|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
© 2002 American Society for Clinical Oncology Randomized, Controlled Trial of Written Emotional Expression and Benefit Finding in Breast Cancer PatientsByFrom the Department of Psychology, University of Kansas, Lawrence, KS. Address reprint requests to Annette L. Stanton, PhD, Department of Psychology, 426 Fraser Hall, University of Kansas, Lawrence, KS 66045-2160; email: astanton{at}ukans.edu
PURPOSE: Expressing emotions and finding benefits regarding stressful experiences have been associated in correlational research with positive adjustment. A randomized trial was performed to compare effects of experimentally induced written emotional disclosure and benefit finding with a control condition on physical and psychological adjustment to breast cancer and to test whether outcomes varied as a function of participants’ cancer-related avoidance. PATIENTS AND METHODS: Early-stage breast cancer patients completing medical treatment were assigned randomly to write over four sessions about (1) their deepest thoughts and feelings regarding breast cancer (EMO group; n = 21), (2) positive thoughts and feelings regarding their experience with breast cancer (POS group; n = 21), or (3) facts of their breast cancer experience (CTL group; n = 18). Psychological (eg, distress) and physical (perceived somatic symptoms and medical appointments for cancer-related morbidities) outcomes were assessed at 1- and 3-month follow-ups. RESULTS: A significant condition x cancer-related avoidance interaction emerged on psychological outcomes; EMO writing was relatively effective for women low in avoidance, and induced POS writing was more useful for women high in avoidance. Significant effects of experimental condition emerged on self-reported somatic symptoms (P = .0183) and medical appointments for cancer-related morbidities (P = .0069). Compared with CTL participants at 3 months, the EMO group reported significantly decreased physical symptoms, and EMO and POS participants had significantly fewer medical appointments for cancer-related morbidities. CONCLUSION: Experimentally induced emotional expression and benefit finding regarding early-stage breast cancer reduced medical visits for cancer-related morbidities. Effects on psychological outcomes varied as a function of cancer-related avoidance.
A NUMBER OF controlled experiments have demonstrated the benefits of expressive disclosure on research participants’ physical and psychological health. Developed by Pennebaker and Beall,1 the method typically involves random assignment of individuals to one of two conditions, with instructions to write or speak about their deepest thoughts and feelings regarding a stressful experience or about innocuous topics (control) over several brief experimental sessions. Dependent variables include such indices as subsequent medical appointments, immune parameters, and subjective well-being. A meta-analysis2 of 13 investigations of written emotional disclosure yielded a reliable improvement in health outcomes. However, all 13 studies included physically healthy participant samples. Recently, the effects of expressive disclosure have been examined in medical patient samples. In a recent trial,3 patients with asthma or rheumatoid arthritis were instructed to write over a three-session period either about the most stressful event of their lives or about emotionally neutral topics. As compared with control participants, asthma patients in the experimental group evidenced significantly improved lung function as assessed through spirometry at 4-month follow-up, and rheumatoid arthritis patients showed significant improvement in overall disease activity indicated in clinical examinations conducted by a rheumatologist. Other research4 also revealed benefits for rheumatoid arthritis patients of audiotaped emotional disclosure versus a control condition on affective disturbance and reported physical functioning in daily activities but no effect of the experimental manipulation on pain or joint condition. It seems that emotional disclosure can carry benefits for both physically healthy individuals and those diagnosed with chronic disease. However, only one known published study of experimentally manipulated emotional disclosure has been conducted with cancer patients.5 That study, using a relatively small sample, revealed no effect of a three-session home-based writing intervention on psychological adjustment. However, physical health outcomes were not assessed. Further, other studies with medical patient samples3,4 have induced expression not about the relevant disease, but rather regarding a stressful experience of the participants’ choice. Finally, no known study with a medical sample has compared the effects of emotional disclosure to other potentially useful experimental interventions. An intriguing comparison condition involves inducing participants to write about the benefits of their experience with cancer. Descriptive studies reveal that many individuals who confront serious disease are able to extract benefits from their experience, often including perceptions of strengthened intimate relationships, favorable personality change, and changes in life priorities.6 Given the demonstrated relation of naturally elected benefit finding with adjustment in cancer patients7,8 and the promotion of emotional expression or benefit finding in many popular clinical interventions for cancer patients,9-12 experimental tests of the utility of these strategies are particularly important. The first goal of this randomized, controlled trial was to test the hypothesis that written emotional disclosure and benefit finding would produce enhanced physical health-related outcomes (ie, medical appointments for cancer-related morbidities, self-reported physical symptoms) and psychological well-being (ie, health-related quality of life, positive and negative affect) relative to a control condition in which participants wrote solely about the facts of their breast cancer experience. A secondary goal was to test the relative effectiveness of the writing conditions as a function of women’s cancer-related avoidance (ie, intentional attempts to avoid thoughts and feelings regarding cancer). Some evidence suggests that expressive disclosure may be most effective for individuals who use more approach-oriented, expressive coping strategies than for those who are more nonexpressive or have deficits in identifying and processing emotion.13,14 Accordingly, an interaction between experimental condition and cancer-related avoidance was postulated such that participants low on avoidance would benefit more from emotional disclosure than would avoidant women. Whether the effects of experimentally induced benefit finding would vary as a function of avoidance also was examined.
Patients Participants were 60 women who met inclusion criteria of having a first diagnosis of stage I or II breast cancer and being within 20 weeks after completion of medical treatments (ie, surgery, radiotherapy, or chemotherapy). Exclusion criteria were diagnosis of recurrent or metastatic disease and inability to read or write English. As the outcomes of most interest in written expressive disclosure research,15 variables relevant to physical health were selected as the primary dependent variables. Appropriate sample size (n > 54) was determined16 on the basis of comparisons across treatment arms of these two measures (ie, medical appointments for cancer-related morbidities and self-reported physical symptoms) (power = 0.80; f2 = 0.35; P < .05). In addition to the 60 participants that composed the sample for analyses, 14 women were recruited for the study but did not complete it for reasons as follows (Table 1): two women declined on study recruitment, four consented to the study but did not complete the baseline questionnaire, five completed the baseline questionnaire but then terminated participation before randomization, two were diagnosed with cancer recurrence and terminated participation after completing the baseline questionnaire but before attending the writing sessions (thus, they were not aware of condition assignment), and one declined further participation after completing one control condition writing session. Thus, 81% of the women introduced to the research completed it, and only one who terminated participation was aware of her condition assignment. Those who declined or terminated participation most often did not report a reason or reported that they were too busy. Nonparticipants (n = 8) completing the baseline questionnaire did not differ significantly on any baseline variable from study participants.
Procedure After approval by institutional review boards at four participating oncology clinics (one academic medical center, one hospital-based clinic, and two group practices) in the midwestern United States, women were introduced to the study by research staff or medical personnel at a clinic visit during their medical treatment for stage I or II breast cancer. Women were informed that the study’s purpose was to "learn more about how women adjust to having breast cancer" and that they would be randomized to one of three conditions that would involve writing about their experience with breast cancer. On medical treatment completion, women were telephoned by a study coordinator to describe the trial further and schedule participation. The study coordinator did not have access to the randomization schedule, which was kept in a separate office. Participants were mailed an informed consent form and a baseline questionnaire packet, which they returned by mail or at the time of their first writing session. The informed consent form included no mention of expected benefits from the writing conditions. The randomization schedule was created by the first author using a computerized random numbers generator. Sequentially numbered envelopes were used to conceal allocation, and the study coordinator determined and revealed each patient’s condition assignment to trained research assistants just before the patient’s first writing session. Once randomization was performed, no change in treatment allocation took place. Trained research assistants conducted individual writing sessions in patients’ homes, the first author’s laboratory, or a participating medical institution. Because they verbally conveyed the writing instructions to participants, assistants were informed of condition assignment immediately before each participant’s first writing session. However, assistants followed a standard script, which made no mention of expected benefits of any condition. Assistants were not aware of research hypotheses. The conditions all involved a writing task, with instructions over each of four sessions to write about the participant’s (1) "deepest thoughts and feelings" regarding her experience with breast cancer (expressive condition; EMO group) following standard instructions from Pennebaker,1 (2) positive thoughts and feelings regarding her experience with breast cancer (benefit-finding condition; POS group), or (3) facts regarding her cancer and its treatment (fact-control condition; CTL group). All writing instructions are available from the first author. As an example, writing instructions for the EMO group were as follows: "What I would like you to write about for these four sessions are your deepest thoughts and feelings about your experience with breast cancer. I realize that women with breast cancer experience a full range of emotions, and I want you to focus on any and all of them. In your writing, I want you to really let go and explore your very deepest emotions and thoughts. You might think about all the various feelings and changes that you experienced before being diagnosed, after diagnosis, during treatment, and now. Whatever you choose to write, it is critical that you really focus on your deepest thoughts and feelings. Ideally, I would like you to focus on feelings, thoughts, or changes that you have not discussed in great detail with others. You might also tie your thoughts and feelings about your experiences with cancer to other parts of your life, ie, your childhood, people you love, who you are, or who you want to be. Again, the most important part of your writing is that you really focus on your deepest emotions and thoughts. The only rule we have is that you write continuously for the entire time. If you run out of things to say, just repeat what you have already written. Don’t worry about grammar, spelling, or sentence structure. Don’t worry about erasing or crossing things out. Just write." The sessions occurred within a 3-week period for each participant and were scheduled at the participant’s convenience. Patients were informed of their condition assignment immediately before beginning writing at the first session, and they engaged in the 20-minute writing task at each of four sessions. Physiological arousal (heart rate and skin conductance) and self-reported arousal and affect were monitored during each session, but these data are not reported here. At their final writing session, participants were given a form on which to record any medical visits over the subsequent 3 months, which they returned at the 3-month assessment. One month and 3 months after the final writing session, participants also completed measures of quality of life, positive affect, negative affect, and somatic symptoms. Participants received and returned all follow-up questionnaires by mail. Each participant was compensated $100 for her time.
Measures Administered at study entry and the follow-up assessments, the Functional Assessment of Cancer Therapy (FACT)20 scale also was included as a measure of health-related quality of life. This measure, which assesses perceived life quality in physical, social, relationship with doctors, emotional, and functional domains, possesses adequate psychometric properties.20 The items are rated on a 5-point scale (0 = not at all; 4 = very much). A total score on the 28 items was used to indicate health-related quality of life. Its correlations with the POMS indexes ranged in absolute magnitude from 0.56 to 0.66 (P < .0001). Although the FACT contains one subscale to assess the domain of physical health, it was included in the psychological variable cluster to preserve the integrity of the often used total score and because the majority of FACT items are relevant to psychosocial adjustment. Physical health-related outcomes. Negative somatic symptoms were assessed on a measure developed by Pennebaker.21 At baseline and follow-ups, patients reported the number of days in the past month on which they had experienced each of nine somatic symptoms (eg, coughing/sore throat), and a total score was calculated. Patients prospectively recorded all medical visits (including dental and eye exams) during the period from the last writing session through the subsequent 3 months. They also recorded the medical provider and reason for each visit (eg, check-up with medical oncologist). A subset of these reports was confirmed through medical records with the patients’ consent. Rate of agreement of patients’ reports and medical records was 92%, lending confidence to the accuracy of patients’ reports. These medical appointments were coded as a function of reason for the visit (ie, routine and nonroutine cancer-related and non–cancer-related appointments) by raters unaware of participants’ condition assignment. We were interested in medical appointments for cancer-related problems (eg, lymphedema, breast symptoms, or possible recurrence), excluding scheduled check-ups, as an indicator of morbidity associated with breast cancer and its treatment. The other categories of medical appointments (ie, cancer-related scheduled medical check-ups, other scheduled medical check-ups, or nonroutine medical appointments for other problems, such as flu symptoms) were combined for analysis. Self-reported somatic symptoms and medical appointments for cancer-related morbidities at 3 months yielded a correlation of 0.26 (P = .0469). Correlations of these two indicators with the psychological adjustment measures at 3 months ranged in absolute magnitude from 0.05 (P > .10) to 0.40 (P < .005), and all were in the direction of poorer health status being associated with poorer psychological adjustment. Cancer-related avoidance. Two psychometrically sound scales were used to indicate avoidance of cancer-related thoughts and feelings. First, as in our other studies,18 three subscales (ie, denial, mental disengagement, and behavioral disengagement) of the COPE,22 a psychometrically sound measure of coping processes, were combined to indicate avoidance-oriented coping. Women were instructed to respond to items regarding "what you have been doing to cope with your experience of breast cancer." Rated on 4-point scales, sample items are "I refuse to believe that it has happened" and "I sleep more than usual to think about it less." Second, the Impact of Events Scale23 eight-item avoidance subscale assesses intentional, avoidant responses specifically targeted to a stressful event, in this case "your experience with cancer" over the past 7 days (eg, "I tried not to think about it" or "I avoided letting myself get upset when I thought about it or was reminded of it"). These two indexes of cancer-related avoidance, correlated at 0.50 (P < .0001), were standardized and summed. Manipulation check and essay ratings. An independent rater, unaware of condition membership, judged whether each transcribed essay, ordered randomly, conformed to condition instructions. In addition, after each writing session, participants rated how personal their essays were, how much they revealed emotions in their essays, and how much writing increased understanding of their experience on 7-point scales (1 = not at all; 7 = a great deal/extremely). At the 1-month and 3-month follow-ups, participants rated how valuable the experiment was to them and how much they expected their participation to have long-lasting positive and negative effects, also on 7-point response scales. Preliminary repeated-measures analyses of variance indicated that neither scores for the four sessions nor scores from follow-up questionnaires differed significantly across time. Accordingly, means across sessions and across follow-ups were calculated for each variable.
Statistical Analysis
Sample Characteristics Participants had a mean age of 49.53 years (SD, 12.16 years; range, 21 to 76 years) and educational level of 15.20 years (SD, 2.48 years). Most (93%) were white (5% African American; 2% Asian American), 68% were employed outside the home, and 78% were married. Average diagnosis duration was 28.37 weeks (SD, 9.95 weeks; range, 12 to 52 weeks). Mastectomy was received by 30%, breast conservation by 62%, and both surgical procedures by 8%. Most (75%) had received chemotherapy, 67% had received radiotherapy, 17% had undergone reconstructive surgery, and 52% were taking a selective estrogen receptor modulator (eg, tamoxifen) at the time of the study. Forty-three percent had attended a support group or had consulted a mental health professional regarding breast cancer at least once at the point of study entry.
Manipulation Check and Essay Ratings
A MANOVA including experimental condition, cancer-related avoidance, and their interaction as independent variables was conducted on the participants’ ratings of the degree to which their essays were personal, revealing, and increased their understanding, each of which was averaged across writing sessions. The effect for experimental condition was significant (Wilks’
A MANOVA conducted on the two ratings of perceptions of value and long-lasting positive effects of the writing task averaged across 1-month and 3-month follow-ups was significant for the experimental conditions (F[4,104] = 3.18; P = .0165). Table 2 demonstrates that POS participants rated the experiment as significantly more valuable (95% CI, 0.75 to 2.86) and as having more long-lasting positive effects (95% CI, 0.66 to 2.75) than did CTL participants, with EMO participants’ ratings falling between the two groups. Listed in Table 2, an analysis of variance conducted on the rating of long-lasting negative effects was not significant and indicated that the groups did not expect enduring negative effects from their participation (mean < 1.7 on a 7-point scale).
Analyses on Psychological Outcomes
Although analyses revealed no significant main effects for condition or cancer-related avoidance, a significant multivariate effect emerged for the condition x avoidance interaction at 1 month (F[6,88]= 5.14; P = .0001) and at 3 months (F[6,92] = 2.92; P = .0118). In both cases, these were accompanied by a significant univariate interaction on POMS distress (F[2,48] = 7.37; P = .0016) at 1 month and 3 months (F[2,50] = 4.80; P = .0124). Presented in Fig 1 (top), the interaction plots for distress at 1 month revealed that avoidance was related significantly to distress in the expressive disclosure condition (F[1,17] = 14.28; P = .0015), such that induced expressive disclosure produced a decrease in distress for women who were low on cancer-related avoidance, whereas high avoidance was related to higher distress in that condition. A marginal effect in the opposite direction emerged for women in the benefit-finding condition (F[1,14] = 4.13; P = .0615), revealing that higher avoidance was associated with decreased distress and lower avoidance with higher distress. Cancer-related avoidance was unrelated to distress in the CTL condition (F[1,15] = 2.37; P = .1444). Results were similar but somewhat weaker at 3 months (Fig 1, bottom), such that cancer-related avoidance was associated significantly with distress only in the expressive disclosure condition (F[1,16] = 7.46; P = .0148). Although the univariate interactions were not significant, predicted condition means on the other two psychological adjustment variables (FACT and POMS vigor) as a function of avoidance were in the same direction as those for POMS distress.
Analyses on Physical Health-Related Outcomes Health-related outcomes are listed in Table 4. In these analyses, main effects of experimental conditions were significant at 3 months. The groups did not differ on self-reported somatic symptoms at 1 month, and no significant condition x avoidance interaction emerged on physical health-related outcomes. At 3 months, the multivariate condition effect on health-related outcomes was significant (F[4,100] = 3.18; P = .0166). Both self-reported physical symptoms and prospectively recorded medical appointments for cancer-related morbidities yielded a significant effect of experimental conditions at 3 months. The EMO group evidenced a significant decrease in physical symptoms compared with the CTL group, and the POS group participants’ symptom scores fell between the other groups. The 95% CIs for the comparison of the expressive disclosure condition with the control condition on self-reported somatic symptoms ranged from -3.5 to -22.8 and did not include 0, indicating that it is statistically reliable. A reduction of approximately 4 to 23 days in the past month on which negative physical symptoms are experienced also seems clinically meaningful (note that total number of days during which symptoms were experienced potentially was fewer for both groups because several symptoms experienced on any one day each was counted as a symptom day).
EMO and POS participants had significantly fewer medical appointments for cancer-related morbidities than did the CTL group participants during the 3 months after study completion. As expected, no between-groups differences emerged on all other medical appointments combined over the 3 months. This effect also was obtained when baseline self-reported somatic symptoms were controlled statistically, as well as when demographic and medical variables were controlled. The CIs for both the comparison of the expressive disclosure condition (95% CI, -3.0 to -0.6) and the benefit-finding condition (95% CI, -2.5 to -0.1) with the control condition did not include 0. Thus, at the population level, participants in an expressive disclosure condition would be expected to have at least 0.6 fewer and as many as three fewer medical appointments for cancer-related problems than participants in the CTL condition, a finding that seems both statistically reliable and clinically meaningful. Although statistically reliable, the difference between the finding-benefit and the control conditions may be less clinically meaningful (0.1 to 2.5 fewer appointments in the benefit-finding condition).
Findings support the hypothesis that promoting expression of the full range of thoughts and feelings in breast cancer patients produces benefit with regard to the primary dependent variable of physical health-related outcomes. At 3 months, women who expressed their emotions through writing reported significantly fewer negative physical symptoms and had fewer medical appointments for cancer-related morbidities than did CTL participants. These findings are consistent with longitudinal, naturalistic research revealing the benefits of coping through emotional expression in breast cancer patients18 and with experimental studies of expressive disclosure in other samples.2,3 Women who wrote about the positive consequences of their experience fell between the other two groups on somatic symptoms, and they had significantly fewer medical appointments for cancer-related morbidities at 3 months than did the CTL group. That the POS condition had salutary effects on health is consistent with recent studies25 revealing that participants’ use of positive emotion words in essays is associated with health benefits in studies in which they are induced to write about their deepest thoughts and feelings and with correlational26 and experimental27 research on benefit finding in other samples. The narrow advantage of expressive disclosure over benefit finding on health-related outcomes requires replication in a larger sample to assess its reliability and potency. How did the induction of emotional expression and benefit finding confer health benefits? Emotional expression may serve as a vehicle for clarifying and pursuing goals.13,18,28 Thus, expressive writing may have prompted women to make efficient use of their scheduled medical appointments or take other proactive actions to address medical concerns and consequently to require fewer medical appointments for cancer-related morbidities. Another possibility is that the EMO and POS conditions buffered women’s reactions to physical symptoms and that CTL participants expressed lingering distress through somatic channels. Direct physiological mechanisms (eg, immune modulation29-31 and decreased sympathetic activation13,32,33) for the effects of the interventions also are possible, although the links between such mechanisms and the outcomes assessed in this trial require study. In addition, EMO and POS conditions may have had distinct mechanisms of action. Mediating processes between such expressive interventions and health outcomes warrant investigation. A separate manuscript is planned to investigate potential mediators of the obtained effects. In contrast to the main effects of condition on physical health-related outcomes, significant benefits of the writing conditions on psychological outcomes varied as a function of participants’ cancer-related avoidance (ie, intentional attempts to avoid thoughts and feelings related to cancer). The current sample of breast cancer patients reported more positive psychological adjustment at study entry than evidenced by other cancer patient samples,7,18-20 perhaps rendering significant improvement in quality of life across participants unlikely. However, participants displayed wide variability in psychological adjustment across assessment points, particularly as indicated by POMS distress. Such variability renders it important to identify moderators of intervention effects (note that a meta-analysis2 of written expressive disclosure experiments also revealed more variable psychological than reported physical health effects). Indeed, induced expressive disclosure was more beneficial with regard to psychological distress for women low on cancer-related avoidance, and benefit finding conferred a somewhat greater advantage for more avoidant women. Certainly, asking women to explore their deepest thoughts and feelings regarding cancer may be threatening when they have devoted considerable effort to avoiding just those experiences. Rather, requesting that they generate benefits from the experience of cancer likely required less painful processing and hence may have produced more favorable psychological effects. In contrast, women low in avoidance may embrace an expressive disclosure intervention, using their already developed skills to process and express the full range of emotions effectively. The benefit-finding condition may be somewhat less useful for women low in avoidance because it discourages full emotional processing and expression. We should note that inducing participants to focus on the benefits of their breast cancer experience did not seem to produce extreme emotional suppression, which has been demonstrated in previous research to yield negative adaptive consequences.32 Only two participants in that condition noted in written comments that they felt emotionally constrained by it. Interestingly, both were low in cancer-related avoidance. In contrast, several participants noted that they were initially surprised and then pleased at being requested to write about positive consequences, given that they had been more likely to express negative emotions up until study entry. Given previous processing of negative emotions, the POS manipulation may have encouraged a balanced emotional evaluation of both the positive and negative aspects of women’s experience with breast cancer. Certainly, the obtained benefits of the POS condition might be dependent on the timing of the study in the trajectory of the breast cancer experience. An instruction to generate benefits shortly after diagnosis or in the midst of treatment might have been much more difficult for women to perform effectively and might have induced suppression of negative emotions, yielding more negative outcomes. We strongly endorse cautions34-35 against interventions11 promoting a sole focus on positive thinking in cancer patients. Although the present trial demonstrated benefits of writing about positive consequences, a rather superficial focus on urging positive thinking may enable maladaptive forms of avoidance.36 Moreover, imposing the suggestion of specific benefits may be perceived as insensitive or as minimizing the affected person’s experience.6 The benefit-finding condition may have been effective because it prompted self-generation of cancer-related benefits rather than external imposition of them and because it occurred after women had had the chance to process negative emotions. Clearly, the most appropriate format and timing of interventions to allow women to consider positive consequences of what typically is viewed as a stressful or traumatic experience require study. Limitations of this trial deserve mention. A first potential limitation includes a primary reliance on participant self-report. However, questionnaire measures were psychometrically sound and empirically validated. Further, that medical appointments for cancer-related morbidities and self-reported physical symptoms comprised valid indicators of physical health outcomes is supported by: (1) the finding that a subset of self-reported medical appointments corresponded closely with medical records, (2) the positive association of the two indicators of health outcomes, and (3) their positive relations with psychological distress and poorer quality of life. It also should be noted that self-reported physical symptoms (and psychological distress) declined from study entry to the 3-month follow-up for all groups, indicating that the intervention effects resulted from a greater decline in symptoms in the experimental conditions than in the CTL condition rather than from amplification of symptoms in the CTL group. A second limitation regards generalizability of the findings. Participants evidenced somewhat more positive psychological adjustment than other samples,7,18-20 and generalizability to very distressed or less motivated women requires study. Further, applicability of the findings to men, individuals with metastasized disease, and cancer patients at other points in the treatment trajectory and of diverse backgrounds warrants examination. Extension of the findings to longer-term psychological adjustment and cancer morbidity and mortality also is necessary. It also should be noted that three women terminated participation after randomization (two had not yet been informed of condition assignment). Although 81% of women introduced to the research completed it, and completers did not differ from noncompleters on baseline variables, it must be acknowledged that the noncompleters’ missing data precluded a true intent-to-treat analysis. A third limitation involves the relatively small sample size. Although power was greater than 0.80 to detect effects of the size obtained in this study, it was not adequate to detect small to medium effects.16 In addition, although consistent multivariate effects emerged on the major dependent variables, effects were not entirely uniform across single dependent variables, and confidence intervals for main effects were relatively broad. Further, outcomes differed on dependent variables relevant to physical versus psychological status, with the former revealing main effects of experimental condition and the latter yielding moderated effects. Stability and clinical import of these effects require study with larger samples. Given these limitations, applied implications of the results must be advanced judiciously. Findings suggest that emotional expression and benefit finding may be among the effective ingredients in multimodal interventions delivered in various formats (eg, group therapy9,10,12) and argue for their inclusion in intervention packages. Results also allow the tentative suggestion that specific writing interventions can be implemented productively with women treated for early-stage breast cancer. Although adherence cannot be guaranteed without oversight by a clinician, such interventions carry the advantages of being relatively brief, easily administered, and beneficial with regard to physical health outcomes. However, with regard to alleviating psychological distress, prescriptions for expressive writing may require tailoring to patient characteristics. Women who report avoiding cancer-related thoughts and feelings may respond more positively to a benefit-finding intervention than to expressive disclosure. Perhaps promoting more prolonged expressive disclosure than was conducted in this study may yield benefits for avoidant women, given that repeated exposure can prompt habituation to negative emotions.37 A more intensive expressive disclosure intervention may yield both physical and psychological benefits for both highly avoidant and less avoidant women. Further, the utility of integrating writing interventions into a clinical situation that promotes subsequent therapeutic discussion of the written material requires study, as does when and for whom particular writing interventions are most effective. Notwithstanding the trial’s limitations, the findings reveal that effective ingredients of physical health maintenance after breast cancer diagnosis and treatment are the ability to express emotions surrounding one’s experience and to generate benefits accrued from the experience. The positive consequences of induced emotional expression and benefit finding were apparent even several months after cancer diagnosis. Not only may benefits accrue for the breast cancer patient herself, but also such psychosocial interventions may promote more efficient medical care use.
Supported by the United States Army Medical Research and Materiel Command under grant no. DAMD-17-94-J-4244.
1. Pennebaker JW, Beall SK: Confronting a traumatic event: Toward an understanding of inhibition and disease. J Abnorm Psychol 95: 274-281, 1986[CrossRef][Medline] 2. Smyth JM: Written emotional expression: Effect sizes, outcome types, and moderating variables. J Consult Clin Psychol 66: 174-184, 1998[CrossRef][Medline]
3. Smyth JM, Stone AA, Hurewitz A, et al: Effects of writing about stressful experiences on symptom reduction in patients with asthma or rheumatoid arthritis: A randomized trial. JAMA 281: 1304-1309, 1999 4. Kelley JE, Lumley MA, Leisen JCC: Health effects of emotional disclosure in rheumatoid arthritis patients. Health Psychol 16: 331-340, 1997[CrossRef][Medline] 5. Walker BL, Nail LM, Croyle RT: Does emotional expression make a difference in reactions to breast cancer? Oncol Nurs Forum 26: 1025-1032, 1999[Medline] 6. Affleck G, Tennen H: Construing benefits from adversity: Adaptational significance and dispositional underpinnings. J Pers 64: 899-922, 1996[CrossRef][Medline] 7. Stanton AL, Snider P: Coping with a breast cancer diagnosis: A prospective study. Health Psychol 12: 16-23, 1993[CrossRef][Medline] 8. Carver CS, Pozo C, Harris SD, et al: How coping mediates the effect of optimism on distress: A study of women with early stage breast cancer. J Pers Soc Psychol 65: 375-390, 1993[CrossRef][Medline] 9. Fawzy FI, Fawzy NW, Hyun CS, et al: Malignant melanoma: Effects of an early structured psychiatric intervention, coping, and affective state on recurrence and survival 6 years later. Arch Gen Psychiatry 50: 681-689, 1993[Abstract] 10. Spiegel D, Bloom JR, Kraemer HC, et al: Effect of psychosocial treatment on survival of patients with metastatic breast cancer. Lancet 2: 888-890, 1989[CrossRef][Medline] 11. McWilliams P: You Can’t Afford the Luxury of a Negative Thought. Los Angeles CA, Prelude Press, 1997 12. Antoni MH, Lehman JM, Kilbourn KM, et al: Cognitive-behavioral stress management intervention decreases the prevalence of depression and enhances benefit finding among women under treatment for early-stage breast cancer. Health Psychol 20: 20-32, 2001[CrossRef][Medline] 13. Stanton AL, Kirk SB, Cameron CL, et al: Coping through emotional approach: Scale construction and validation. J Pers Soc Psychol 78: 1150-1169, 2000[CrossRef][Medline] 14. Lumley MA, Tojek TM, Macklem DJ: The effects of written emotional disclosure among repressive and alexithymic people, in Lepore SJ, Smyth JM (eds): The Writing Cure: How Expressive Writing Promotes Health and Emotional Well-Being. Washington DC, American Psychological Association, 2002, pp 75-95 15. Pennebaker JW: Putting stress into words: Health, linguistic, and therapeutic implications. Behav Res Ther 31: 539-548, 1993[CrossRef][Medline] 16. Cohen J: Statistical Power Analysis for the Behavioral Sciences ( ed 2 ). Hillsdale NJ, Erlbaum, 1988 17. McNair DM, Lorr M, Droppleman LF: EITS Manual for the Profile of Mood States. San Diego CA, Educational and Industrial Testing Service, 1971 18. Stanton AL, Danoff-Burg S, Cameron CL, et al: Emotionally expressive coping predicts psychological and physical adjustment to breast cancer. J Consult Clin Psychol 68: 875-882, 2000[CrossRef][Medline] 19. Stanton AL, Estes MA, Estes NC, et al: Treatment decision making and adjustment to breast cancer: A longitudinal study. J Consult Clin Psychol 66: 313-322, 1998[CrossRef][Medline] 20. Cella DF: F.A.C.T. Manual: Functional Assessment of Cancer Therapy (FACT) Scales. Chicago IL, Rush-Presbyterian-St Luke’s Medical Center, 1994 21. Pennebaker JW: The Psychology of Physical Symptoms. New York NY, Springer-Verlag, 1982 22. Carver CS, Scheier MF, Weintraub JK: Assessing coping strategies: A theoretically based approach. J Pers Soc Psychol 56: 267-283, 1989[CrossRef][Medline]
23. Horowitz M, Wilner N, Alvarez W: Impact of event scale: A measure of subjective stress. Psychosom Med 41: 209-218, 1979 24. Aiken LS, West SG: Multiple Regression: Testing and Interpreting Interactions. Newbury Park CA, Sage, 1991 25. Pennebaker JW, Mayne TJ, Francis ME: Linguistic predictors of adaptive bereavement. J Pers Soc Psychol 72: 864-871, 1997 26. Tennen H, Affleck G: Finding benefits in adversity, in Snyder CR (ed): Coping: The Psychology of What Works. New York NY, Oxford University Press, 1999, pp 279-304
27. King LA, Miner KN: Writing about the perceived benefits of traumatic events: Implications for physical health. Pers Soc Psychol Bull 26: 220-230, 2000 28. Frijda NH: Emotions are functional, most of the time, in Ekman P, Davidson RJ (eds): The Nature of Emotion: Fundamental Questions. New York NY, Oxford University Press, 1994, pp 112-122 29. Lutgendorf SK, Antoni MH, Kumar M, et al: Changes in cognitive coping strategies predict EBV-antibody titre change following a stressor disclosure induction. J Psychosom Res 38: 63-78, 1994[CrossRef][Medline] 30. Pennebaker JW, Kiecolt-Glaser J, Glaser R: Disclosure of traumas and immune function: Health implications for psychotherapy. J Consult Clin Psychol 56: 239-245, 1988[CrossRef][Medline] 31. Petrie KJ, Booth RJ, Pennebaker JW, et al: Disclosure of trauma and immune response to a hepatitis B vaccination program. J Consult Clin Psychol 63: 787-792, 1995[CrossRef][Medline] 32. Gross JJ, Levenson RW: Hiding feelings: The acute effects of inhibiting negative and positive emotion. J Abnorm Psychol 106: 95-103, 1997[CrossRef][Medline]
33. Pennebaker JW, Barger SD, Tiebout J: Disclosure of traumas and health among Holocaust survivors. Psychosom Med 51: 577-589, 1989 34. Holland JC, Lewis S: The Human Side of Cancer: Living With Hope, Coping With Uncertainty. New York NY, Harper Collins, 2000 35. Rittenberg CN: Positive thinking: An unfair burden for cancer patients? Support Care Cancer 3: 32-39, 1995 36. Lutgendorf SK, Anderson B, Larsen K, et al: Cognitive processing, social support coping, and distress in gynecologic cancer patients. Cancer Res Ther Cont 8: 123-137, 1999 37. Foa EB, Kozak MJ: Emotional processing of fear: Exposure to corrective information. Psychol Bull 99: 20-35, 1986[CrossRef][Medline] Submitted August 18, 2000; accepted July 11, 2002. This article has been cited by other articles:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
INTRODUCTION
2 analyses for categorical variables were conducted on demographic variables (ie, age, education, marital status, and employment status), cancer-related parameters (ie, stage, diagnosis duration, medical treatments received, and group or individual psychological support participation), and dependent variables (ie, quality of life, positive and negative affect, and somatic symptoms) at baseline. No significant differences emerged as a function of experimental condition or cancer-related avoidance on any demographic, cancer-related, or dependent variable. When relevant, baseline values on dependent variables were used as covariates to control for nonsignificant chance variation between groups at baseline. An important question potentially affecting the interpretation of findings is whether the groups were equivalent at study entry with regard to independently sought psychological support (ie, support group or mental health professional) and whether the writing conditions might prompt differential seeking of psychological support. Analyses revealed that the participants who had sought support at study entry (43%) or during the study (additional 9%) were distributed equally across conditions, 
. Only when the multivariate F was significant were the univariate analyses examined to determine the locus of the effect, controlling for the relevant baseline value. Means and SEs reported are adjusted for the covariate and for slightly unequal cell sizes. Significant condition x cancer-related avoidance interactions were interpreted via an accepted strategy.
