Mindfulness‐based interventions for psychological wellbeing and quality of life in men with prostate cancer: A systematic review and meta‐analysis

Mindfulness‐based interventions (MBIs) are increasingly being encouraged for managing treatment‐related symptoms but much less is known about the extent to which mindfulness is effective in relieving the psychosocial distress experienced by men with prostate cancer. A systematic review was conducted to synthesise the literature on MBIs for psychological wellbeing and quality of life improvement in men with prostate cancer.


| BACKGROUND
Prostate cancer is the second most frequently diagnosed malignancy in men after lung cancer with a global prevalence of over 1 million cases every year, and contributes to 3.8% of all cancer-related deaths. 1,2 The incidence of prostate cancer vary with the highest age-standardized prevalence recorded in Australia (79.1 per 100,000), followed by the United States of America (73.7 per 100,000) and across Europe (62.1 per 100,000). 1,2 In the United Kingdom, prostate cancer is the most diagnosed cancer with over 47,000 men diagnosed every year. 3 The availability of diagnostic tests to support early prostate-specific antigen (PSA) screening means that these prevalence is projected to double by 2040. 1,4,5 Similarly, given the therapeutic advancement in oncology, associated decrease in mortality rate is anticipated and many patients living longer after diagnosis. 5 Whilst this may be reassuring for patients, the uncertainty about patients quality of life remains an issue, as most men experience adverse health outcome associated with prostate cancer treatment.
Although all recommended treatments for prostate cancer take into consideration the age, comorbidities, and probable side effects, any chosen treatment is likely to cause incontinence, erectile dysfunction, impaired bowel function and fatigue in patients. 6,7 Hence, a diagnosis of prostate cancer often raises significant psychosocial distress in the form of anxiety and depression which further impacts the individual's day-to-day life. 7 In addition, issues related to the side effects of prostate cancer treatment, fear of disease progression for those on active surveillance and uncertainty about the disease reoccurrence after treatment further raises serious concerns on interventions required to support the psychological needs of patients with prostate cancer.
Patients living with prostate cancer suffer from a range of issues including urinary incontinence, sexual dysfunction, fatigue and gynaecomastia (a complication of hormonal therapies of the likes of androgen deprivation therapy) which can lead to anxiety, depression and poor quality of life. 8,9 Hence, psychological interventions to enhance the well-being of patients through their ability to recognise and manage the mental and emotional complications associated with the condition are particularly needed. 10,11 Mindfulness, a form of meditation, adopts a non-judgemental approach to paying attention to the present moment and becoming more aware of how our feelings, thoughts and physical state originates from our current situation. 12 The philosophy behind mindfulness is based on a person's ability to accept their current circumstance and control their emotions by becoming less reactive to unpleasant events, 13,14 and guides the individual towards accepting the status quo with a view of developing the attitude to de-emphasise its existence. 14 Mindfulness-based interventions (MBIs) improve quality of life, adherence to treatment anxiety and depression symptoms, and a range of other health emotionally distressing complications prostate cancer treatment. 10,11,15 The exploration of mindfulness in psychotherapy could be attributed to its therapeutic success in individuals with emotional disorders and its ability to minimise the chances of relapsing into depression. 13,16 There is an extensive body of literature on the psychological needs of men diagnosed with prostate cancer, who often present with anxiety, depression, and stress with the associated negative effect on their social interaction and quality of life. 17

| METHODS
The current study was conducted in accordance with the Centre for Reviews and Dissemination guidance. 24 Review findings was reported following the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) 25 in Figure 1. The PRISMA checklist for this review is presented in supplementary Table S1. The protocol for this review was registered with PROSPERO (registration number: CRD42020207788). Ethical approval was not sought as this study does not include confidential data or information likely to make participants identifiable.

| Criteria for studies
Given that this is the first review on this topic, we anticipated there will be sparse literature to limit inclusion of studies of participants at any specific stage of the disease journey or focused studies of men on any particular type of treatment for prostate cancer. Therefore, men (≥40 years) with localised or advanced prostate cancer who are on active surveillance, undergoing radiotherapy, chemotherapy, hormonal therapy, or had radical prostatectomy after diagnosis were considered for inclusion, regardless of the cultural or geographical context. Studies including participants with other cancer types were excluded.

| Comparison
Where applicable, studies with control arm that compared MBIs with usual care or treatment-as-usual, standard care, wait-list control, placebo and other active controls or interventions was eligible.

| Outcome
The primary outcomes measure were psychological outcomes (specifically including anxiety, depression, psychological distress and cancer-specific distress) and quality of life. The secondary outcome was post-traumatic growth.

| Study type
Studies that employed quantitative research methods such as randomised controlled trials (RCTs), non-randomised studies (NRS), controlled before-after studies, interrupted time series, cohort studies and case-control studies.

| Study selection and data extraction
Titles and abstracts of studies not available in English language were first translated using Google Translate to determine its eligibility as recommended by Balk et al. 26 However, studies were included if there were an available English language version from journal publishers. Identified citations were uploaded into the reference manager EndNote X9.3 (Clarivate Analytics) to remove duplicates. Titles and abstracts of all identified studies from the searched databases were screened independently by DAN and UOA against the eligibility criteria. Conference abstracts were also checked to establish if it has been published. Disagreements were resolved through discussion. Data extracted from eligible studies encompassed study characteristics (first author, year of publication, country and sample size); intervention (the type of intervention and control, duration of intervention in weeks); patient characteristics (mean age and current treatment); and patient reported outcome (PRO) and length of follow up (see Table 1). Excluded studies and reasons for exclusion were also provided in supplementary Table S7.

| Risk of bias and quality assessment
Two quality appraisal tools were selected based on the study designs of the included studies. A modified version of the Effective Practice and Organisation of Care (EPOC) 31 risk of bias (RoB) tool for randomised trials was used for quality assessment of RCTs. Included RCTs were assessed for bias in the following domain: selection bias (difference between baseline characteristics, poor random sequence generation and allocation concealment), detection bias (different outcome assessment), performance bias (blinding of participants, study personnel and outcome assessment), attrition bias (Incomplete outcome data) and reporting bias (selective reporting of outcomes and publication bias or spin). RoB in each domain and within each study was rated as low, unclear, or high (see supplementary Table S3 for the EPOC RoB domains). The National Institute of Health (NIH) quality assessment tool for before-after (pre-post) studies was used to assess NRS with no control group. A 'yes', 'no', 'cannot determine', or 'not applicable' or 'not reported' response was assigned to each question. Studies were rated good, fair or poor (see supplementary   Table S4).

| Data analysis
Similar study outcomes were grouped together, and statistically homogenous study results pooled in a meta-analysis using STATA software (Version 16). Narrative synthesis was used to summarise the findings from heterogeneous results. Randomised trials were eligible for inclusion in a meta-analysis. Entry into the metaanalysis was also restricted to effect measures collected from questionnaires that have been validated for the intended outcome. [32][33][34][35][36] The effect size of NRS was evaluated using the procedure recommended by Estrada et al. 37 Where included studies adopted different scales for measuring PROs, Murad et al. 38 emphasised the need to standardise the outcomes before pooling in a meta-analysis. Therefore, effect size of the standardised mean difference was expressed as (Cohen's d, [95% confidence intervals]).
With small, moderate and large effect sizes interpreted as Cohen's The heterogeneity among studies were measured using I 2 statistic. 39 I-squared (I 2 ) values less than 50% implies a low level of heterogeneity among the intervention effects, while an I 2 value greater than 50% signified a large degree of heterogeneity. 40 Where I 2 is greater than 50%, the DerSimonian-Laird randomeffects model was recommended. 40

| Study selection
The approach used for selecting included studies was summarised in a PRISMA flow diagram to ensure transparency in the screening and study selection process (see Figure 1) gov. After the removal of duplicates, a total of 1242 titles and abstracts were retained and screened for eligibility. A further 1179 studies did not meet the inclusion criteria. In total, 63 papers were subject to a full review of which 59 papers were excluded with reasons for exclusion recorded (see supplementary Table S7). Four studies comprising randomised trials (n = 3) and non-randomised studies (n = 1) met the inclusion criteria. No research summary on MBIs specific to prostate cancer patients, which was based on purposely paying attention and non-judgementally to their current experience was identified.

| Characteristics of included studies
MBIs used in the included studies are ACT, 29 MBCT 27,28 and MBSR 30 (see Table 1). Two studies 27,28 were conducted in Australia, and the other two studies 29,30 27,28 Up to 9.6% of the men were undergoing chemotherapy, 28 and 18.6% were on active surveillance. 28,30 The questionnaires utilised for measuring anxiety, depression, cancer-specific distress and quality of life is presented in supplementary Table S5. Of all the four included studies, one 29 compared the effect of ACT on participants undergoing penile injection therapy-based rehabilitation after radical prostatectomy with standard care. Therefore, quality of life was measured as sexual selfesteem and relationship quality (SEAR). 36 Two studies provided an account of men's experiences with treatment by performing an indepth interview on participants. 27,30 Another potential source of bias acknowledged in the two studies was incomplete outcome data. 29,30 Erectile function was not assessed in a longer term by Nelson et al. 29 Moreover, the study by Victorson et al. 30 did not assess if patients considered their experience to be traumatic due to fear of progression. Furthermore, a post-treatment analysis of fear related to disease progression was required to further support findings from the exploratory follow-up analysis. All studies considered intention to treat in their statistical analysis.

| Risk of bias and quality of included studies
Chambers et al. 28 further carried out per protocol analysis on study participants that completed more sessions.
The included NRS was a before-after study (pre-post) with no separate control group. The main source of bias was the unavailability of a control arm. Consequently, it might not be possible to make valid claims regarding the effect of an intervention by relying on a comparison between pre-and post-intervention result because findings may not infer causality. The single-arm study was judged to be of a low quality for using insufficient sample size; not measuring the clinical characteristics of participants at baseline, and no proof of blinding of participants and personnel. Since no baseline assessment of participants was carried out to determine individual level characteristics, the study result may not apply to a wider population.
Overall, the NRS was rated fair in the NIH quality appraisal tool (see supplementary

| Psychological wellbeing
Chambers et al. 27  However, there was a striking difference in the level of statistical significance in the study by Victorson et al. 30 with p < 0.05. Consequently, high-quality studies show a potential to change the initial evidence of effectiveness as observed in Table S6.

| Post-traumatic growth
There was an observed increase in Chambers et al. 28 (p = 0.086) and Victorson et al. 30

| Summary of evidence
The result of PROs for mindfulness intervention reveals that strong randomised studies may weaken or strengthen the evidence presented. While the NRS 27 showed a relatively large effect size, reported findings may not infer causality because of the perceived high degree of confounders associated with the study design. Furthermore, single-arm pre-post studies may constitute a low quality with the perceived overestimated effect size compared to randomised trials. 42 Besides, Cuijpers et al. 43 and Fitzpatrick-Lewis et al. 44 advised that reviewers should be cautious about using pre-post effect sizes because results may not be generalisable to a broader population since individual characteristics of study participants are not taken into consideration. RoB across studies with respect to number of recorded high risk: (0%-30% = low risk; 30%-70% = moderate risk; 70%-100% = high risk).

T A B L E 2 Risk of bias assessment across included studies
RoB within studies with respect to numbers of recorded low risk: (5/5 = low risk; 4/5 a = low risk; 3/5 b = moderate risk; 2/5 c = moderate risk; 1/5 = high risk; 0/5 = high risk). It is somewhat surprising that the control group experienced no observed positive changes in the outcomes of interest, as the study was likely to be contaminated in the control arm which received a mindfulness book.
Chambers et al. 28 reported no improvement in psychological outcomes and quality of life; and neither did MBCT influence mindfulness ability in men with advanced prostate cancer. In addition, there was also no significant difference between the two study arms after a per-protocol analysis was conducted at each time point. Though increasing age is associated with a less likelihood of receiving a restorative treatment for prostate cancer, ageing is also associated with multiple comorbid conditions including but not limited to congestive heart failure, hypertension, and diabetes. These chronic diseases are also associated with increased anxiety, depression and poor quality of life. 46

| Clinical implications and future research directions
This is a timely review providing an insight on the psychological needs,

ACKNOWLEDGMENTS
The authors declare no source of funding.

CONFLICT OF INTEREST
The authors declare no conflict of interests.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are openly available in PsychArchives at https://doi.org/10.23668/psycharchives.4763, reference number [54].