Objective: This systematic review aimed to determine the effectiveness of psychoeducation, cognitive behavioural therapy (CBT) and social support interventions used in the rehabilitation of breast cancer (BC) patients.
Methods: We conducted a systematic literature search to identify randomised controlled trials of female BC patients who underwent different psychosocial interventions during or after primary cancer treatment. The methodological quality of all studies was independently assessed by two reviewers. Studies with low quality, less than 20 participants in each group, patients with metastatic cancer, data not presented separately for BC and studies that included other cancer types were excluded.
Results: Among 9617 identified studies, only 18 RCTs published between 1999 and 2008, including 3272 patients were finally included in this systematic evaluation. Outcome measures were categorised into quality of life (QoL), fatigue, mood, health behaviour and social function. Six trials examined psychoeducation had inconsistent results, both during and after the primary treatment. Seven trials examined the effect of CBT, four of which given after primary treatment (range 6–12 weeks) demonstrated improvements in QoL; the other three CBT studies given during primary treatment (range 9–20 weeks) had inconsistencies. Five studies addressed social support and showed no conclusive impacts of this intervention.
Breast cancer (BC) is the most prevalent cancer in women, including 1.15 million new invasive BC cases worldwide 1. Modern tailored treatment regimens have resulted in longer treatment periods, lasting up to 6 years after the diagnosis of BC, which may lead to more adverse side effects, including distress and pain 2. The increases in BC incidence and improved survival have given an increase in number of long-term BC survivors, and more long-term side effects due to both disease and treatment.
Clinical outcomes seem to differ among patients with apparently similar tumour stage at time of presentation. Whether psychosocial factors affect survival is unclear 3–6, but there is an urgent need to find adequate rehabilitation methods to better manage the sequelae of BC patients. Optimal recovery and prevention of treatment complications are main goals of rehabilitation.
Approximately 50% of all BC patients experience emotional distress, and 25% substantial psychological distress 7, 8. Distress symptoms appear to be mediated through thoughts about illness and prognosis, with cognitive intrusions by thoughts of recurrence, death, and disability 8, 9. These factors have been the rationale for the use of psychosocial interventions in the rehabilitation of BC patients, but whether these interventions actually facilitate the rehabilitation process is unclear 3–5.
The most frequent interventions are psychoeducation 10, cognitive behavioural therapy (CBT) 8, and social and emotional support 11. A previous systematic review 12 summarised current meta-analysis literature on treatment outcomes for CBT in a wide range of disorders. Large effects were found in unipolar depression, generalised anxiety disorder, panic disorder with or without agoraphobia, social phobia, and post-traumatic stress disorder. The effects of CBT on marital distress, anger, and chronic pain were moderate.
Many symptoms above are applicable to BC patients, and using CBT might be useful. However, few systematic reviews have evaluated the use of CBT in cancer patients in general and BC patients in particular. A recent meta-analysis examined distress and pain in BC patients, and supported individual CBT use 8. Moreover, studies addressing the usefulness of psychosocial rehabilitation measures in BC patients have been inconclusive 3–5, 13.
The main aim of the present study was therefore to conduct a systematic review for evaluating the present knowledge regarding the effect of psychosocial interventions in BC rehabilitation, and if possible, point to future important research priorities.
Based on the Norwegian Knowledge Centre for the Health Services (NOKC) handbook guidelines for systematic reviews 14, systematic electronic search of Cochrane Library, The Centre for Reviews and Dissemination databases, Medline, Embase, Cinahl, PsycINFO, AMED, and PEDro was performed, using the search strategy published earlier 15. Only randomised controlled trials (RCTs) studying the effect of psychosocial interventions in relation to BC rehabilitation with ⩾20 female BC patients after finishing surgery and adjuvant treatment in each group were included. Studies with metastatic BC were excluded.
Psychosocial interventions were divided into three categories: Psychoeducation, CBT, and social and emotional support. Outcome measures were categorised as: (i) quality of life (QoL); (ii) mood, including anxiety, depression, and cancer-related stress; (iii) clinical symptoms, including fatigue, nausea, dizziness; (iv) health behaviour, including coping, work disability and leisure time activities; and (v) social functioning, including spouse relations and friends.
The working group independently selected the trials for the systematic review according to given national and international guidelines (NOKC) 14, using the study abstract to assess the eligibility based on the inclusion criteria regarding design, subjects, and intervention.
The methodological quality of all studies was independently assessed by two reviewers 14. Discrepancies between the reviewers were resolved by a third independent reviewer, and disagreements were discussed in consensus meetings. The checklist consists of 11 questions based on the risk of bias assessments according to Cochrane handbook (see Appendix A).
Data collection and analysis
To measure clinical symptoms, health behaviour, and social functioning, a variety of different measurements were used as mean scores between the intervention and the control group, or mean difference (SD), as well as post-scores and follow-up results if available. Some of the most common measurement scales have been reviewed by Mandelblatt 16. The outcome effects between the groups were subjected to a two-way analysis of variance (ANOVA) and the indicated F-scores were converted post-hoc to standardised mean differences with confidence intervals. With several follow-up measurements, we reported the first measurement after the interventions, and reported to which extent the differences were sustained in the following measurements. Owing to the heterogeneity of the data statistical pooling for meta-analysis could not be performed. Results were summarised in narrative form.
A total of 9617 abstracts were identified, but only 191 were retrieved for further assessment, and 137 of these were excluded due to poor study quality or a failure to fulfil the inclusion criteria 15. Of the remaining 54 publications, 18 RCTs (22 publications) between 1999 and 2008, including 3272 patients receiving psychosocial interventions with moderate to high study quality were included in the analysis (Figure 1). Only three publications were characterised as high quality 17–20 (Appendix A).
In measuring QoL, 3 of 18 RCTs used FACT-B (17%), 3 used EORTC-QLQ-C30 (17%), 1 study used SF-36 (5.5%), 1 used QLI (5.5%), and 1 used Euro-QoL-5D (5.5%). Measuring mood, 8 of 18 studies used POMS (44%), and 6 used HADS (33%). Two studies used both POMS and HADS 21, 22. To measure clinical symptoms, health behaviour, and social functioning a variety of different measurements were used (Tables 1–3; list of abbreviations in Appendix B).
Table 1. Results of randomised control trials on the efficacy of psychoeducation (a) during primary breast cancer treatment (b) after primary breast cancer treatment
Table 2. Results of randomised control trials on the efficacy of cognitive behaviour therapy (CBT) (a) during primary breast cancer treatment (b) after primary breast cancer treatment
Table 3. Results of randomised controlled trials on the efficacy of social and emotional support (a) during primary breast cancer treatment (b) interventions after primary breast cancer treatment
Six RCTs (seven publications) (n = 1318) examined the effect of ‘psychoeducational information’, which was provided by computers, internet, videotapes, trained cancer educators or in combination. Three RCTs (four publications) (n = 437) examined the effect of the interventions when provided during primary BC treatment (range 8 weeks–13 months); no treatment effects (Table 1a). Three RCTs (n = 881) examined the effect of psychosocial rehabilitation after primary BC treatment (range 12 weeks–6 months) (Table 1b). No treatment effect was found on QoL or mood; however a short-term benefit for fatigue was observed. The outcome measures ‘health behaviour’ or ‘social function’ were not considered in these studies.
Cognitive behavioural therapy
Seven RCTs (nine publications) (n = 903) examined the effect of CBT. Three of the studies (four publications) (n = 616) examined the effect of the intervention during primary treatment (range 9–20 weeks). All found some mood changes (e.g. anxiety, ‘event-related distress’, and depression). A modest short-term benefit on fatigue was found in one study (intervention 9 weeks) 23 (Table 2a).
Four RCTs (five publications) (n = 287) examined the effect of CBT after primary BC treatment (range 6–12 weeks). Different QoL scores, such as FLIC 24, QLI 25, QoQ-C33 global QoL-scales 26, and FACT-B 27 were used (Table 2b). All showed short-term improvement on QoL (Table 2b), but only half showed effects on mood.
Social and emotional support
Five RCTs (six publications) (n = 1051) examined the effect of social and emotional support. Four studies (five publications) (n = 1001) assessed the effect during primary BC treatment (range 6 weeks–4 months), but revealed dubious effects (Table 3a); only one study (n = 50) assessed the effect of social and emotional support (6 weeks) after primary cancer treatment 22, 27. Significant improvement was seen on the POMS scale, but not on HADS and MAC scales (Table 3b).
This systematic review addressing the efficacy of different psychosocial interventions in BC rehabilitation observed limited and indifferent documentation on the effect of psychoeducation and social and emotional support, but documented some benefit of CBT on QoL.
Different aspects were focused on in the interventions intended to be applicable in clinical settings. In many of the CBT studies, additional procedures (support groups 25, existential group psychotherapy 17, educational strategies 26) were included. In addition, one study addressed stress reduction, coping skills, problem solving, and disease information 28, another focused mainly on coping skills 29, whereas one included health education, coping skills training, stress management, and psychological support 22.
Many studies had methodological shortcomings, i.e. lack of primary endpoints and calculations of statistical power. Results were reported as pre–post within-group effects 26, 28, 30, 31 or as a post-hoc analysis of subgroups (two- or three-way interaction effects, such as time × group × ‘low stress’ 28). One study reported growth-curve values without any control group and with poor availability of absolute scores 29. The intervention and control arms were imbalanced with respect to disease stage in two studies 17, 32, and few studies had intention to treat analysis 17, 26, 29, 33. Many trials had small sample size, thus a type II error is conceivable.
The internal validity was moderate in many of the studies due to the application of few standardised intervention programs with multimodal treatments and inadequate procedure reporting. Owing to the spectrum of outcome measures, and their respective instruments, calculating both internal consistencies/uncertainties and treatment effects is difficult. The assessment instruments were predominantly well known and validated in BC populations with some generic (e.g. SF-36 and POMS) and other cancer-specific inventory scales (e.g. FACT-B and EORTC-QLQ-C30), but some are less well known (UCLA-3) 32. Nevertheless, most of the studies made use of inventories that are thought to have good internal consistencies with adequate Cronbach alphas and reliability with test–retest and inter-rater validations.
Importantly, our findings support previous systematic reviews in this field 34, but we have included more recent studies than previous reviews. Limited evidence exists regarding psychosocial interventions. However, results related to CBT use appear promising with respect to their short-term effects on depression, anxiety, and QoL, as shown for distress managements in BC patients 8. Other systematic reviews and meta-analyses have reached divergent conclusions about effects of psychosocial interventions on QoL outcome for cancer patients, and all underline that more well-designed studies are needed 8, 35–37. Thus, there is no clear evidence to guide decisions on how psychosocial interventions should be designed to assist the BC rehabilitation. Whether all BC patients or selected groups with specific characteristics (age, ethnicity, stage of disease, early or late distress etc.) will benefit more than others from different types of psychosocial interventions remains unclear. Specific subgroup analysis, however, indicate that patients with high initial cancer-related stress might benefit more from social and emotional support, but this needs to be investigated in future studies 28.
Complex interventions, including exercise and psychosocial intervention, could be beneficial to BC patients 38. However, a wide range of uni- and multi-dimensional outcome measures and a variety of duration of intervention period were used in the included studies, which complicate any comparisons between studies. Therefore, estimating any combined effect by including results from individual studies in a meta-analysis, or even compare different studies for outcome measures, is difficult. The lack of control groups (i.e. no rehabilitation or intervention) may be a reason for the relatively minor effects of the interventions. However, not offering any intervention to women in need of rehabilitation seems unethical. A better understanding of the variability of the effects of the different interventions seems to be of major importance for optimising rehabilitation approaches.
The evaluation of the effectiveness of psychosocial interventions is complicated by small sample sizes, different outcome measures and assessment methods, and short follow-ups. A wide range of endpoints are normally used to evaluate the signs and symptoms of distress, including emotional, cognitive, physiological, and behavioural aspects 8. In addition, studies used a wide range of psychosocial and psychometric inventories and doses of interventions 8, 35–37. Therefore, summarising the overall results from different studies is difficult. This field of research is challenging, and the call for more studies is important. The consequences of small and heterogeneous studies result in little relevant knowledge for considering which psychosocial factors are beneficial for BC patients.
Implication for further research
Limited evidence is available from RCTs among BC patient to guide choices for BC rehabilitation interventions. However, this review presented some promising results: Findings support that CBT intervention decreases depression, anxiety, and improves the overall QoL. However, insufficient evidence exists to define optimal psychoeducational interventions beyond an information package that all cancer patients usually receive, and insufficient evidence is provided in these studies to determine the most beneficial social and emotional support interventions, even though some subgroups with high initial cancer-related stress seem to benefit, in particular. Limitations in present included studies are mostly the lack of a non-intervention control group. More studies are clearly necessary, and future research is needed to improve the understanding of structured interventions. Support groups should be examined, and psychological interventions should be tailored to meet the individual needs among distressed cancer patients, including adequate intervention doses. Importantly, the use of more valid inventories, and studies designed with better methodological quality, and collaboration in multicentre studies between and within countries, is needed.
This review had some promising results, especially for the use of CBT to improve QoL when continuing for 6–12 weeks (weekly sessions) after the primary treatment of BC patients. However, there is still insufficient documentation concerning important issues in breast cancer rehabilitation.
Methodological quality of randomised controlled trials on effectiveness of rehabilitation interventions for breast cancer patients
Appendix B: List of abbreviations
Affects balance scale
Loss of behavioural and emotional control
Brief symptom inventory
Cancer rehabilitation evaluation system
Cancer behaviour inventory
Cognitive behavioural therapy
Centre for Epidemiologic Studies—Depression Scale
95% confidence interval
Dealing with illness inventory
The European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire
Functional assessment of cancer therapy—breast
Functional assessment of cancer therapy—fatigue scale
Functional assessment of cancer therapy—general
Family assessment device
Functional living index
Fatigue symptom inventory
Global severity index
Hospital anxiety depression scale
Impact of event scale
Revised impact of events scale
Interview Schedule for Social Interaction
Mental adjustment to cancer scale
Multidimensional fatigue inventory
Mental health inventory—18
Monash Interview for Liaison Psychiatry
Measure of current status;
Memorial symptom assessment scale
National Cancer Institute print material
Profile of mood states
Positive states of mind
Perceived stress scale
Perceived social support scale
Post-traumatic growth inventory
Quality of life
Quality of life index
QLQ-C30 + 3
The European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire
Randomised controlled trial
Revised Piper fatigue scale
Rosenberg self-esteem scale
Structured clinical interview for DSM-III-R
Symptom experience scale
Medical outcome survey short form—36
Standardised mean difference
Social network index
State-trait anxiety inventory
State-Trait Anxiety Inventory—State
Total mood disturbance.
University of California, Los Angeles, Loneliness Scale—version 3