New Challenges in Psycho‐Oncology Research III: A systematic review of psychological interventions for prostate cancer survivors and their partners: clinical and research implications

Suzanne K. Chambers, Melissa K. Hyde, David P. Smith, Suzanne Hughes, Susan Yuill, Sam Egger, Dianne L. O'Connell, Kevin Stein, Mark Frydenberg, Gary Wittert, Jeff Dunn

worry, low mood 15 and diminished quality of life (QoL). 16 Partners of PCa survivors also experience ongoing psychological concerns and changes in their intimate relationships 17 ; with these impacts driven in part by the man's level of distress, sexual concerns and physical QoL. 18 In 2011, our group published the first criterion-based systematic review of psychosocial interventions for men with PCa and their partners. 19 We concluded that group cognitive-behavioural interventions and psycho-education appeared to be helpful in promoting better psychological adjustment and QoL for men with localised PCa, and coping skills training for female partners may improve their QoL. However, data were limited by inconsistent results and low study quality. In response to the increasing burden of PCa, uncertainties about optimal psychosocial care, and additions to the literature, we updated and extended this review with the intent of determining benefit and acceptability, and considering intervention content and format. In brief, we considered the range of psychosocial and psychosexual interventions that may be optimal, and for whom.

| METHODS
Two clinical questions guided the review 20  cognitive-behavioural (cognitive restructuring, behaviour change, cognitive-behavioural stress management), relaxation (relaxation techniques, meditation), supportive counselling (counselling/psychotherapy, health professional discussion), peer support (peer support, social support including discussion within a group of peers), communication (skill development to encourage communication with partners, health professionals or generally) and decision support (aids or tools to assist decisions about PCa treatment or use of sexual aids). The review and reporting of results were guided by the PRISMA statement. 21 Ethical approval was not required.

| Search strategy
Our prior review (until December 1, 2009) identified 195 articles that met criteria for the current study. 19

| Selection criteria
Studies were included if the following pre-specified criteria were met: • Randomised controlled trial design.
• ≥80% of participants were men diagnosed with PCa (no restrictions on disease stage or time since diagnosis) and/or partners/ carers of men with PCa or results for men with PCa and/or partners/carers were reported separately.
• Outcome(s) reported were psychosocial (including psychological, relationships, decision-making), health-related QoL, and sexuality outcomes (including sexual function, bother, and use of erectile dysfunction aids or treatments). Mediator outcomes such as cognitive reframing and coping were not included.
• Outcomes were assessed using validated scales or scales adapted from these.
• Intervention(s) were compared with usual care or supportive attention or no intervention, and/or another intervention(s) with different psychosocial or psychosexual components, and/or the same intervention components with a different mode(s) of delivery. Multimodal interventions such as lifestyle interventions were only included if they had a psychosocial or psychosexual component.
• Published in English language.
Two authors reviewed titles and abstracts and excluded irrelevant articles and duplicates. Full-text articles that potentially met criteria were then retrieved and reviewed by one author. A random sample of 5% of articles was assessed for inclusion by 2 authors with 100% agreement achieved.

| Data extraction
One author extracted pre-specified study characteristics (eg, participant demographics, PCa treatments, intervention content, delivery and results) and another checked each extract. To support data extraction, published descriptions of interventions were content analysed to create a framework of common psychosocial or psychosexual intervention components (Appendix B).

| Risk of bias
The Cochrane Collaboration's tool was used to assess risk of bias regarding sequence generation, allocation concealment, blinding of participants and personnel collecting outcome data, incomplete outcome data, selective outcome reporting, and other sources (eg, difference in follow-up between arms). 22 Blinding is difficult to achieve in psychological trials where consent mechanisms require participants to understand differences in treatments, which are often clearly discernible to the participant (eg, therapist-delivered intervention vs self-help materials). 19 On this basis, blinding was excluded from assessment. Clinical trial registries at https://clinicaltrials.gov/, http://www. isrctn.com/, and http://www.anzctr.org.au/ were searched for protocols of included studies to identify pre-specified outcomes and determine whether there was a risk of bias from selective outcome reporting. Differences in evaluations were resolved by discussion and where necessary adjudication by a third author.

| Intervention acceptability
The criteria of Yanez et al 23 were used to identify and evaluate aspects of interventions that indicate acceptability: ≥40% recruitment rate, ≥70% retention at end of intervention or follow-up (or <30% withdrawal), and ≥70% average intervention attendance.

| Analyses
It was anticipated that some trials may be underpowered. 19 Thus, an intervention was considered potentially beneficial compared with usual care or better than another intervention if for at least one reported outcome (at the longest reported follow-up), there was in favour of the intervention(s): (i) a statistically significant difference between arms; (ii) a moderate or large standardised effect size (eg, Cohen's d ≥ 0.5, η 2 ≥ 0.06); or (iii) a difference in mean score changes from baseline calculated by ANCOVA or multiple linear regression between arms ≥10% of the scale of the differences in means. For a given measurement scale, results from subscales were only considered in the absence of an overall score.

| Search results
In all, 6631 citations were identified of which 161 full-text (including 16 identified from reference lists) were retrieved and evaluated FIGURE 1 PRISMA flow diagram of study selection for systematic review as well as 195 articles from the prior review. 19 Of the total 356 full-text articles assessed for inclusion, 68 articles met criteria and reported a total of 57 RCTs. Forty-one RCTs reported in 51 articles (2 publications for 10 studies) included only patients (Q1); 1 RCT included only partners (Q2); 15 RCTs reported in 16 articles (2 publications for 1 study) included patients and partners (Q1 and Q2) ( Figure 1). Most studies were excluded because of study design or population not meeting criteria, or results for patients or partners/carers were not reported. Clinical trial registry searches identified 47 trials: 25 completed (16 included in the review); 20 ongoing; 2 terminated (slow accrual, funding unavailable).

| Trial characteristics
Included trials randomised 8378 men (range 27-740; 48% of trials had <100 participants), and 1313 partners (range 27-263; 57% of trials had <100 participants; >90% partners were female in 14 trials; >80% partners were spouses in 12 trials). Most (67%) trials were conducted in North America. In 10 trials (4 including partners), participation was determined by socio-demographic background (eg, African-American), emotional state (eg, distress), or QoL (eg, urinary or sexual dysfunction, ADT treatment side-effects, fatigue). When reported, mean or median age was below 65 years in 49% of trials for patients and below 65 years in 100% of trials for partners. In approximately half of trials (57% of patient trials, 40% of partner trials) reporting college/university education, >50% of participants were university/college educated. In 25 trials (45%), men were diagnosed with or treated for localised disease in the previous 6 months (14 trials enrolled men prior to treatment or treatment decision). Men with recurrent or metastatic disease and their partners were included in 16% and 21% of trials, respectively.
The number of relevant outcomes measured by trials varied from 1 to 16 (patient) and 2 to 12 (partner). Most common outcomes for patients were sexual bother and/or function and mental health; and for partners were relationships, general and cancer-specific distress. Trials

| Intervention effects
Three trials reported couple-focused interventions that, compared with usual care, increased partner distress about sexual function, 24 worsened partner challenge appraisal, 25 and reduced relationship satisfaction and intimacy for partners who had high levels of these constructs at baseline 26 (Appendix D). By contrast, for patients, all intervention effects indicated improvement. Four trials included outcomes of interest 27-30 but did not report comparative results and were excluded. The remaining 29 trials (21 person-focused: 20 patients, 1 partner and patient; 8 couple-focused) showed a benefit for psychosocial or psychosexual outcomes (

Decision making
Six trials improved patient decision-making mostly for men diagnosed with early stage disease and/or recruited prior to treatment. Decision support, aid, or navigation reduced patient uncertainty, 31,32 conflict, 33 and regret 34,35 about their treatment decision, and a combined online psycho-educational intervention and moderated peer forum also reduced regret. 36,37 Patient self-efficacy or confidence in their decision-making was increased by decision navigation 34 and interactive education interventions. 38

| DISCUSSION
Psychosocial and psychosexual intervention can improve decisionrelated distress, mental health, domain-specific, and health-related QOL in men with PCa. Combinations of educational, cognitive behavioural, communication, and peer support have been most commonly applied and found effective; followed by decision support and relaxation; and to a much lesser extent supportive counselling. These components were often used in a multi-modal approach, and delivered through both face-to-face and remote technologies, with therapist, nurse or peer support. In sum, multi-modal psychosocial and psychosexual care for men with PCa, particularly localised disease, is both acceptable and effective.
The evidence is less clear for the female partners of these men and couples as a dyadic unit. Couple-focused interventions were the least acceptable approach and almost half of the couple interventions produced poorer outcomes for partners. When couple interventions were effective, they improved relationship outcomes for the partner but not the man; had a positive effect on the partner's mental health but conversely; improved sexuality outcomes for the man but not the partner.
No interventions improved sexuality outcomes for female partners.
Based on these results, effective and acceptable interventions for female partners and couples remain an area of uncertainty. It may be that couples interventions have been primarily focused on the PCa survivor's needs, leaving the partner's concerns poorly managed. This is an area where significant further work is required to understand the needs and preferences of couples, and to determine approaches to improve sexual and relationship satisfaction for both partners.
Limitations of the research to date include small sample sizes; low statistical power; suboptimal statistical methods in some studies; inconsistency in measurement approaches; a lack of diversity in participants-particularly with regards to gay and bisexual men; men with advanced PCa; and men from socio-economically deprived; and non-Anglo-Saxon backgrounds. Long-term survivorship outcomes (>2 years) are yet to be addressed. In addition, intervention components were often described in a vague way such that it was not always clear what was actually delivered; and treatment fidelity and therapist adherence was in most studies not well described. Strengths of the current review by comparison with previous reviews include a departure from a narrow focus on specific intervention type(s), single outcomes, or sub-groups; a consideration of acceptability as well as statistical significance; and examination of not only intervention effectiveness but also who benefits by considering the influences of socio-demographic and medical characteristics of men and their partners; intervention format and delivery; and acceptability.  # Searches S17 S3 AND S15 Published date: 2009-2016; English language; Exclude MEDLINE records S16 S3 AND S15 S15 S4 OR S5 OR S6 OR S7 OR S8 OR S9 OR S10 OR S11 OR S12 OR S13 OR S14 S14 TX allocat* random* S13 (MH "Quantitative Studies") S12 (MH "Placebos") S11 TX placebo* S10 TX random* allocat* S9 (MH "Random Assignment") S8 TX randomi* control* trial* S7 TX ((singl* n1 blind*) or (singl* n1 mask*)) or TX ((doubl* n1 blind*) or (doubl* n1 mask*)) or TX ((tripl* n1 blind*) or (tripl* n1 mask*)) or TX ((trebl* n1 blind*) or (trebl* n1 mask*)) • Behaviour change: Goal setting and problem solving or behavioural maintenance.

| Clinical implications
• Cognitive behavioural stress management: intervention identified as CBSM.

Supportive counselling
• Counselling/psychotherapy (as identified by the study author): counselling or therapy offered as part of the intervention including sexual therapy, excludes cognitive-behavioural approaches.
• Health professional discussion: discussion with a health professional (excludes counselling/psychotherapy, routine/standard care).
Peer support • Peer support: shared experience with a peer who also has PCa (includes support groups, social support).
• Social support: mentions social support generally and may also include informal peer support in a group setting, or does not specify type.