Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
5 The Effectiveness of Psychosocial Interventions for Women with Breast Cancer H ealth-care providers are offering and women are seeking a range of interventions to manage psychosocial distress following a diag- nosis of breast cancer, but there is no strong evidence-based con- sensus indicating which interventions are effective in reducing distress and improving quality of life. This chapter provides a critical review of the most relevant and highest quality studies of the effectiveness of psychosocial inter- ventions in breast cancer. Included are published randomized trials and selected non-randomized studies that were designed to evaluate the effec- tiveness of psychosocial interventions for women with breast cancer.1 Excluded is the psychosocial literature in other cancers or in mixed groups of cancer patients, in which only some patients have breast cancer. Also not covered is the extensive literature relating to psychological issues surround- ing breast cancer screening or identification of high-risk individuals. The effectiveness of the treatment of mental disorders such as major depression is not examined, although, according to a recent review, pharmacologic treatment of cancer patients who are clinically depressed is effective if anti- depressants are administered appropriately (Agency for Healthcare and Research and Quality, 2002). Underlying psychiatric disorders may worsen 1This chapter is based on a commissioned paper by Pamela J. Goodwin, M.D., M.Sc., FRCPC, Marvelle Koffler Breast Centre, Department of Medicine, Division of Epidemiology, Samuel Lunenfeld Research Institute, Mount Sinai Hospital, University of Toronto and pre- sented at the National Cancer Policy Board workshop âMeeting Psychosocial Needs of Women with Breast Cancerâon October 28â29, 2002, at the Institute of Medicine in Washington, DC. 95
96 MEETING PSYCHOSOCIAL NEEDS OF WOMEN WITH BREAST CANCER or be exacerbated following a diagnosis of cancer, and psychopharmaco- logic treatment can reduce distress and improve quality of life. The chapter begins with a brief history of psychosocial intervention research in breast cancer followed by a discussion of methodological issues that are crucial to assessing the effectiveness of psychosocial interventions. An enumeration of research priorities to address identified knowledge gaps is outlined in Chapter 8. HISTORY OF PSYCHOSOCIAL INTERVENTION RESEARCH In the past decade, considerable resources have been allocated to re- search into the psychosocial aspects of breast cancer. Much of this research has focused on describing the emotional experience of women with breast cancer and developing interventions that reduce the psychosocial distress and improve coping and adjustment. Early intervention studies were done by Ferlic and colleagues (1979) and by Heinrich and Schag (1985), both with positive psychosocial results. A report by David Spiegel et al. that appeared in The Lancet in 1989 suggested, for the first time, that a psycho- logical intervention (supportiveâexpressive group therapy) might prolong survival in women living with metastatic breast cancer (Spiegel et al., 1989). This observation heightened interest in survival effects of psychosocial in- terventions and led to a series of intervention studies in breast cancer (Classen et al., 2001; Cunningham et al., 1998; Edelman et al., 1999a; Goodwin et al., 2001) and other cancers (Fawzy et al., 1993; Ilnyckyj et al., 1994; Kuchler et al., 1999; Linn et al., 1982), conducted mainly in the 1990s. These studies attempted to confirm a beneficial effect of psychologi- cal interventions on survival. None of the subsequent studies in metastatic breast cancer has identified a survival effect of a range of psychological interventions (although one is ongoing) (Cunningham et al., 1998; Edelman et al., 1999a; Edmonds et al., 1999; Goodwin et al., 2001). Three studies in other malignancies (leukemia, melanoma, and mixed GI cancers) have re- ported survival benefits for brief interventions, delivered around the time of or shortly after diagnosis or during treatment (Fawzy et al., 1993; Kuchler et al., 1999, Richardson et al., 1990). These results have not been repli- cated. Nonetheless, throughout much of the 1990s, the focus of some mem- bers of the psycho-oncology research community shifted from evaluation of psychological effects of interventions to survival and intermediate biomedi- cal effects, including enhancement of immune function. A growing recogni- tion has developed that the survival effect in metastatic breast cancer that was seen by Spiegel et al. (Spiegel et al., 1989) has not been replicated and that little evidence supports the prognostic importance of intermediate out- comes such as cortisol levels that have been linked to psychosocial interven- tions, although it has been suggested that the Spiegel study at least indicated
EFFECTIVENESS OF PSYCHOSOCIAL INTERVENTIONS 97 that these kinds of psychosocial interventions did not damage patientsâ cop- ing skills in a way that would limit survivals. Spiegel has undertaken a replication study currently in progress. Research focus is once again returning to the psychological status and quality of life of women with breast cancer, and to the identification of interventions that favorably influence their psychological and social func- tioning. An example is the report by Kissane and colleagues (2003) of a randomized, controlled trial of cognitiveâexistential group therapy for women with early breast cancer. It found the women in the intervention group showed significantly lower anxiety, and improved family function. The authors reported self-growth and increased knowledge of cancer and its treatment. A number of excellent reviews of psychosocial interventions in breast cancer, and cancer in general, have been published recently (Burke and Kissane, 1998; Fawzy et al., 1995; Meyer and Mark, 1995; Newell et al., 2002; Rimer et al., 1985; Wallace, 1997). The majority of these reviews concluded that there are important benefits associated with the use of psychological interventions of various types in various cancer settings, including breast cancer. Fawzy et al. reviewed research into education, behavioral training, individual psychotherapy, and group interventions in cancer patients in general (Fawzy et al., 1995). They concluded that there was evidence of benefit for all of these approaches, reporting that cancer patients may benefit from a range of psychological intervention programs, and suggesting specific interventions at different points along the cancer trajectory. Meyer and Mark conducted a meta-analysis of psychosocial interventions in adult cancer patients (Meyer and Mark, 1995). Effect sizes (treatment mean minus control mean divided by pooled standard deviation) of 0.19â0.28 for emotional adjustment, functional adjustment, and global measures were identified. These effect sizes were in the range considered clinically important, suggesting significant benefit from the interventions. Burke and Kissane prepared a review of psychosocial inter- ventions provided to breast cancer patients by specialist providers (includ- ing both published and unpublished data), drawing fairly extensively from the general cancer literature (Burke and Kissane, 1998). They concluded that the benefits of psychosocial interventions in breast cancer were clearly established and recommended that such interventions be an âintegral part of comprehensive medical care.â They presented a series of recommenda- tions for future research and made more specific recommendations re- garding integration of psychosocial interventions into the treatment set- ting. These were published in an overview paper outlining the need to challenge the health delivery system to improve and implement psychoso- cial and supportive care (Redman et al., 2003). Two targeted reviews con- cluded beneficial effects were present for relaxation and imagery interven-
98 MEETING PSYCHOSOCIAL NEEDS OF WOMEN WITH BREAST CANCER tions (Wallace, 1997) and for educational interventions in patients with cancer (Rimer et al., 1985). All of these reviews identified important areas for additional research. One recent review reached somewhat different conclusions (Newell et al., 2002). The authors of this review undertook an extensive review of psychological therapies in all types of cancerâthey applied a series of rigorous methodological standards and retained only those reports that scored one-third or more of the total possible points. This resulted in exclusion of the majority of published studies. They also took a very conservative approach to evaluation of benefits, requiring that at least half of the outcome measures for a specific attribute (e.g., mood) yield significant results for the effect to be classified as significant. This as- sumes that all of the questionnaires were equally responsive to change, a situation that is not likely. They examined the short, medium, and long- term impact of interventions on a large number of outcomes including anxiety, depression, hostility, general or overall affect, stress or distress, general or overall functional ability or quality of life, coping or coping skills, vocational or domestic adjustment, interpersonal or social rela- tionships, sexual or marital relationships, nausea, vomiting, pain, fa- tigue, overall physical symptoms, conditioned nausea, conditioned vom- iting, survival, and immune outcomes. They did not distinguish the effects of interventions among different types of cancers. These authors concluded that âonly tentative recommendations about the effectiveness of psychosocial therapies for improving cancer patientsâ outcomesâ were possible and recommended that future trials adhere to minimum meth- odological standards. Examples of their tentative recommendations in- cluded the use of music therapy to reduce anxiety and enhance overall affect, group therapy to enhance coping skills, counseling to enhance social or interpersonal relationships, and hypnosis to reduce conditioned nausea and vomiting. Further research was recommended into the ben- efits of these and other forms of therapy, including cognitive behavioral therapy, education, individual therapy, and interventions involving sig- nificant others. This review was thorough, but it did not focus on a specific type of cancer or a specific type of treatment, and its strict appli- cation of exacting methodologic criteria (some of which may not have been well suited to psychosocial research) places it at the conservative end of the spectrum of analyses and may have resulted in important effects of psychosocial interventions being missed or undervalued. Nev- ertheless, the Board realizes that attention needs to be paid to stronger research designs in future studies. Relatively few studies among those reviewed in this chapter may relate to a specific intervention, so that expanded research will be needed to strengthen evidence for a particular treatment.
EFFECTIVENESS OF PSYCHOSOCIAL INTERVENTIONS 99 METHODOLOGICAL ISSUES There are a number of key methodological characteristics that are critical to the success of studies assessing the effectiveness of psychosocial interven- tions. Many of these methodological characteristics are similar in studies evaluating biomedical or psychosocial interventions. Others, notably those related to measurement and to standardization of the intervention, are unique challenges for researchers conducting psychosocial intervention trials. Study Design The gold standard design for studies that evaluate effectiveness of vari- ous interventions is the randomized controlled trial. The purpose of ran- domized allocation of patients is to avoid bias or confounding in the assign- ment of study treatments, thereby ensuring that patients in each arm of the study are as comparable as possible with respect to all characteristics except for the specific intervention(s) being studied. When baseline patient charac- teristics are potentially associated with effects of the intervention being stud- ied (e.g., severity of anxiety or depression), methodological approaches to ensure balance between study arms (e.g., stratification) are recommended as part of the randomization process. This is particularly important in stud- ies with a small sample size. Although randomized trials have been recognized as the gold standard since the mid-twentieth century, concerns have been raised that they may not be the most appropriate study design for evaluation of psychosocial interventions in cancer patients, in part because patient commitment to the intervention is deemed an important predictor of benefit from the interven- tion (Cunningham et al., 1998). It has been argued that if patients are will- ing to be randomized to a control arm, they are not as strongly committed to the intervention, and this will lead to an underestimation of the benefits of the intervention. Although this is possible, it is difficult to overcome self- selection biases (which may be powerful predictors of psychosocial out- comes) without using a randomized study design. It has also been argued that cancer patients may not be willing to accept randomization. This might occur not only in psychosocial trials but also in biomedical intervention trials. However, the success of the many random- ized trials reviewed below argues against this as a major obstacle to success- ful conduct of randomized psychosocial intervention trials in breast cancer. When a randomized design is selected, it is important that the random- ization not be compromised for logistical, or other, reasons. In the studies reviewed here, an example of compromised randomization occurred in an otherwise well conducted trial of the contribution of coaches to support groups for women with breast cancer when re-randomization was permit-
100 MEETING PSYCHOSOCIAL NEEDS OF WOMEN WITH BREAST CANCER ted in the two intervention arms (but not the control arm) if logistics of group meetings for the original randomized allocation were not convenient (Samarel et al., 1997). Another potential example occurred when assign- ment to a specialist nursing intervention versus control was according to the week women were admitted for surgery (weeks, not patients, were ran- domized) (Maguire et al., 1980). Theoretically, advance knowledge of whether the intervention was being administered in a given week could have influenced admission dates for individual women, leading to non-com- parable study groups. The issue of selection of an appropriate control or comparison arm is also challenging in randomized psychosocial intervention trials. Although many researchers select a âno treatmentâ or âstandard treatmentâ control group, others argue that the control group should receive a similar amount of attention to that received by the intervention group, so that beneficial effects of attention (as opposed to the âactiveâ components of the interven- tion) are not falsely labeled as benefits of the intervention. Selection of a no- treatment (or standard treatment) control arm versus attention control arm should probably reflect the specific study question. If the purpose of the study is to evaluate the overall benefits of adding the psychosocial interven- tion being studied to routine clinical care, then a no-treatment (or standard treatment) control group would be most appropriate. If, however, research- ers are attempting to delineate which aspects of the intervention (e.g., atten- tion versus teaching of coping skills or a specific cognitiveâbehavioral thera- peutic approach) are important determinants of benefit, then selection of an attention control group may be more appropriate. The selection of control conditions will become an even more challenging issue as the benefits of psychosocial interventions are increasingly accepted by the medical com- munity and these interventions become standard care; these interventions will then become the control or comparison arm in future studies. To facili- tate this selection of comparison arms, agreement will be necessary regard- ing which interventions are considered standard care. Some studies reviewed used âwait-listâ controls. These controls were offered the study intervention after all study measurements were completed. Other studies passively offered educational materials to control subjects. Although these approaches may enhance acceptability of the study design to investigators, potential participants, and ethics committees alike, they are not truly no-treatment approaches and their use may tend to diminish treatment effects or, in the case of âwait-listâ controls, make it difficult to examine long-term effects of interventions. These issues should be consid- ered in both design and interpretation of trials. Finally, there are situations when non-randomized designs may be de- sirable. These include studies whose purpose is to develop and standardize an intervention, studies whose purpose is to demonstrate that an interven-
EFFECTIVENESS OF PSYCHOSOCIAL INTERVENTIONS 101 tion is feasible and can be delivered in a standard fashion by more than one investigator, and pilot studies looking for early evidence of treatment effec- tiveness, before full commitment is made to an expensive, long-term ran- domized trial. Non-randomized designs may also be important for studies seeking to obtain descriptive information about patient experiences or in- formation about change over time, as well as for studies examining prog- nostic effects of psychosocial status. Such non-randomized designs may also be used when randomization is difficult, such as in the evaluation of peer support, novel and/or alternative therapies (e.g., reiki, yoga or tâai chi), or participation in Internet chat support rooms where access to the interven- tion may be available to controls outside of the study. However, when as- sessment of treatment effectiveness is the primary purpose of a study, a randomized design is optimal. Use of alternative designs may lead to biased or inaccurate results. Study Population Many published psychosocial intervention studies involve patients with more than one type of cancer or patients with the same type of cancer in various stages of their illness (for example, early stage curable and late stage metastatic). Although, in the long term, intervention studies may demon- strate that the effectiveness of psychosocial interventions does not vary across cancer type or cancer stage, this is not known a priori. More sharply focused trials include patients with one type of cancer, and, in most circum- stances, one stage of cancer. In the reports reviewed here, only randomized trials that restricted study entry to breast cancer patients are included. Once the population of interest is identified (e.g., women with recently diagnosed breast cancer), the study population should be as representative of that population as possible. The characteristics of the women in the trials re- ported here are listed (Phase/Stage of Disease) in Appendix B. Although the review of Newell and colleagues (Newell et al., 2002) was critical of low reporting of randomization (about 25 percent) in the trials they reviewed, in the present case with the exceptions noted earlier, randomization ap- pears acceptable. Furthermore, the populations offered the trials were drawn consecutively from, responded to advertisements or other solicitations to, or were identified through records of, patient groups similar to those that might be offered the intervention in routine practice, although this was not specifically stated in those words, with the caveat that many of the studied populations were in academic health centers. It would be interesting to study patients in a larger cross-section of settings. It is possible that attributes of the study population, such as age, medi- cal treatment received, or baseline psychological characteristics will be pre- dictors of effectiveness of psychosocial interventions. This was the case in a
102 MEETING PSYCHOSOCIAL NEEDS OF WOMEN WITH BREAST CANCER recent study in which the psychological benefits of supportiveâexpressive group therapy on mood were present in women who were distressed at study entry, but not in non-distressed women (Goodwin et al., 2001). Pa- tient attributes that are potential predictors of benefit from the study inter- vention should be identified and their contribution to study outcomes care- fully examined. Ideally, randomization should be stratified for these key baseline characteristics, as was discussed above. Finally, attempts are warranted to design research on larger sample sizes. About half the 31 trials reviewed in this chapter involved fewer than 100 subjects, in many cases 50 or fewer, weakening their power to provide strong evidence. The largest trial involved 312 women randomized to four arms and did not demonstrate lasting effects (Helgeson et al., 1999, 2000, 2001). The recent study from Australia by Kissane and colleagues random- ized 303 women to group and relaxation versus relaxation alone. Positive findings of reduced anxiety and better function were found in this large cohort. Study Intervention Standardization of the study intervention is a particular challenge when the intervention is psychosocial in nature. As much as possible, investiga- tors should use a well described, standardized therapy that has the potential to be delivered in routine practice by other trained and qualified practitio- ners who will be the end-users of the study results. When describing the intervention, the skills and the qualifications of the therapists must be in- cluded along with details of the intervention. Furthermore, it should be demonstrated that the therapy is delivered in a standardized fashion throughout the study. Ideally, investigators should provide evidence of com- pliance with the intervention as well as evidence of the competence of thera- pists delivering it. They should also report the dose/duration of exposure to the study intervention, and designs of longer duration should be considered for future studies. These issues are more straightforward when interven- tions are biomedical because investigators can simply describe the number of doses and number of milligrams of a drug that are administered. The challenge is much greater when the intervention is psychosocial. The interventions that were investigated in the reports reviewed here often involved more than one potentially active component. For example, group therapy, relaxation/self-hypnosis, and teaching of coping skills were included in a single intervention (Cunningham et al., 1998). When interven- tions have multiple components, it may be difficult for investigators to determine whether one, or all, of the components is responsible for the outcome. This is not a problem if it is the overall effectiveness of the inter- vention that is being investigated. However, if evidence of effectiveness of
EFFECTIVENESS OF PSYCHOSOCIAL INTERVENTIONS 103 individual components is sought, it may be necessary to study each compo- nent separately. Study Measures Measurement of psychological outcomes is as exact a science as mea- suring biomedical outcomes, and requires the same attention to accuracy. There are many well validated, standardized instruments available that measure quality of life and psychosocial status. However, not all of these instruments measure attributes that are likely to be altered by specific study interventions (see Chapter 3 for a description of selected instruments). For example, an intervention targeting anxiety would not necessarily be ex- pected to enhance role functioning. Furthermore, not all of the available instruments are sufficiently sensitive to detect clinically important change resulting from an intervention that is, in fact, efficacious. This was demon- strated in a recent randomized trial of supportiveâexpressive group therapy in metastatic breast cancer (Goodwin et al., 2001). Beneficial effects of the intervention on mood were readily detected using the Profile of Mood States. Similar benefits were not detected using the Emotional Functioning Subscale of the European Organization for Treatment and Research of Cancer (EORTC) QLQ-C30 (see Chapter 3). The selection of quality of life and psychosocial outcome measures in randomized trials in breast cancer pa- tients, and the ability of these instruments to identify changes when various interventions are delivered, has recently been reviewed by Ganz and Goodwin (Ganz and Goodwin, 2003). It is recommended that investigators select, whenever possible, stan- dardized, well validated, and well accepted outcome measures that are sen- sitive to clinically important change and that target attributes that are likely to be influenced by the intervention. For example, if the intervention is designed to reduce anxiety, then an instrument that specifically measures anxiety and is responsive to changes in anxiety should be used. As noted above, a general health-related quality of life (HRQOL) instrument that measures emotional functioning in a general way with a small number of items and limited response categories may miss important changes in mood that can be detected with other instruments. Thus, use of these general HRQOL instruments as sole outcome measures in psychosocial interven- tion studies is not recommended. At times, it will be necessary for investiga- tors to develop new instruments or new modules of existing instruments in order to ensure they are targeting key attributes. When this is the case, care should be taken to validate these new measures before they are used as study outcomes. However, wholesale development of new instruments when acceptable instruments already exist is not recommended because it reduces the ability to compare results across studies.
104 MEETING PSYCHOSOCIAL NEEDS OF WOMEN WITH BREAST CANCER An additional challenge in psychosocial intervention research relates to compliance with study measurements. A balance must be reached between patient burden and comprehensiveness. Every effort should be made to en- courage patients to comply with study measurements. The problem of miss- ing data points has been a troublesome issue in psychosocial research. This is particularly challenging when patients are seriously ill, or undergoing complex medical treatments, or when multiple outcome measures are being used. Blinding should be used in outcome assessments to the extent possible, given the study design. When outcomes are self-reported, patients should be unaware of the specific study hypothesis (if possible), and research staff who score the questionnaires should be unaware of subjectsâ randomization allocation. When outcomes are interview-based, interviewers should be blinded to randomization allocation. Finally, cost and feasibility assessments to allow the possible balancing of benefits against costs of psychosocial interventions or programs should be included insofar as is possible and consistent with the abilities of the investigators, time and money resources available, and capacities of the settings. It is recognized that this is easier said than done, and that with the exception of an occasional study or mention in reports is rarely accom- plished (Koocher and Pollin, 2001; Simpson et al., 2001). Outcome/Analysis Each randomized trial should have a clearly stated hypothesis and a clearly stated primary outcome; sample size calculations and statistical analyses should reflect this primary study question. The description of the primary outcome should include a description of the specific questionnaire and/or questionnaire item/scale as well as the specific time-point(s) in the study that will be used to define this outcome. If more than one primary study outcome (or time-point) is selected, allowances must be made for multiple testing in statistical analysis and sample size calculations. At times, the issue of multiple time-points can be overcome using statistical methods, such as repeated measures analysis of variance or slopes analysis, that in- corporate measurements performed at multiple times into a single statistical analysis. Thus, although studies may use multiple instruments administered at multiple time-points with a plan for multiple approaches to analysis, the primary endpoint, time-point, and analytic approach must be stated a priori, and the sample size and significance cut-points must reflect these decisions. Other endpoints and/or analyses should then be viewed as secondary or hypothesis-generating. This will overcome the common criticism of psycho- social intervention studies that use of multiple outcomes, all of which are treated equally in the analysis, leads to multiple, and at times conflicting,
EFFECTIVENESS OF PSYCHOSOCIAL INTERVENTIONS 105 study results with a high experiment-wide type-1 error (the probability of incorrectly accepting a study hypothesis that is not true) (Newell et al., 2002). In evaluating outcomes, it might also be useful to consider whether the length of follow-up should be related to the length of time that the treatmentâs effect is expected to be needed. For example, if the intervention being tested is intended to deal with a current problem, like getting through physical treat- ments like surgery, radiation, or chemotherapy, a long follow-up may not be required and could lead to an erroneous conclusion that the treatment was not effective because its effects did not last until the final follow-up. Also, since trials of psychosocial interventions evaluated here are often applied to the whole population of women with breast cancer, they may include some women who would not have, or would not develop, psychosocial problems. This might dilute results and explain lack of effect in some trials. The majority of the published psychosocial intervention studies in breast cancer reviewed here have selected psychosocial outcomes as their primary outcome measure, but some have also included biomedical outcomes, nota- bly survival, treatment response, and immune factors. As with psychosocial outcomes, when biomedical outcomes are being studied, the primary out- come must be identified a priori. The only exceptions to this rigorous ap- proach occur in hypotheses generating studies such as pilot studies, or in studies designed to standardize interventions. Descriptive qualitativeâana- lytic studies of cancer patientsâ narratives, or correlative studies examining inter-relationships of psychosocial variables that are designed to generate, rather than confirm, hypotheses are also excluded. Nonetheless, such stud- ies provide important hypotheses that are generated and subjected to con- firmation in future trials. REVIEW OF THE LITERATURE Methods Literature searches were conducted using computerized databases (MedLine, Cancer Lit) between March 2002 and August 2002 using the following headings: 1. breast cancer and psychological intervention 2. breast cancer and psychosocial intervention 3. breast cancer and relaxation intervention 4. breast cancer and hypnosis intervention 5. breast cancer and group therapy intervention 6. breast cancer and individual therapy intervention 7. breast cancer and psychotherapy intervention
106 MEETING PSYCHOSOCIAL NEEDS OF WOMEN WITH BREAST CANCER Searches were also made for reviews in each of these areas. The articles were obtained, and their reference lists were reviewed to identify additional reports. This process continued in an iterative fashion until no new reports were identified. Studies reported in abstract form only were not included as they provided insufficient information on methodology for adequate criti- cal review. The reports thus identified were included in this review if they described randomized trials of psychosocial interventions in breast cancer patients, and they provided information on psychosocial and/or biomedical out- comes. Pilot studies were included if they met these criteria, but not if they did not compare study arms (i.e., presentation of before/after data within study arms as the only outcome was not adequate). Studies in mixed cancer patients were not included (though some of these studies are briefly cited if they provide key data not otherwise available). In total, 31 randomized trials of psychosocial interventions meeting these criteria were identified, some of which were reported in a series of publications (Cunningham et al., 1998; Edelman et al., 1999b; Goodwin et al., 2001; Maguire et al., 1983, 1980; Samarel et al., 1997; Spiegel et al., 1989). Potentially important problems with randomization in two of these trials were noted above (Maguire et al., 1980; Samarel et al., 1997). These trials are summarized in Table 5-1 according to year of publication and phase of breast cancer (early, metastatic, healthy survivor). The majority of the published studies (24 of 31) have focused on early phases of breast cancer, that is, initial diagnosis and treatment, including surgery, chemo- therapy, and radiation therapy. The pace of publication is increasing, with about 80 percent (25 of 31) of the studies appearing since 1995. This un- TABLE 5-1 Randomized Trials of Psychosocial Interventions in Breast Cancer Year of Publication 1980 to 1985 to 1990 to 1995 to 2000 to Phase of Illness 1984 1989 1994 1999 mid-2002 Total âEarlyâ* 2 1 1 9 **11** 24 âMetastaticâ 1 0 1 2 2 6 âHealthy Survivorsâ 0 0 0 1 0 1 Total 3 1 2 12 13 31 **Includes all reports that enrolled patients at diagnosis during initial treatment (surgery, chemotherapy, radiation therapy. **One trial included a small number of women with metastatic breast cancer (Targ and Levine, 2002).
EFFECTIVENESS OF PSYCHOSOCIAL INTERVENTIONS 107 doubtedly reflects a growing interest in psychosocial intervention in breast cancer patients in the scientific, medical, and lay communities. The studies are further subdivided in Table 5-2 according to the type of intervention studied and the country in which the research was conducted. The majority of this psychosocial intervention research in breast cancer has been conducted in the United States (16 of 31 studies, 52 percent), and 13 of the studies examined group interventions (Antoni et al., 2001; Bultz et al., 2000; Classen et al., 2001; Cunningham et al., 1998; Edelman et al., 1999a, 1999b; Edmonds et al., 1999; Fukui et al., 2000; Goodwin et al., 2001; Helgeson et al., 1999, 2000, 2001; Richardson et al., 1997; Samarel and Fawcett, 1992; Samarel et al., 1993, 1997; Spiegel and Bloom, 1983; Spiegel et al., 1989). These group interventions varied in nature; some were supportive only, some involved manual-based supportiveâexpressive therapy, some examined cognitiveâbehavioral interventions or a combina- tion of cognitiveâbehavioral and supportive interventions, while one exam- ined psycho-education and peer discussion and one investigated a mindâ bodyâspirit complementary medicine intervention. The next largest group of studies evaluated individual therapies (Allen et al., 2002; Burton et al., 1995; Cimprich, 1993; Lev et al., 2001; Maguire et al., 1983, 1980; Marchioro et al., 1996; Maunsell et al., 1996; McArdle et al., 1996; Richardson et al., 1997; Ritz et al., 2000; Sandgren et al., 2000; Walker et al., 1999; Wengstrom et al., 2001, 1999). The interventions used in these studies were diverse and included telephone support and screening, a series of nursing interventions, couples therapy, cognitiveâbehavioral interven- tions, and others. Relaxation/hypnosis with or without imagery was also a major focus of several reports; in total six studies examined interventions of this type as their primary focus (several others included relaxation/hypnosis as part of multifaceted intervention) (Arathuzik, 1994; Bridge et al., 1988; Kolcaba and Fox, 1999; Molassiotis et al., 2002; Richardson et al., 1997; Walker et al., 1999). Although only one study formally evaluated the use of education as a form of psychosocial support, a number of other studies provided educational materials in a passive fashion to women randomized to control groups (Fukui et al., 2000; Helgeson et al., 2000, 2001). The study that formally evaluated education was also the only randomized trial to evaluate facilitated peer discussion; one other study (McArdle et al., 1996) evaluated support from a lay breast cancer organization. Given the characteristics of these studies, their results should be highly relevant to women with breast cancer receiving care in the United States today. The focus of many studies on the early phases of breast cancer re- flects the fact that this is often a stressful time for women diagnosed with breast cancer. It is a phase of breast cancer that affects all women, and it is a phase of the illness in which intervention is facilitated by the fact that women attend for treatment on a regular basis. The paucity of studies in
108 TABLE 5-2 Randomized Trials of Psychosocial Interventions in Breast Cancer Type(s) of Intervention Investigated (some studies investigated more than one intervention) Group Individual Relaxation Phase of Illness Country Education* Therapyâ Therapy** Imagery Other*** âEarlyâ (n=24) United States 12 1 8 10 5 11 Canada 4 UK/Europe 6 Asia 2 âMetastaticâ (n=6) United States 3 0 5 1 1 0 Canada 2 Australia 1 âHealthy Survivorsâ (n=1) United States 1 0 0 0 0 1 Total 1 13 11 6 12 * Educational materials provided passively to control groups not included. ** Includes telephone support, nursing intervention, cognitive-behavioral interventions, other. *** Includes telephone screening in early breast cancer, videotapes in healthy survivors. â Includes supportive or supportive-expressive group therapy, cognitive-behavioral interventions, peer discussion.
EFFECTIVENESS OF PSYCHOSOCIAL INTERVENTIONS 109 healthy survivors is a potential concern; the one study that has been re- ported used this group as a convenience sample to evaluate the psychosocial benefit derived from viewing a videotape that projected an empathic inter- action (enhanced compassion) (Fogarty et al., 1999). This study did not specifically address psychosocial issues of healthy survivors. One additional study evaluated the benefits of managing menopausal symptoms in breast cancer survivors using a comprehensive menopausal assessment with subse- quent recommendations for management (Ganz et al., 2000). Because the main focus of this trial was on medical symptom management, it was not included as a psychosocial intervention trial, even though a portion of the intervention was psychosocial in nature (pharmacologic interventions were also used). Survivors are a relatively understudied group in general (although research on survivorship is growing), and descriptive studies that focus on identifying psychosocial problems in this group and that determine the prevalence of these problems, as well as those at risk, are probably needed to guide development of interventions that can be tested in full-fledged ran- domized trials. Review of Clinical Trials This section of the chapter reviews each of the 31 randomized trials of psychosocial interventions in women with breast cancer identified in the literature search. Trials are first grouped into those taking place in the early phases of breast cancer, metastatic breast cancer, and during survivorship. Key characteristics, methodologic limitations and findings of each trial are presented (a detailed summary of each trial is included in Appendix B). Trials are then regrouped according to the intervention approach used (e.g., group, individual, relaxation/hypnosis). Table 5-3 provides definitions of the most common types of interventions used in these trials. Early Breast Cancer The first, and largest, group of trials involves those conducted in early breast cancer (see Table B-1). These 24 trials include patients enrolled at cancer diagnosis and those enrolled during initial treatment, including sur- gery, chemotherapy, and radiation therapy. One of these studies (Targ and Levine, 2002) included a small number of patients with metastatic breast cancer. In general, patients on these trials were within 1 to 2 years of breast cancer diagnosis and, in most trials, they were within a few weeks or months of diagnosis. These trials are listed chronologically in Table B-1 along with information on the number of women studied, the type of intervention evaluated, and the duration of the intervention, as well as study outcomes and the length of follow-up.
110 MEETING PSYCHOSOCIAL NEEDS OF WOMEN WITH BREAST CANCER TABLE 5-3 Definitions of Psychosocial Interventions Supportive-expressive psychotherapy A group therapy developed by Spiegel and colleagues in which social support is given in the context of encouraging expression of feeling, following a manualized format. Supportive-existential psychotherapy A group therapy developed by Kissane and colleagues which provides both social support and encouragement to explore existential fears, e.g., recurrence, and the need for meaning in the face of serious illness. Psychoeducational interventions Providing information to individuals and groups about illness in a social, supportive interaction. Interpersonal psychotherapy/counseling An intervention that uses a manual approach to focus on role changes and problems caused by illness, exploring constructive efforts to accommodate. Cognitive-behavioral interventions (See Chapter 4, Box 4-1). Use of cognitive restructuring, distraction, mental guided imagery, coping-enhancing interpretations along with common behavioral interventions of relaxation, hypnosis, desensitization. Mind-body-spiritual interventions This approach by Targ and Levine combines psychosocial and complementary medicine interventions (dance, meditation, imaging). Restorative interventions A structured nursing intervention to help women give attention to and participate in activities while reducing attention to negative aspects of illness and treatment resulting in attentional fatigue. Supportive group with coaching A group intervention in which women were âcoachedâ to improve self-concept, physical function, roles, and relationships. SOURCE: Jimmie Holland, personal communication. Maguire et al. reported the first randomized psychosocial intervention trial in breast cancer (Maguire et al., 1983, 1980), but randomization was not performed on individual patients. Instead, weeks during a 24 month period were randomized. Thus, it is possible surgeons or patients may have been aware of randomization allocation for a given week and that knowl- edge may have influenced admission dates. This might have led to bias in treatment allocation and non-comparability of study arms. The interven-
EFFECTIVENESS OF PSYCHOSOCIAL INTERVENTIONS 111 tion involved individual counseling by a nurse before and after mastectomy, with subsequent home visits every 2 months until the woman had âadapted well.â The focus of the intervention was on adaptation to the scar, arm morbidity, and breast prosthesis as well as encouraging women to be open regarding the effect of the surgery on themselves and their emotional status. The comparison patients received usual treatment. Outcomes were evalu- ated using an interviewer administered Present State Exam and Life Events Scale, as well as a linear analog scale of mood. Patients were followed for up to 12 to 18 months. Counseling failed to prevent morbidity; however, the nursesâ regular monitoring of the women led to more appropriate psy- chiatric referral, and subsequent psychiatric intervention reduced psychiat- ric morbidity as well as anxiety and depression. This early trial provides the first evidence of benefit of a psychosocial intervention in women with breast cancer. Christensen randomized 20 husband and wife pairs after mastectomy to receive four weekly counseling sessions or to a control condition (Christensen, 1983). The intervention focused on the impact of mastectomy on the coupleâs relationship, and it was tailored to each coupleâs needs. No overall treatment effects were identified, although the small sample size may have precluded identification of clinically important effects. Despite the absence of overall effects, adjusted analyses suggested that there might be some benefits in terms of enhanced sexual satisfaction and psychological status in both husbands and wives and reduced depression in wives. Fol- low-up was shortâonly 1 week post intervention. Effects of this short in- tervention on longer-term outcomes cannot be assessed. Cimprich studied an individualized, but structured nursing intervention designed to enhance attentional capacity in women with recently diagnosed breast cancer (Cimprich, 1993). The restorative intervention was designed to âminimize or prevent attentional fatigue through regular participation in activities that engage fascination or have other restorative properties.â The intervention resulted in significantly improved attentional capacity and to- tal attentional score over a 90-day period. There were no effects on mood. Bridge et al. randomized 154 women undergoing local radiotherapy for early stage breast cancer to one of three arms: (1) structured teaching of relaxation techniques including diaphragmatic breathing, supplemented by audiotapes; (2) the above plus imagery of a peaceful scene; or (3) a control arm in which individuals were seen individually and encouraged to talk about themselves (Bridge et al., 1988). The interventions took place weekly for 30 minutes for a total of 6 weeks. They were delivered by the research- ers, one of whom was a psychiatrist. Immediately post intervention, the relaxation with imagery intervention resulted in lower total mood distur- bance on the Profile of Mood States than relaxation alone, and both inter- ventions resulted in enhanced mood compared to the attention control arm.
112 MEETING PSYCHOSOCIAL NEEDS OF WOMEN WITH BREAST CANCER Similar findings were found for the individual item ârelaxed.â This study provides evidence of short-term benefits of relaxation, and it suggests that the addition of imagery to muscular relaxation further enhances psycho- logical benefits. Burton et al. randomized 200 women who were about to undergo mas- tectomy to an interview with a psychologist (a Present State Examination that focused on worries, concerns, or beliefs), or to the interview plus a half-hour individual psychotherapy session provided by a surgeon that dealt with the effects of breast cancer on the womanâs life situations and her feelings, or to the interview with a 30-minute chat with the surgeon that focused on patientsâ holidays and the like, or to a control arm (Burton et al., 1995). Psychological outcomes were obtained up to 1 year after the inter- vention. The preoperative interviews with the psychologist resulted in a lasting reduction in body image distress, reduced overall distress, anxiety, and depression, reduced upset regarding the loss of the breast as well as enhanced use of fighting spirit coping style. The psychotherapy versus the chat was superior among patients who had experienced stressful life events, but not in the overall study group. The control subjects in this study were not informed that they were participating in a study until the final 1-year measurement. Therefore, no baseline data were available on controls. Thus, results should be interpreted with caution, as the comparability of interven- tion and control groups at baseline was not demonstrated. It is notable that 80 women randomized to the intervention declined. There were no such decliners in the control group because the control subjects did not know that they were part of a study. This further increases the possibility that study groups were not comparable at baseline. Maunsell et al. randomized 259 women with newly diagnosed, local, or regional breast cancer to receive telephone screening every 28 days or to routine care (Maunsell et al., 1996). Telephone screening, performed by a research assistant, involved administration of the General Health Question- naire. If a score >5 was obtained, patients were referred to a social worker who then intervened by telephone to confirm distress, identified the cause for the distress, and offered support. Screening continued for 1 year. No significant effects of this brief screening intervention performed by a non- mental health professional were identified, despite the use of a fairly large number of outcome measures. Marchioro et al. randomized 36 women with newly diagnosed non- metastatic breast cancer to weekly, psychologist led individual cognitive psychotherapy combined with bi-monthly family counseling, or to standard care (Marchioro et al., 1996). The duration of the intervention was not described. Follow-up continued to 9 months. The intervention was reported to reduce depression and enhance quality of life. Changes in some aspects of personality were also identified. This study provides evidence of psycho-
EFFECTIVENESS OF PSYCHOSOCIAL INTERVENTIONS 113 logical benefit for an individual intervention; information on duration of the intervention would have been helpful. McArdle et al. randomized 272 women at the time of breast cancer diagnosis to receive either (1) support from a breast cancer nurse, (2) sup- port from a voluntary organization (Tak Tent) that provided individual counseling and support groups, (3) support from both the nurse and the voluntary organization, or (4) usual care (McArdle et al., 1996). Outcomes (general health, social function, and depression) were best in those random- ized to receive support from the nurse, providing clear evidence of benefit for support provided by a breast cancer nurse specialist; no benefit was seen in those randomized to the voluntary organization, and dropout rates were also higher among this latter group. In a small pilot study, Richardson et al. randomized 47 subjects to a non-structured support group that focused on stress reduction, reducing feelings of isolation, and enhancing self-esteem or to an imagery relaxation group that received instruction on relaxation, imagery, and basic breathing with supplemental discussion of stress and coping. In the latter arm, group meetings were supplemented by one individual relaxation/imagery session (Richardson et al., 1997). A third arm received standard care. The support group and relaxation imagery group sessions occurred weekly for 6 weeks, each lasting 1 hour. They were led by Masterâs level social workers; a hyp- notist participated in the relaxation imagery group. Although this was a small pilot study, there was evidence of enhanced coping skills in both the support and the relaxation/imagery groups (more significant in the former), and there was evidence that women in both intervention arms sought more support from others than women in the control arm. There was also a greater acceptance of death in women participating in the support groups. There was no evidence of an effect of either intervention on quality of life, mood, or immune parameters and no evidence of a differential effect of the support group versus the imagery relaxation group. Psychological outcomes were measured in the short-term only, the final evaluation was conducted immediately after the intervention was completed. No information was available on long-term effects. Samarel et al. randomized 228 women with recently treated stage I or II breast cancer to a support group, a support group with coaching (each woman was invited to bring a significant otherâa spouse, friend, or family member), or to a control arm (Samarel and Fawcett, 1992; Samarel et al., 1993, 1997). The support groups were led by a nurse and social worker and were structured and designed to assist women to adapt in physiological, self-concept, role function, and interdependent response modes. The inter- vention took place weekly for 8 weeks. Psychological outcomes were mea- sured up to 8 weeks after the end of the intervention. It was reported that the addition of coaching to the support groups resulted in higher quality
114 MEETING PSYCHOSOCIAL NEEDS OF WOMEN WITH BREAST CANCER relationships at the end of the intervention compared to the other arms of the study, but this effect did not persist 8 weeks later. The interventions had no effect on symptom distress, mood, or functional status, although an effect on symptom distress and mood might have been expected in view of the large size of the study and the use of psychological measures shown to be responsive to change in other studies (Profile of Mood States, Symptom Distress Scale). In this study, randomization was not straightforward since women randomized to one of the two intervention arms could be re-ran- domized to one of the two remaining arms if they were unable to attend the assigned group for any reason, but re-randomization was not possible for women allocated to the control arm. This may have led to non-comparabil- ity of the study groups and interference with detection of potential psycho- logical benefits. Kolcaba and Fox randomized 53 women to use a guided imagery au- diotape daily during radiation therapy and for three weeks later or to a no treatment control arm (Kolcaba and Fox, 1999). The audiotape included 20 minutes of verbal guided imagery focusing on comfort, as well as 20 minutes of soft jazz. Psychosocial follow-up continued until 3 weeks after radiation was completed. The intervention significantly improved comfort (psycho-spiritual, environmental, social) throughout the duration of the study. Walker et al. randomized 96 women with newly diagnosed locally ad- vanced breast cancer to relaxation training with guided imagery or to stan- dard care (Walker et al., 1999). The relaxation training included the use of audiocassettes and cartoon images of host defenses destroying cancer cells. The relaxation component focused on progressive muscular relaxation and cue controlled relaxation. The first 20 women in the study received âliveâ training, and the remaining women used the audiocassettes and cartoon images alone. Women were encouraged to use these materials daily during six chemotherapy cycles. Follow-up was to the end of the sixth chemo- therapy cycle. The intervention reduced overall emotional repression as well as unhappiness, and it enhanced global quality of life. It had no effect on mood. There was no effect of the intervention on tumor size or response to chemotherapy. This relatively simple intervention had a beneficial psycho- logical effect in women with locally advanced breast cancer, but no bio- medical effect was observed. WengstrÃ¶m et al. randomized 134 women to a nursing intervention based on Oremâs model for self-care versus standard care (WengstrÃ¶m et al., 1999, 2001). The intervention, based on individual patient contact, in- cluded education; support and guidance regarding self-care, psychological support, coping strategies, body image; and treatment. Patients received a total of five 30-minute sessions weekly for 5 weeks. They were followed throughout radiation and for 3 months post radiation. The intervention
EFFECTIVENESS OF PSYCHOSOCIAL INTERVENTIONS 115 group had fewer distress reactions as measured using the Impact of Events Scale, but there were no overall effects on quality of life or toxicity. In women over 59 years of age, the intervention was reported to result in âstronger motivation to be emotionally involved.â Sandgren et al. randomized 62 women to telephone-based individual cognitiveâbehavioral therapy or to assessment only (Sandgren et al., 2000). The telephone-based cognitiveâbehavioral therapy, delivered by Masterâs level clinical psychology students, provided support, addressed coping skills and strategies to manage anxiety and stress, and helped patients solve prob- lems. Cognitive restructuring and emotional expression were encouraged. Women randomized to the intervention received weekly interventions for 4 weeks, then every 2 weeks for 6 weeks. Each intervention averaged 20 to 25 minutes. Women were followed for a total of 10 months. There were no consistent effects of the intervention over time. Some borderline and tran- sient effects were seen; for example, there was an early reduction in stress followed by a late increase in stress, and early benefits for mental health but early deterioration in physical functioning. These inconsistent and border- line results may reflect the small sample size of this study or the lack of experience of the therapists. Further investigation in a larger group of pa- tients, or with more experienced therapists, might yield more stable results. Bultz et al., in a unique study, randomized 36 partners of breast cancer patients to a psycho-educational group led by two psychologists that took place weekly for 6 weeks or to a control arm (Bultz et al., 2000). Both patients and their partners were followed until 3 months after the comple- tion of the intervention. There were no statistically significant effects. Ritz et al. randomized 210 women who had been diagnosed with either invasive or non-invasive breast cancer within the preceding 2 weeks to receive advanced practice nursing interventions or to standard care (Ritz et al., 2000). The advanced practice nursing interventions included written and verbal in- formation about breast cancer. They addressed what to expect in consulta- tion with physicians; they assisted with decision-making, provided support, answered questions, and enhanced continuity of care. The intervention was provided in person, by telephone, and during home visits. The number of contacts per patient and the duration of these contacts were not specified. Patients were followed for a total of 24 months. The intervention was re- ported to reduce uncertainty at 1, 3, and 6 months with the effect being greatest in unmarried women. There were benefits in a number of uncertainty subscales including complexity, inconsistency, and unpredictability. There was no overall effect on mood, although a beneficial effect on mood was seen at 1 and 3 months in women without a family history of breast cancer. There were no effects on quality of life or on overall health-care costs. This nursing intervention, which in some centers may be considered the standard of care, reduced uncertainty, at least in the short term.
116 MEETING PSYCHOSOCIAL NEEDS OF WOMEN WITH BREAST CANCER Fukui et al. randomized 50 Japanese women to cognitiveâbehavioral group therapy or to a wait-list control arm (Fukui et al., 2000). Women had been diagnosed with high-risk invasive breast cancer 4 to 18 months earlier, and none had received chemotherapy. The cognitiveâbehavioral group therapy was led by a psychiatrist and a psychologist; it included education, coping skills training, stress management (including muscle relaxation and guided imagery), and psychosocial support. Sessions lasted 90 minutes and were held weekly for 6 weeks. The intervention significantly improved vigor and overall mood, and it significantly enhanced fighting spirit coping at the end of 6 weeks, but effects were marginal 6 months later. Additionally, there were no significant effects on depression or anxiety. This brief inter- vention appeared to have short-term, but not long-term, beneficial psycho- logical effects. Helgeson et al. randomized 312 women to four arms using a factorial design. The three treatment arms consisted of (1) education, (2) facilitated peer discussion, and (3) education and facilitated peer discussion. There was also a non-treatment control arm (Helgeson et al., 1999, 2000, 2001). The education intervention was group-based and involved a lecture as well as a question and answer session with an expert presenter and an oncology nurse or social worker. Interaction between group members was not en- couraged. The peer discussion was also in a group format with a nurse or social work facilitator who did not direct the conversation but encouraged women to focus on expression of feelings and self-disclosure. Both interven- tions lasted 60 minutes weekly for 8 weeks. Follow-up continued for 48 months. The education intervention resulted in a number of psychological benefits over both facilitated peer discussion and control. These benefits included enhanced vitality, social functioning, and mental health as well as reduced bodily pain. There were no early or late benefits for the facilitated peer discussion group, and the beneficial effects of the educational interven- tion dissipated over time. This study demonstrated early benefits for an educational intervention but provided no evidence of benefit for a facili- tated peer discussion group. This is the only randomized trial that included a peer discussion group (although professional facilitators were present). The findings at 48 months are harder to interpret because of attrition in the groups over time. Simpson et al. randomized 89 women with early stage breast cancer who had completed their treatment up to 2 years earlier to either a group intervention or to a control arm that received a book Helping YourselfâA Workbook for People Living with Cancer (Simpson et al., 2001). The inter- vention arm participated in structured group psychotherapy, 90 minutes weekly for 6 weeks, with weekly themes. A psychiatrist and two survivors led the groups, which included up to 10 subjects. Follow-up continued for 2 years. This brief intervention reduced depression, enhanced mood and qual-
EFFECTIVENESS OF PSYCHOSOCIAL INTERVENTIONS 117 ity of life, and diminished psychiatric symptoms at 2 years but not at 1 year. Furthermore, health care billing was reduced by $147 (Canadian dollars) in each intervention subject (a 23.5 percent reduction). These results suggest that a brief intervention may result in psychological benefits and reduction of health-care costs. Lev et al. randomized 53 women with stage I or II breast cancer who were receiving their first cycle of chemotherapy to an âefficacy enhancing interventionâ or to receive an educational booklet on cancer chemotherapy (Lev et al., 2001). The efficacy enhancing intervention included a 5-minute videotape, a self-care educational booklet, and five monthly efficacy en- hancing counseling interventions over 8 months. Patients were followed for a total of 8 months. No formal tests of statistical significance were per- formed. The authors reported effect sizes only: small to large effect sizes for functional or social concerns on the Functional Assessment of Cancer TherapyâB, as well as small to large effect sizes on the Symptom Distress Scale and a scale called âStrategies Used by Patients to Promote Health.â The lack of statistical significance testing and a dropout rate over 50 per- cent make interpretation of these results highly problematic. Antoni et al. randomized 136 women with stage 0 to II breast cancer who were within 8 weeks of surgery to a structured group intervention that included cognitiveâbehavioral and stress management therapy that met weekly for 2 hours for a total of 10 weeks or to a control arm (Antoni et al., 2001). The intervention also included didactic/experiential exercises as well as homework. Women randomized to the control arm participated in a 1- day seminar during which they received information relating to stress reac- tions and successful adjustment. The interventions and seminars were led by psychology fellows and doctoral students. Only the 100 women who completed all psychological assessments were included in the analysis. Thus, the larger group of women who were randomized was not fully described. The authors reported that there were no overall effects of the intervention, although they noted that it reduced the prevalence of moderate depression (but not mean depression scores) and that it increased benefit finding and optimism, the latter effect being maximal in women who had low baseline optimism scores. A large number of secondary analyses were presented. The selection of only the 100 women who completed all psychological as- sessments may have led to non-comparable study groups, and may have made it difficult to identify overall effects of the intervention similar to those found by other investigators. Molassiotis et al. randomized 71 women receiving their first cycle of chemotherapy for stage I to III breast cancer to a relaxation/imagery inter- vention or to a control arm (Molassiotis et al., 2002). The intervention involved progressive muscle relaxation training and guided imagery that was provided in six standardized sessions with a therapist (1 hour before
118 MEETING PSYCHOSOCIAL NEEDS OF WOMEN WITH BREAST CANCER chemotherapy and days 1 to 5 post chemotherapy). This intervention was supplemented with individual audiocassettes as well as a 30-minute video teaching program. Psychological assessments were obtained up to 14 days post chemotherapy. The intervention resulted in a significant decrease in total mood disturbance, as well as a significant decrease in the duration of nausea and vomiting and a trend towards lower frequency of nausea and vomiting; however, it had no effect on the intensity of nausea and vomiting. These findings suggest that this brief, simple intervention had important effects on symptom control and the psychological well-being of women receiving adjuvant chemotherapy. Allen et al. randomized 164 women under the age of 50 who were receiving adjuvant chemotherapy for stage I to III breast cancer to a brief intervention with an oncology nurse or to a control arm (Allen et al., 2002). The intervention consisted of two in-person and four telephone sessions over a 12-week period. The nurse provided problem-solving skills training as well as informational materials. Women were followed up to 8 months. There were no overall effects of the intervention in either univariate or multivariate analyses. Some subgroup effects were seen; in particular, women who had good baseline, inate problem-solving ability had a signifi- cant decrease in the number and severity of difficulties they experienced. A similar effect was not seen in women with poor baseline, inate problem- solving abilities. This suggests that a more focused and/or more intensive intervention may be necessary in women with poor baseline problem-solv- ing abilities, and it highlights the importance of targeting interventions to the needs and characteristics of specific populations. Finally, Targ and Levine (2002) recently reported an important study comparing a standard psychoeducational group to a mindâbodyâspirit group that included a number of complementary and alternative medicine (CAM) interventions, including dance, meditation, imagery, and ritual (Targ and Levine, 2002). The standard group with two leaders, one of whom was a psychologist, met for 90 minutes weekly for 12 weeks. The CAM group, led by nurses and social workers, met twice weekly for 150 minutes each time for 12 weeks. The CAM sessions included health discussions, medita- tion and guided imagery, experiential work, group discussion, and a move- ment/dance program. Despite that greater time commitment, dropouts were fewer and satisfaction was greater among women randomized to the CAM group. Psychosocial status including mood, coping, and HRQOL improved in both groups. The only significant between-group difference was in a novel measurement, spiritual integration, or the degree to which a person makes spirituality an integral part of their coping. Because the standard group had more dropouts than the CAM group even before that intervention began, it was suggested that the study may have attracted participants who were particularly interested in CAM interventions. Although the authors con-
EFFECTIVENESS OF PSYCHOSOCIAL INTERVENTIONS 119 cluded there was âequivalence on most psychosocial outcomes,â this study does describe the acceptability of complementary and alternative medicine interventions to women with breast cancer. Metastatic Breast Cancer There have been six randomized trials of psychosocial interventions in metastatic breast cancer (Arathuzik, 1994; Classen et al., 2001; Cunningham et al., 1998; Edelman et al., 1999a, 1999b; Edmonds et al., 1999; Goodwin et al., 2001; Spiegel et al., 1989, 1981; Spiegel and Bloom, 1983). Many of these trials have included both survival outcomes and psy- chosocial outcomes. As noted earlier, this interest in effects of psychosocial outcomes on survival reflects the publication by Spiegel et al. of an unex- pected survival benefit for a group psychosocial intervention in metastatic breast cancer (Spiegel et al., 1989). Spiegel et al. reported the first randomized trial of a psychosocial inter- vention in metastatic breast cancer (Spiegel et al., 1989, 1981; Spiegel and Bloom, 1983). He randomized 86 women to participate in weekly expres- siveâsupportive group therapy that included hypnosis for pain control or to a control arm. There was a high dropout rate pre-intervention (28 of 86 women dropped out before the intervention began, largely because of ill- ness). The group therapy sessions lasted for 90 minutes and were led by a psychiatrist and social worker. Women were asked to participate for 1 year, longer if possible. The intervention was reported to improve mood, to re- duce maladaptive coping responses, and to reduce phobias. Furthermore, in an unplanned survival analysis conducted after prolonged follow-up, the intervention was reported to significantly increase survivalâthe mean sur- vival being 36.6 months in the intervention arm and 18.9 months in the control arm. Although these results were exciting and provided the first evidence of psychological benefit for a psychological intervention in meta- static breast cancer, while also showing the feasibility of a psychological intervention in this seriously ill population of patients, the study was not designed to examine survival effects. Fox pointed out in a critique of the study that using SEER data for survival of women from metastatic breast cancer in that geographic area at that time resulted in the same survival curve for both the women in the intervention group and the women in the SEER database. In fact, for unknown reasons, the control group had a poorer survival. The apparent survival advantage appeared to be attribut- able to this fact (Fox, 1998). Replication of these results, particularly the results of the unplanned survival analysis, was recommended and is under- way by Spiegel and the subject of the Goodwin study described below. Arathuzik reported a very small randomized trial, essentially a pilot study, involving 24 women who were randomized to relaxation/visualiza-
120 MEETING PSYCHOSOCIAL NEEDS OF WOMEN WITH BREAST CANCER tion, relaxation/visualization plus cognitiveâbehavioral therapy, or to a con- trol arm (Arathuzik, 1994). The relaxation and visualization was a struc- tured intervention that involved individual sessions during which women were taught progressive relaxation and encouraged to perform visualiza- tion exercises. The cognitiveâbehavioral component was also delivered in- dividually, and it included discussion and a written handout describing 23 methods of pain distraction. Each intervention was administered once by the nurse investigator. There were no between-group differences in pain or mood. However, both treatment groups reported an interesting finding, an enhanced perceived ability to decrease pain. The lack of significant between- group effects on pain or mood might reflect the extremely small sample size in this study. Edelman et al. randomized 121 women with metastatic breast cancer to a group cognitiveâbehavioral therapy program or to a control arm (Edelman et al., 1999a, 1999b). The cognitiveâbehavioral intervention included a sup- port group that met weekly for 8 weeks, then monthly for 3 weeks, and involved one family session. There was a focus on cognitive restructuring, relaxation, communication and coping strategies, group interaction and support, relationships, and self-image. The group meetings were delivered by two therapists, one of whom was a psychologist, and were supplemented by a structured, manual-based set of homework exercises. Survival and psy- chosocial effects were examined. Patients were followed for at least 12 months post intervention; there was evidence of improved mood (depres- sion, total mood disturbance) and enhanced self-esteem at the end of therapy, but no benefits were present at 3 and 6 months. There was no evidence of survival benefits. Edmonds et al. and Cunningham et al. randomized 66 women to a group intervention that involved discussion, supportive strategies, and cognitiveâbehavioral therapy (Cunningham et al., 1998; Edmonds et al., 1999). The latter included 20 cognitiveâbehavioral assignments. In addi- tion, participants were asked to attend a coping skills weekend and to practice relaxation exercises. The group intervention lasted 35 weeks for 2 hours each week and was led by a psychologist and either a social worker or psychology doctoral student. The control arm received a home cognitiveâbehavioral therapy package which included a coping skills workbook, relaxation tapes as well as supportive phone calls at 2, 4, 5, 10, and 12 months. Patients were followed for 12 months. This interven- tion resulted in âlittleâ psychometric benefit compared to the control arm. Intervention subjects experienced increased anxious preoccupation coping and reduced helplessness coping. The control arm involved an important degree of intervention. This makes interpretation of the ab- sence of a psychological benefit from the group intervention difficult. No survival effects were seen in this study.
EFFECTIVENESS OF PSYCHOSOCIAL INTERVENTIONS 121 Classen et al. have reported psychological outcomes of a randomized trial attempting to replicate Spiegelâs earlier report of a survival benefit of supportiveâexpressive therapy in metastatic breast cancer (Classen et al., 2001). One hundred and twenty-five women with metastatic breast cancer were randomized to the same weekly supportiveâexpressive group therapy used in Spiegelâs original study, each session lasting 90 minutes. Women were encouraged to participate to the end of life. Sessions were led by psy- chiatrists, psychologists, and social workers. The control arm was offered educational materials. Psychological outcomes were reported; the interven- tion significantly reduced traumatic stress symptoms but had no significant effects on overall mood. However, if the final assessment performed during the last year of life was excluded, a significant improvement in mood was seen. Survival data have not yet been reported. Goodwin et al. reported the results of a multicenter randomized trial of group supportiveâexpressive therapy in 235 women with metastatic breast cancer (Goodwin et al., 2001). Women randomized to the intervention arm received supportiveâexpressive therapy following Spiegelâs manual-based description. The therapy fostered support and encouraged emotional ex- pressiveness and confronting effects of illness and change in self-image. Women were encouraged to examine roles, relationships, and the life alter- ing nature of their illness, and to enhance communication and coping. Women attended weekly 90-minute sessions led by psychiatrists, psycholo- gists, social workers, and nurse clinicians until death. The intervention was reported to enhance mood (total mood, anger, anxiety, depression, confu- sion) and to reduce pain. No survival effects were seen despite greater than 90 percent power to identify the survival effects that had been reported in 1989 by Spiegel et al. This study represents the largest attempt to replicate Spiegelâs survival resultsâthe successful delivery of the intervention is evi- denced by the psychological benefits that were seenâhowever, an impor- tant effect of the intervention on survival was ruled out. Healthy Survivors Only one trial has been reported in healthy breast cancer survivors (not included in Tables B-1 or B-2). In this trial, Fogarty et al. randomized 123 subjects to view a dramatized videotape of a treatment consultation that enhanced compassion, or a dramatized videotape of a treatment consulta- tion that did not enhance compassion; women who viewed the enhanced compassion videotapes reported reduced anxiety but also reduced informa- tion recall (Fogarty et al., 1999). There were also differences in perception of compassion and physician attributes in the expected direction. Women with no history of breast cancer participated as a second comparison group; results from these women were similar to those in breast cancer survivors.
122 MEETING PSYCHOSOCIAL NEEDS OF WOMEN WITH BREAST CANCER Summary of Effectiveness of Psychosocial Interventions According to Type of Intervention As was discussed in the preceding section, many of the published re- ports have identified psychological benefits of a variety of psychosocial in- terventions. These studies are reorganized in Appendix B, Table B-3 into three groups. The first group involves studies that primarily evaluated re- laxation and/or imagery, whether the approach was individual or group based. The second category of interventions involved therapy delivered in a group setting; the specific type of therapy varied across studies. It included cognitiveâbehavioral therapy, structured group therapy, supportiveâexpres- sive group therapy, education with or without peer discussion, and mindâ bodyâspirit groups. In the final section of Table B-3, studies of interven- tions in individuals are listed. These individual interventions were delivered in person or by telephone. Interventions included psychological assessment, telephone screening, cognitiveâbehavioral counseling, and several of these studies evaluated a variety of nursing interventions, focusing mainly on women with early stage breast cancer. Consideration was given to grouping studies by type of therapy evaluated (e.g., supportive, cognitiveâbehavioral). However, many of the interventions included components of more than one type of therapy (e.g., Cunningham), and some could not be easily classified into a specific therapeutic approach (e.g., telephone screening). Further- more, the classification of interventions into group or individual approaches reflects current interest in group approaches for women with breast cancer. Studies involving couples counseling or partners of breast cancer patients are not included in these tables (Bultz et al., 2000; Christensen, 1983). Studies evaluating relaxation with or without imagery appear in the first section of Table B-3. Studies in early breast cancer are listed first (Arathuzik, 1994; Bridge et al., 1988; Kolcaba and Fox, 1999; Molassiotis et al., 2002; Richardson et al., 1997; Walker et al., 1999). With a single exception (Arathuzik, 1994), all of these studies provided evidence of effec- tiveness of the intervention, including improvements in mood, relaxation, comfort, overall quality of life, acceptance of death, enhanced coping, as well as reduced nausea and vomiting. The effective studies were all con- ducted in early phase breast cancer, the majority during initial treatment. Most of these interventions were short-term (6 days to 6 months), and the majority involved the use of relaxation audiotapes. One study utilized a group intervention (that study is also listed in the group intervention sec- tion of Table B-3) (Richardson et al., 1997). Taken together, these studies provide evidence of the effectiveness of relaxation with or without imagery in relieving psychologic distress or side effects of treatment (including nau- sea) in the early phases of breast cancer, effects that were present during active treatment with radiation and/or chemotherapy. The failure of
EFFECTIVENESS OF PSYCHOSOCIAL INTERVENTIONS 123 Arathuzik to identify significant effects may reflect the fact that the inter- vention was delivered in one session, that the study involved only 24 women (the power may have been inadequate to identify significant effects), or that the study was conducted in the setting of metastatic breast cancer, a setting in which a longer intervention may be needed (Arathuzik, 1994). Because of these limitations, further research would be needed to clarify the effec- tiveness of relaxation as a sole modality or combined with brief (single session) cognitiveâbehavioral therapy in metastatic breast cancer. The second section of Table B-3 lists studies that evaluated a variety of group interventions. Studies of supportive therapy (including a study that compared supportive therapy to mindâbodyâspirit group therapy) are listed first, followed by those of supportiveâexpressive therapy, and then those that evaluated interventions having cognitiveâbehavioral therapy as a com- ponent (studies in early breast cancer are listed first in each grouping). These studies are roughly evenly split between early stage (seven studies) and meta- static (five studies) breast cancer. Four of these studies included cognitiveâ behavioral therapy either as the sole treatment modality, or as a major treatment modality. Three utilized a manual-based form of therapy: sup- portiveâexpressive therapy, combined with relaxation. One evaluated the benefits of therapist facilitated peer discussion groups, comparing them to education alone or to education combined with therapist-facilitated peer discussion, and one investigated a mindâbodyâspirit (CAM) group inter- vention. The interventions in early stage breast cancer tended to be shorter (6 to 12 weeks) than those conducted in metastatic breast cancer (5 months to end of life). Despite the differences among these studies, the majority demonstrated important psychosocial benefits, although some of the studies conducted in early stage disease evaluated psychological outcomes for only a brief period of time (6 weeks) (Richardson et al., 1997). The benefits reported in these studies included improved mood, enhanced coping, reduced phobias, re- duced traumatic stress symptoms, enhanced vitality, social and role func- tioning, reduced severity of psychiatric symptoms, enhanced quality of life, and enhanced spiritual integration. One study reported prolongation of sur- vival (this is discussed in greater detail in the next section) (Spiegel et al., 1989). In at least one study beneficial effects of the intervention (an educa- tional intervention) dissipated over time (Helgeson et al., 2001). In contrast to these largely positive findings, there were some studies that did not iden- tify significant overall effects. For example, Edmonds et al., studying 66 patients evaluating supportive plus cognitiveâbehavioral group therapy ver- sus home cognitiveâbehavioral study, reported âlittle psychometric effectsâ although there were some minor changes in coping style (Edmonds et al., 1999). The fact that the control group in this study also received interven- tions such as a home cognitiveâbehavioral study package and supportive
124 MEETING PSYCHOSOCIAL NEEDS OF WOMEN WITH BREAST CANCER telephone calls may have contributed to these negative results. Antoni et al. also failed to identify overall effects for a structured cognitiveâbehavioral group intervention, although they did find some beneficial effects in sub- groups (Antoni et al., 2001). In evaluating Antoniâs results, it must be noted that only the 73 percent of randomized women that completed all study assessments were included in the analysis; this could have introduced bias and made it difficult to identify important effects of the intervention. Two studies evaluated the benefits of facilitated peer discussion or the use of coaches in support groups. Samarel, evaluating the addition of coaches to a structured support group, found higher quality relationships in the short-term when coaches were incorporated into the groups; however, these effects did not last beyond 8 weeks (Samarel and Fawcett, 1992; Samarel et al., 1993, 1997). In this study, some patients were re-random- ized if their initial randomization resulted in an allocation that posed logis- tical difficulties. Because of this, these results must be interpreted with cau- tion. Helgeson is the only investigator to have evaluated a facilitated peer discussion group in women with breast cancer (Helgeson et al., 1999, 2000, 2001). No benefits of facilitated peer discussion were identified in this large study in which women were followed for up to 4 years after randomization. Firm conclusions about facilitated peer discussion groups will require more than a single study. Targ and Levine demonstrate comparability of psychoeducational group support to a complex CAM group support that addressed the mindâ bodyâspirit continuum for most key psychosocial and HRQOL outcomes (Targ and Levine, 2002). Enhanced satisfaction and spiritual integration were reported in those who participated in the CAM group. Overall, there was little evidence that the type of group therapy used led to different therapeutic effects; psychological benefits of various types were reported in cognitiveâbehavioral group interventions, supportiveâex- pressive group interventions, and other interventions. There have been no head-to-head comparisons of the two most commonly utilized types of sup- port groups: cognitiveâbehavioral and supportiveâexpressive. Such a com- parison might lead to an enhanced understanding of the relative benefits of each type of intervention. On the basis of evidence currently available, se- lection of one type of intervention could probably be based on therapist and/or patient preference. Finally, individual interventions have been listed in the third section of Table B-3. All of these were conducted in early breast cancer. The interven- tions evaluated in these studies were quite variable, including psychological interviews, telephone screening, individual cognitiveâbehavioral therapy, and problem skills training as well as a variety of nursing interventions. These nursing interventions yielded somewhat mixed results. In a very early study, Maguire et al. showed that nurse specialist counseling increased rec-
EFFECTIVENESS OF PSYCHOSOCIAL INTERVENTIONS 125 ognition of the need for psychiatric referral and that subsequent psychiatric intervention reduced psychiatric morbidity (Maguire et al., 1980, 1983). McArdle et al. identified important psychosocial and quality of life benefits for support from a breast cancer nurse compared to support from a volun- tary organization of usual care (McArdle et al., 1996). Similarly, WengstrÃ¶m et al. reported fewer distress reactions and a stronger motivation to be emo- tionally involved in women who received a nursing intervention based on Oremâs model for self-care (Walker et al., 1999; WengstrÃ¶m et al., 1999, 2001). Ritz et al., evaluating advanced practice nursing interventions, iden- tified reduced uncertainty, as well as beneficial subgroup effects on mood in those without a family history of breast cancer (Ritz et al., 2000). Cimprich reported beneficial effects of an unusual intervention designed to restore attentional capacity (Cimprich, 1993). In contrast, Sandgren et al. reported no overall beneficial effects of telephone-based cognitiveâbehavioral therapy (Sandgren et al., 2000). Similarly, Allen et al., evaluating problem-solving skills training (two sessions in person, four by telephone), identified no overall effects, although subgroup analysis suggested that there may be some benefit in women with good problem-solving skills at baseline (Allen et al., 2002). Finally, Lev et al., evaluating nurse counseling supplemented with videotape and a self-care booklet, reported âsmall to large effect sizesâ for quality of life outcomes and psychiatric symptoms but did not perform any significance testing in their small study (Lev et al., 2001). A single study of individual cognitiveâbehavioral counseling identified improved depression and quality of life in women receiving the interven- tion, similar to the findings reported for group cognitiveâbehavioral therapy (Marchioro et al., 1996). Similar effects were noted for a brief (one session) psychological interview, with or without a subsequent brief psychotherapy session (Burton et al., 1995). Maunsellâs study of telephone screening for psychologic distress with referral of distressed patients to a social worker did not lead to any identifiable psychological benefits (Maunsell et al., 1996). Use of a health-care professional to provide screen- ing (as opposed to a research assistant) and closer coordination of referrals to a therapist (rather than forwarding the patientâs name to a social worker who attempted to contact the patient by telephone) may have led to better screening and counseling and different results. Further research in this area is needed. Taken together, these studies of individual interventions have identified some important psychological benefits; however, their results are not as consistent as those of relaxation/imagery interventions or group interventions. It is probably too early to make comparisons across types of interven- tions, particularly since there have been no head-to-head comparisons of individual versus group interventions. It is notable that trials of individual interventions have been conducted in early stage breast cancer only, that
126 MEETING PSYCHOSOCIAL NEEDS OF WOMEN WITH BREAST CANCER some of the interventions have been extremely brief, and that the variety of the interventions studied has been much broader than the variety of inter- ventions used in the group setting. Furthermore, although many reports in the metastatic setting in particular have examined long-term benefits, some, especially those in early stage breast cancer, have examined extremely short- term benefits (Arathuzik, 1994; Goodwin et al., 2001; Molassiotis et al., 2002; Richardson et al., 1997; Spiegel et al., 1989; Walker et al., 1999). In general, there is fairly good support for a short to intermediate term benefit for many group and individual interventions in early breast cancer (up to six months). However, the long-term benefits of interventions (apart from group interventions in the metastatic setting) have not been adequately evaluated. Review of Selected Studies Using Non-Randomized Designs There is an extensive literature describing a variety of psychosocial in- terventions in cancer using non-randomized designs. Some of these studies focus on breast cancer patients. Others, including some that used random- ized designs, involve patients with many different types of cancer. These studies are too numerous to review in detail; a small selection that high- lights key points will be reviewed here. Devine and Westlake have published an excellent meta-analysis2 of over 100 studies of psycho-educational care in cancer patients (Devine and Westlake, 1995). These studies were conducted in a variety of settings, in patients with a variety of cancers, and they evaluated a range of interven- tions. The majority of the studies reviewed were conducted in the United States, and the most common setting for the intervention was the outpatient treatment department. Some of the interventions involved education only; others included muscle relaxation, guided imagery, and other behavioral interventions. Statistically significant benefits were found for anxiety, de- pression, overall mood, nausea, vomiting, pain, and knowledge in the meta- analysis. It was concluded that psycho-educational care was of benefit to adults with cancer, but that it was difficult to differentiate the effectiveness of various types of psycho-educational care. The results of this meta-analysis are consistent with the findings of Helgeson et al. reported above in which a psycho-educational group inter- vention administered to women with breast cancer was found to be effec- tive for a variety of psychosocial outcomes (Helgeson et al., 1999, 2000, 2001). They are also consistent with a recent non-randomized study by Sepucha et al. in which breast cancer patients were assigned either to a 2The term âmeta-analysisâ refers to the statistical analysis of the results from many indi- vidual studies for the purpose of integrating the findings.
EFFECTIVENESS OF PSYCHOSOCIAL INTERVENTIONS 127 control intervention (productive listening) in which a researcher listened to and prompted patients to reflect on their experiences communicating with physicians or to an experimental intervention (consultation planning) in which a researcher actively assisted patients in preparing for an upcoming consultation, generating a printed agenda, and developing techniques to improve communication with their physicians (Sepucha et al., 2002). Pa- tients who received the experimental intervention reported significantly higher satisfaction, a sentiment that was echoed by the patientsâ physicians, suggesting that psycho-educational interventions and/or interventions that prepare patients for consultations with their physicians in a structured fash- ion may result in psychological benefit. A series of studies conducted by Hosaka et al. in Japan evaluated a structured intervention that included psycho-education, problem-solving, psychological support, relaxation training, and guided imagery (Hosaka et al., 2001a, 2001b, 2000a, 2000b, 2000c). Using non-randomized before- after designs, these investigators first demonstrated psychological benefits of a 5 week intervention in women with node-negative breast cancer who did not have adjustment disorders but not in those with positive nodes or adjustment disorders. They later added three sessions at two monthly inter- vals and were able to show persistent benefit in all women regardless of nodal status or presence of an adjustment disorder. Additionally, they iden- tified a series of psychological attributes predictive of improvement (good relationship with doctors, family support/understanding, no co-morbid adjustment disorders). Although these results should be confirmed in a ran- domized trial, this pilot work suggested characteristics of women who needed a longer intervention and the efficacy of a longer intervention. There are a number of interventions of potential psychosocial value that have not undergone adequate scrutiny in a randomized trial. Some patients are interested in these interventions, and some are used with in- creasingly frequency; for example, Internet chat groups, art therapy, music therapy, yoga, reflexology, and tâai chi. Evidence supporting the benefits of these interventions is lacking. Targ and Levine (2002) evaluated a multi- factorial mindâbodyâspirit group intervention and suggested that these al- ternative therapy interventions will be well received by women with breast cancer and that they may provide psychosocial benefit similar to more tra- ditional psychosocial interventions. Haun and colleagues reported results of a non-randomized comparison of women awaiting breast biopsy who were exposed to either a 20 minute music-based intervention or to usual care (Haun et al., 2001). The music intervention involved a selection of ânew age musicâ from which participants chose a 20-minute segment which they listened to via headphones. Women who received the intervention had a lower level of anxiety (measured using the State Trait Anxiety Inventory) and a lower respiratory rate. Similar information regarding other comple-
128 MEETING PSYCHOSOCIAL NEEDS OF WOMEN WITH BREAST CANCER mentary interventions is likely to be forthcoming. Formal evaluation of the benefit, both in the long-term and short-term, of some of these interven- tions will be of interest. Ashbury et al. evaluated a Canadian Cancer Society post-mastectomy peer visitor program, Reach to Recovery (Ashbury et al., 1998). They re- ported that patients were generally satisfied with the program and presented evidence of enhanced quality of life for participants. This non-randomized study suggests a beneficial effect of peer support shortly after breast cancer diagnosis. Finally, there is a growing number of reports of benefit from a range of psychological interventions in breast cancer. In this literature, either the intervention or the subject population is novel. For example, Spiegel et al. have evaluated supportiveâexpressive group therapy in early stage breast cancer using a pre-, post-test design; they indicate that mood is enhanced, anxiety and depression are reduced, and traumatic stress symp- toms reduced over a 6 month period (Spiegel et al., 1999). Because the study design was pre- and post-test and not randomized, it cannot be concluded that these benefits arose from the supportiveâexpressive group therapy intervention and not from spontaneous improvement as women adapted to their diagnosis. Similar benefits were seen in the metastatic setting, suggesting that at least some of these benefits might have arisen from the group intervention. Donnelly et al. carried out a pilot study of interpersonal psychotherapy administered by telephone to cancer patients receiving high dose chemotherapy and their partners (Donnelly et al., 2000). They reported the feasibility of the intervention and that partici- pants were satisfied with the intervention. Further exploration of this ap- proach could be useful for a number of reasons. It uses a therapeutic approach that has not been evaluated in a randomized trial in breast can- cer patients (interpersonal psychotherapy), the intervention is telephone- based (some of the randomized trials of telephone-based interventions have not been successful in breast cancer patients), and it intervenes not only with breast cancer patients but also with partnersâa group that has not received as much attention in the randomized trials setting. This brief, and selective, presentation of findings outside of the clini- cal trials setting is not intended to be comprehensive. Instead, it is in- tended to demonstrate that there is currently a broad spectrum of interest- ing and potentially useful research that is looking at a spectrum of novel interventions. Additional research is recommended to address research gaps, to evaluate the generalizability of research findings across gender and cancer type, and to examine the translation of research findings into community practice. The Boardâs priorities for future research are de- scribed in Chapter 8.
EFFECTIVENESS OF PSYCHOSOCIAL INTERVENTIONS 129 SUMMARY The results of 31 randomized trials in women with breast cancer and two meta-analyses, supported by observations in non-randomized studies, provide evidence of beneficial effects of a range of psychosocial interven- tions in both early and metastatic breast cancer. Notably, there is evidence for the benefit of relaxation/hypnosis/imagery interventions in early stage breast cancer, for group interventions in both early and metastatic breast cancer, and for individual interventions, primarily in the early setting. Al- though it needs strengthening, this evidence supports the conclusion that psychosocial interventions can be expected to reduce psychiatric symptoms and improve quality of life in routine clinical care of breast cancer. REFERENCES Agency for Healthcare Research and Quality. 2002. Management of cancer symptoms: Pain, depression, and fatigue: Summary. Evidence Report/Technology Assessment (61):1â9. Allen SM, Shah AC, Nezu AM, et al. 2002. A problem-solving approach to stress reduction among younger women with breast carcinoma: A randomized controlled trial. Cancer 94(12):3089â3100. Antoni MH, Lehman JM, Kilbourn KM, et al. 2001. Cognitiveâbehavioral stress management intervention decreases the prevalence of depression and enhances benefit finding among women under treatment for early-stage breast cancer. Health Psychol 20(1):20â32. Arathuzik D. 1994. Effects of cognitiveâbehavioral strategies on pain in cancer patients. Can- cer Nurs 17(3):207â214. Ashbury FD, Cameron C, Mercer SL, Fitch M, Nielsen E. 1998. One-on-one peer support and quality of life for breast cancer patients. Patient Educ Couns 35(2):89â100. Bridge LR, Benson P, Pietroni PC, Priest RG. 1988. Relaxation and imagery in the treatment of breast cancer. Br Med J 297(6657):1169â1172. Bultz BD, Speca M, Brasher PM, Geggie PH, Page SA. 2000. A randomized controlled trial of a brief psychoeducational support group for partners of early stage breast cancer pa- tients. Psycho-Oncology 9(4):303â313. Burke S, Kissane DW. 1998. Psychosocial impact of breast cancer: A review of interventions by specialist providers: A summary of the literature 1976â1996. NHMRC National Breast Cancer Center, NSW, Australia. [Online]. Available http://www.nbcc.org.au/pages/info/ resource/nbccpubs/psych/interv/contents.htm [accessed October 23, 2003]. Burton MV, Parker RW, Farrell A, Bailey D, Conneely J, Booth S, Elcombe S. 1995. A ran- domized controlled trial of preoperative psychological preparation for mastectomy. Psycho-Oncology 14:1â19. Christensen DL. 1983. Postmastectomy couple counseling: An outcome study of a structured treatment protocol. J Sex and Marital Ther 9:266â275. Cimprich B. 1993. Development of an intervention to restore attention in cancer patients. Cancer Nurs 16(2):83â92. Classen C, Butler LD, Koopman C, et al. 2001. Supportiveâexpressive group therapy and distress in patients with metastatic breast cancer: A randomized clinical intervention trial. Arch Gen Psychiatry 58(5):494â501. Cunningham AJ, Edmonds CV, Jenkins GP, Pollack H, Lockwood GA, Warr D. 1998. A randomized controlled trial of the effects of group psychological therapy on survival in women with metastatic breast cancer. Psycho-Oncology 7(6):508â517.
130 MEETING PSYCHOSOCIAL NEEDS OF WOMEN WITH BREAST CANCER Devine EC, Westlake SK. 1995. The effects of psychoeducational care provided to adults with cancer: Meta-analysis of 116 studies. Oncol Nurs Forum 22(9):1369â1381. Donnelly JM, Kornblith AB, Fleishman S, et al. 2000. A pilot study of interpersonal psycho- therapy by telephone with cancer patients and their partners. Psycho-Oncology 9(1):44â56. Edelman S, Bell DR, Kidman AD. 1999a. A group cognitive behaviour therapy programme with metastatic breast cancer patients. Psycho-Oncology 8(4):295â305. Edelman S, Lemon J, Bell DR, Kidman AD. 1999b. Effects of group CBT on the survival time of patients with metastatic breast cancer. Psycho-Oncology 8(6):474â481. Edmonds CV, Lockwood GA, Cunningham AJ. 1999. Psychological response to long-term group therapy: A randomized trial with metastatic breast cancer patients. Psycho- Oncology 8(1):74â91. Fawzy FI, Fawzy NW, Arndt LA, Pasnau RO. 1995. Critical review of psychosocial interven- tions in cancer care. Arch Gen Psychiatry 52(2):100â113. Fawzy FI, Fawzy NW, Hyun CS, et al. 1993. Malignant melanoma. Effects of an early struc- tured psychiatric intervention, coping, and affective state on recurrence and survival 6 years later. Arch Gen Psychiatry 50(9):681â689. Ferlic M, Goldman A, Kennedy JJ. 1979. Group counseling in adult patients. Cancer 43:760â 766. Fogarty LA, Curbow BA, Wingard JR, McDonnell K, Somerfield MR. 1999. Can 40 seconds of compassion reduce patient anxiety? J Clin Oncol 17(1):371â379. Fox BH. 1998. A hypothesis about Spiegel et al.âs 1989 paper on psychosocial intervention and breast cancer survival. Psycho-Oncology 7(5):361â370. Fukui S, Kugaya A, Okamura H, et al. 2000. A psychosocial group intervention for Japanese women with primary breast carcinoma. Cancer 89(5):1026â1036. Ganz PA, Goodwin PJ. 2003. Quality of life in breast cancer. What have we learned and where do we go from here? In: Lipscomb J, Gotay C, Snyder C, eds. Outcome Assessments in Cancer. Cambridge: Cambridge University Press. Ganz PA, Greendale GA, Petersen L, Zibecchi L, Kahn B, Belin TR. 2000. Managing meno- pausal symptoms in breast cancer survivors: Results of a randomized controlled trial. J Natl Cancer Inst 92(13):1054â1064. Goodwin PJ, Leszcz M, Ennis M, et al. 2001. The effect of group psychosocial support on survival in metastatic breast cancer. N Engl J Med 345(24):1719â1726. Haun M, Mainous RO, Looney SW. 2001. Effect of music on anxiety of women awaiting breast biopsy. Behav Med 27(3):127â132. Heinrich RL, Schag CC. 1985. Stress and activity management: Group treatment for cancer patients and spouses. J Counsel Clin Psychol 53:439â446. Helgeson VS, Cohen S, Schulz R, Yasko J. 1999. Education and peer discussion group inter- ventions and adjustment to breast cancer. Arch Gen Psychiatry 56(4):340â347. Helgeson VS, Cohen S, Schulz R, Yasko J. 2000. Group support interventions for women with breast cancer: Who benefits from what? Health Psychol 19(2):107â114. Helgeson VS, Cohen S, Schulz R, Yasko J. 2001. Long-term effects of educational and peer discussion group interventions on adjustment to breast cancer. Health Psychol 20(5):387â 392. Hosaka T, Sugiyama Y, Hirai K, Okuyama T, Sugawara Y, Nakamura Y. 2001a. Effects of a modified group intervention with early-stage breast cancer patients. Gen Hosp Psychia- try 23(3):145â151. Hosaka T, Sugiyama Y, Hirai K, Sugawara Y. 2001b. Factors associated with the effects of a structured psychiatric intervention on breast cancer patients. Tokai J Exp Clin Med 26(2):33â38.
EFFECTIVENESS OF PSYCHOSOCIAL INTERVENTIONS 131 Hosaka T, Sugiyama Y, Tokuda Y, Okuyama T. 2000a. Persistent effects of a structured psychiatric intervention on breast cancer patientsâ emotions. Psychiatry Clin Neurosci 54(5):559â563. Hosaka T, Sugiyama Y, Tokuda Y, Okuyama T, Sugawara Y, Nakamura Y. 2000b. Persis- tence of the benefits of a structured psychiatric intervention for breast cancer patients with lymph node metastases. Tokai J Exp Clin Med 25(2):45â49. Hosaka T, Tokuda Y, Sugiyama Y, Hirai K, Okuyama T. 2000c. Effects of a structured psy- chiatric intervention on immune function of cancer patients. Tokai J Exp Clin Med 25(4â 6):183â188. Ilnyckyj A, Farber J, Cheang MC WBH. 1994. A randomized controlled trial of psychothera- peutic intervention in cancer patients. Ann CRMCC 27:2:93â96. Kissane D, Bloch S, Smith GC, et al. 2003. Cognitiveâexistential group psychotherapy for women with primary breast cancer: A randomized controlled trial. Psycho-Oncology 12(6):532â546. Kolcaba K, Fox C. 1999. The effects of guided imagery on comfort of women with early stage breast cancer undergoing radiation therapy. Oncol Nurs Forum 26(1):67â72. Koocher, GP, Pollin, I. 2001. Preventive Psychosocial Intervention in Cancer Treatment: Im- plications for Managed Care, in Psychosocial Interventions for Cancer, Baum, A and Anderson, BL. eds. Washington, DC: American Psychological Association. Kuchler T, Henne-Bruns D, Rappat S, et al. 1999. Impact of psychotherapeutic support on gastrointestinal cancer patients undergoing surgery: Survival results of a trial. Hepatogastroenterology 46(25):322â335. Lev EL, Daley KM, Conner NE, Reith M, Fernandez C, Owen SV. 2001. An intervention to increase quality of life and self-care self-efficacy and decrease symptoms in breast cancer patients. Sch Inq Nurs Pract 15(3):277â294. Linn MW, Linn BS, Harris R. 1982. Effects of counseling for late stage cancer patients. Cancer 49(5):1048â1055. Maguire P, Brooke M, Tait A, Thomas C, Sellwood R. 1983. The effect of counselling on physical disability and social recovery after mastectomy. Clin Oncol 9(4):319â324. Maguire P, Tait A, Brooke M, Thomas C, Sellwood R. 1980. Effect of counselling on the psychiatric morbidity associated with mastectomy. Br Med J 281(6253):1454â1456. Marchioro G, Azzarello G, Checchin F, Perale M, Segati R, Sampognaro E, Rosetti F, Franchin A, Pappagallo GL, Vinante O. 1996. The impact of a psychological intervention on qual- ity of life in non-metastatic breast cancer. Eur J Cancer 32A(9):1612â1615. Maunsell E, Brisson J, Deschenes L, Frasure-Smith N. 1996. Randomized trial of a psycho- logic distress screening program after breast cancer: Effects on quality of life. J Clin Oncol 14(10):2747â2755. McArdle JM, George WD, McArdle CS, Smith DC, Moodie AR, Hughson AV, Murray GD. 1996. Psychological support for patients undergoing breast cancer surgery: A randomised study. Br Med J 312(7034):813â816. Meyer TJ, Mark MM. 1995. Effects of psychosocial interventions with adult cancer patients: A meta-analysis of randomized experiments. Health Psychol 14(2):101â108. Molassiotis A, Yung HP, Yam BM, Chan FY, Mok TS. 2002. The effectiveness of progressive muscle relaxation training in managing chemotherapy-induced nausea and vomiting in Chinese breast cancer patients: A randomised controlled trial. Support Care Cancer 10(3):237â246. Newell SA, Sanson-Fisher RW, Savolainen NJ. 2002. Systematic review of psychological thera- pies for cancer patients: Overview and recommendations for future research. J Natl Can- cer Inst 94(8):558â584.
132 MEETING PSYCHOSOCIAL NEEDS OF WOMEN WITH BREAST CANCER Redman S, Turner J, Davis C. 2003. Improving supportive care for women with breast cancer in Australia: The challenge of modifying health systems. Psycho-Oncology 12(56):521â 531. Richardson MA, Post-White J, Grimm EA, Moye LA, Singletary SE, Justice B. 1997. Coping, life attitudes, and immune responses to imagery and group support after breast cancer treatment. Altern Ther Health Med 3(5):62â70. Richardson JL Shelton DR, Krailo M, et al. 1990. The effect of compliance with treatment on survival among patients with hematologic malignancies. J Clin Oncol 8(2):356â364. Rimer B, Keintz MK, Glassman B. 1985. Cancer patient education: Reality and potential. Prev Med 14(6):801â818. Ritz LJ, Nissen MJ, Swenson KK, et al. 2000. Effects of advanced nursing care on quality of life and cost outcomes of women diagnosed with breast cancer. Oncol Nurs Forum 27(6):923â932. Samarel N, Fawcett J. 1992. Enhancing adaptation to breast cancer: The addition of coaching to support groups. Oncol Nurs Forum 19(4):591â596. Samarel N, Fawcett J, Tulman L. 1993. The effects of coaching in breast cancer support groups: A pilot study. Oncol Nurs Forum 20(5):795â798. Samarel N, Fawcett J, Tulman L. 1997. Effect of support groups with coaching on adaptation to early stage breast cancer. Res Nurs Health 20(1):15â26. Sandgren AK, McCaul KD, King B, OâDonnell S, Foreman G. 2000. Telephone therapy for patients with breast cancer. Oncol Nurs Forum 27(4):683â688. Sepucha KR, Belkora JK, Mutchnick S, Esserman LJ. 2002. Consultation planning to help breast cancer patients prepare for medical consultations: Effect on communication and satisfaction for patients and physicians. J Clin Oncol 20(11):2695â2700. Simpson JS, Carlson LE, Trew ME. 2001. Effect of group therapy for breast cancer on healthcare utilization. Cancer Pract 9(1):19â26. Spiegel D, Bloom JR. 1983. Group therapy and hypnosis reduce metastatic breast carcinoma pain. Psychosom Med 45(4):333â339. Spiegel D, Bloom JR, Kraemer HC, Gottheil E. 1989. Effect of psychosocial treatment on survival of patients with metastatic breast cancer. Lancet 2(8668):888â891. Spiegel D, Bloom JR, Yalom I. 1981. Group support for patients with metastatic cancer. A randomized outcome study. Arch Gen Psychiatry 38(5):527â533. Spiegel D, Morrow GR, Classen C, et al. 1999. Group psychotherapy for recently diagnosed breast cancer patients: A multicenter feasibility study. Psycho-Oncology 8(6):482â493. Targ EF, Levine EG. 2002. The efficacy of a mindâbodyâspirit group for women with breast cancer: A randomized controlled trial. Gen Hosp Psychiatry 24(4):238â248. Walker LG, Walker MB, Ogston K, et al. 1999. Psychological, clinical and pathological effects of relaxation training and guided imagery during primary chemotherapy. Br J Cancer 80(1â2):262â268. Wallace KG. 1997. Analysis of recent literature concerning relaxation and imagery interven- tions for cancer pain. Cancer Nurs 20(2):79â87. WengstrÃ¶m Y, Haggmark C, Forsberg C. 2001. Coping with radiation therapy: Effects of a nursing intervention on coping ability for women with breast cancer. Int J Nurs Pract 7(1):8â15. WengstrÃ¶m Y, Haggmark C, Strander H, Forsberg C. 1999. Effects of a nursing intervention on subjective distress, side effects and quality of life of breast cancer patients receiving curative radiation therapyâa randomized study. Acta Oncol 38(6):763â770.