners. One possible explanation for this lack of findings is that, although the intervention was designed to facilitate patient disclosure, patients may have been unable to disclose their thoughts and feelings to their partner (and in the presence of a therapist) to the extent that they would have in a private disclosure protocol. Some patients may have felt inhibited about sharing their thoughts and feelings with their partner and/or the therapist. Alternatively, the sessions' dual focus on both the patient's disclosure and on the dyad's communication processes may have distracted the patient from expressing and processing their thoughts and feelings to the same extent that they would have in a private disclosure protocol. Another possibility is that patients may have believed that their disclosure was burdensome to their partners and that this sense of burden offset any psychological benefits of disclosure. It is noteworthy, however, that there was no evidence that the partner-assisted emotional disclosure sessions led to increased psychological distress for patients or partners. Finally, it is possible that the effects of the intervention were dampened because many couples did not receive all 4 sessions. Future studies of partner-assisted emotional disclosure should consider strategies to increase retention in the intervention, such as offering home-based sessions or targeting patients with less advanced disease. It also may be necessary to increase the number of sessions so that patients become more comfortable with the setting and the communication guidelines and have more opportunity to disclose. In addition, future studies should explore the possibility that the 2 formats of disclosure—partner-assisted versus private—have different effects. Partner-assisted protocols in which patients disclose their concerns to their partner may be more effective in enhancing couples' relationships, whereas private disclosure protocols in which patients write about their thoughts and feelings may be more effective in reducing patients' psychological distress. If so, then treatment can be targeted according to the specific needs of the patient and the couple.

Because of the focus of our intervention on patient disclosure, it may be less surprising that partners did not benefit in terms of reduced psychological distress. Although many partners expressed appreciation for the opportunity to hear the patient discuss his/her concerns, it also was difficult at times for them to listen to the patient describe painful experiences and emotions, particularly because partners were asked to refrain from reassuring the patient or giving advice or suggestions. Thus, any benefits for partners may have been offset by increased burden from listening to the patient's concerns. However, as noted above, partners did not experience increases in distress as a result of the partner-assisted emotional disclosure intervention. Given the high burden of cancer caregiving,43, 44 especially when the patient has advanced disease,45, 46 particular attention should be given to developing interventions that may help alleviate partner distress. Enhancing cancer-related communication between partners ultimately may benefit both patients and partners if it leads to increased relationship functioning and a shared understanding about the meaning of the illness experience. One interesting direction for future research would be to explore the relative benefits for patients and partners of including a more reciprocal approach to disclosure (eg, partners disclosing to patients as well as patients to partners). In addition, longer term effects of the partner-assisted emotional disclosure intervention should be examined, because it is possible that short-term improvements in relationship satisfaction potentially may lead to improvements in individual psychological well being for patients and/or partners.

The current findings build on the private emotional disclosure literature, which has indicated that writing about cancer concerns can lead to benefits for patients with cancer.17-19 To our knowledge, this is the first study of cancer patients that has applied a disclosure paradigm in the context of a patient disclosing to their spouse or intimate partner. There are several reasons that partner-assisted disclosure is particularly appropriate for cancer patients and their partners. The diagnosis and treatment of most cancers places major demands not only on patients but on their partners as well. Although patients and partners must work together to meet these demands, communication, particularly about emotionally related topics, can be challenging even in relationships that are strong overall. The results of the current study are promising, because they demonstrate that a brief intervention targeting patient disclosure to their partner can be beneficial for couples' relationships. Future studies are needed to replicate these findings in patients and partners who are coping with other types of cancer (eg, breast cancer, prostate cancer, or lung cancer).

From a clinical perspective, the findings of this study suggest that psychologists, social workers, and other mental health professionals working with GI cancer patients may want to identify patients who tend to hold back from expressing their cancer-related thoughts and feelings to their partners and may want to work with them and their partners to help facilitate communication. In the current study, the only demographic or medical variable that was associated with holding back was age, indicating that younger patients were more likely to hold back. This is consistent with clinical impressions that younger patients often have more difficulty coping with cancer. Overall, although most couples reported having strong relationships at baseline, many commented that the treatment sessions were helpful in enhancing their relationship further, because they gave them the opportunity to discuss important issues that they previously had avoided. Similar to what has been reported in previous studies,6-8 patients often noted that their greatest need was for someone to listen to them and that it was helpful to be able to express their concerns more openly to their partner. Partners, conversely, often observed that it was difficult not to respond to the patient's concerns with reassurance or problem-solving; however, over the course of the sessions, they learned to appreciate that validating the patient's concerns was more helpful to the patient and also resulted in a higher degree of patient disclosure.

Several limitations need to be kept in mind when interpreting the results of this study. First, the participation rate in this study (25%) was somewhat lower than that obtained in other recent couple-based or caregiver-assisted studies with cancer patients (range, from 34%47, 48 to 43%49), possibly because many of the patients who were approached regarding participation were quite ill. Patient illness also was a significant factor in attrition. Recruitment and retention also most likely were hampered because participation required couples to attend 4 face-to-face sessions. Future studies could test the efficacy of telephone-based or internet-based interventions that could be made available to a wider population of couples. A second limitation is that the couples who participated in this study, on average, reported high levels of relationship quality. The findings may not generalize to couples who are more distressed, who quite likely need a more lengthy and intensive intervention. Third, the results of the study likely reflect the sample of participants who tended to be well educated and have advanced disease. The sex composition of the sample (more men than women patients) also may have influenced the pattern of results. Women reportedly self-disclose more than men, although this difference diminishes in the context of heterosexual relationships14, 50 and was not evident from the baseline data in the current study. Sex differences in response to written emotional disclosure interventions have also been reported.22 Thus, it is possible that the results would not generalize to samples of women patients or to contexts other than heterosexual romantic relationships.

Taken together, the results of this study suggest that future research on partner-assisted emotional disclosure is warranted. This is a novel intervention that builds on both the private emotional disclosure and the cognitive-behavioral marital therapy literature. Although the benefits in the current study were reflected primarily in indices of the quality of the couple's relationship, it is possible that these short-term improvements in relationship quality and intimacy could be associated with longer term improvements in both patient and partner individual functioning. In addition to examining longer term effects, future research also could explore methods of enhancing the effects of the intervention, such as using a mixture of laboratory-based and home-based, partner-assisted disclosure sessions, increasing the number of sessions, and/or exploring the relative benefits of having partners disclose to patients as well as patients to partners. In addition, future research needs to focus on more distressed couples for whom more intensive approaches most likely are necessary. Given the important role that partners play in patients' adaptation to cancer, it is critical to help patients and partners communicate effectively about their cancer-related concerns.

Acknowledgements

We thank Susan Elinoff; Laura Harris, MSW; Regina Ragan, MSW; Rebecca Shelby, PhD; Catherine Moser, PhD; Johanna Bendell, MD; Leigh Howard, RN; and the physicians and staff of the Duke Gastrointestinal Oncology clinic for their assistance. We also extend our gratitude to all of the study participants for their time and effort.

Conflict of Interest Disclosures

Supported by Grant R01 CA100743 from the National Cancer Institute.

Cosponsored by the National Cancer Institute”s Office of Cancer Survivorship, the Office of Cancer Survivorship of the Centers for Disease Control and Prevention, and the American Cancer Society”s Behavioral Research Center.

Ancillary