Introduction
Mindfulness-Based Stress Reduction (MBSR) (Kabat-Zinn,
1990) and Supportive Expressive Group Therapy (SET) (Classen et al.,
1993) are two frequently studied and well-validated psychosocial group interventions for cancer patients. The MBSR program was adapted by us for the treatment of cancer patients and is called Mindfulness-Based Cancer Recovery (MBCR) (Carlson & Speca,
2010). Several clinical trials on MBCR and other mindfulness-based interventions (MBIs) have shown efficacy in improving psychosocial outcomes in cancer patients, such as mood disturbance, stress symptoms and quality of life (e.g., Carlson et al.,
2003; Johannsen et al.,
2016; Lengacher et al.,
2016; Speca et al.,
2000; Wurtzen et al.,
2013). Recent meta-analyses provide further support for the effectiveness of MBIs in cancer with moderate to large effect sizes (Cramer et al.,
2012; Piet et al.,
2012; Zhang et al.,
2016). While SET differs from MBCR in terms of content, focus and theoretical underpinnings, the programs are similar regarding group format, group size and contact hours. In several clinical trials, SET also demonstrated its effectiveness in reducing stress symptoms and improving quality of life (e.g., Classen et al.,
2001; Kissane et al.,
2007).
Gaining a deeper understanding of
how interventions work can advance treatment research (Kazdin,
2007; Moyer et al.,
2012). Mediation analyses may help identify the active components of the intervention and subsequently the intervention can be optimized by tailoring the program accordingly (Kazdin,
2007). Studies examining the mechanisms of an intervention typically focus on the theoretical foundation of the program. SET is based on the idea that participants learn to better cope with their cancer and feel less distressed by expressing emotions and increasing the experience of social support. Clinical trials have examined the effects of SET on potential working mechanisms, finding a decrease in suppression of negative affect and improvements in social functioning (Giese-Davis et al.,
2002; Kissane et al.,
2007). However, to the extent of our knowledge, no study has examined whether changes in these potential working mechanisms mediated the effects of SET.
In MBIs, participants learn to develop non-judgmental and accepting awareness of experiences by practicing mindfulness, which in turn results in a decrease of psychological distress (Kabat-Zinn,
1990; Segal et al.,
2002). A recent meta-analysis of 12 studies concluded that there is consistent moderate evidence for increases in mindfulness as a mechanism of MBIs (Gu et al.,
2015). This mediation process, however, has mainly been studied in comparison with waitlist and usual care control groups. Such research designs cannot rule out whether non-specific factors such as social support may also underlie the positive effects of MBIs (Chiesa,
2011). Independent of the content of an intervention, being in a group with fellow patients and sharing personal experiences can create feelings of social support. In turn, the frequency and quality of social networks has often been positively associated with wellbeing and health (Reblin and Uchino,
2008), potentially explaining the intervention effects. A recent meta-ethnography based on 14 qualitative studies indicated that social support can play an important role in MBIs (Malpass et al.,
2012). Through the practice of non-judgemental awareness, MBI can provide an atmosphere that fosters the allowing and accepting of thoughts and emotions, facilitating openness and sharing experiences with one another (Schellekens et al.,
2016). Such a supportive environment can help patients to facilitate each other’s learning processes (Mackenzie et al.,
2007; Schellekens et al.,
2016).
The aim of the present study was to examine whether change in mindfulness and/or social support mediated the effect of MBCR compared to SET, on change in mood disturbance, stress symptoms and quality of life among distressed breast cancer survivors. As MBCR is mainly focused on practicing mindfulness and SET on facilitating support, we hypothesized that enhancement in mindfulness would mediate the effects of MBCR while enhancement in social support would mediate the effects of SET.
Discussion
The goal of the present research was to examine potential mediators underlying the effect of MBCR and SET on psychological outcomes of breast cancer survivors within a randomized clinical trial. We expected changes in mindfulness to be related to participation in MBCR and social support to be more strongly affected in SET. Surprisingly, social support increased more after MBCR than after SET, and this change partially mediated the effect of MBCR on mood and stress symptoms. While fostering group support in MBCR is more an implicit part of the program than in SET, where it is a central objective, MBCR seems to provide an environment where breast cancer survivors support and feel supported by one another. These findings are in line with qualitative studies (Malpass et al.,
2012), showing how MBIs provide a safe environment, which enhances a sense of community (Schellekens et al.,
2016) and facilitates patients to learn from one another (Mackenzie et al.,
2007; van den Hurk et al.,
2015). MBCR also includes content and experiential practice focused on the cultivation of loving-kindness. Learning to direct kind and compassionate attention towards oneself and others may enhance feelings of relatedness and support within and outside of the group (Birnie et al.,
2010). Note that in the subsample of women who completed both the questionnaires as well as the intervention the effect on social support disappeared. A possible explanation could be the drop in sample size from 139 to 125 women, which limits the power and increases the chance of a type II error occurring.
Thus far, mediation studies on MBIs have focused on working mechanisms within the individual (Gu et al.,
2015) whereas mediators on the level of the group such as social support, seem to potentially be of equal importance. The mediating effect of social support could be seen as a non-therapeutic outcome inherent to a group-based intervention (Chiesa,
2011). However, by comparing MBCR with another group-based training, the increase in social support appears to be an underlying mechanism that might be unique to MBCR and partly explains its positive effects in cancer patients. Note that social support only partly explains the effect of group on mood disturbance and stress symptoms. Other working mechanisms of MBCR affecting outcomes were not addressed in this analysis, but may also be important. These may include the promotion of self-compassion and emotion regulation skills (for an overview, see Gu et al.,
2015).
Surprisingly, no effect was found on mindfulness after MBCR versus SET. This finding is in contrast with several studies showing that the change in mindfulness after MBI participation mediates the effects on several outcomes (Gu et al.,
2015). A possible explanation for these contrasting results might be the use of the MAAS, as it only measures one aspect of mindfulness, i.e. the presence/absence of attention to and awareness of present moment experiences during daily activities, or put differently, the opposite of “running on automatic pilot”. Possibly, SET participants may also become more aware of their emotions, thoughts and behaviours during daily experiences because of the programme’s emphasis on expressing emotions and improving coping skills. Other instruments, such as the Five Facet Mindfulness Questionnaire (FFMQ) (Baer et al.,
2006), might measure mindfulness skills that are more specifically practiced in MBCR than in SET, e.g. Observing and Nonjudging of inner experiences. In our previous work with the FFMQ, MBCR had the strongest effect on the Observing facet compared to other FFMQ facets and MAAS (Labelle et al.,
2014). Baer and colleagues also demonstrated that the Observing facet appears to be particularly sensitive to meditation experience (Baer et al.,
2006,
2008).
A number of limitations should be noted. The study sample consisted solely of women with breast cancer, of which the majority was highly educated and on average they received their cancer diagnosis 2 years prior to participation, limiting the generalizability of these findings. To date, the vast majority of study participants in MBIs for cancer patients have been women with breast cancer (Piet et al.,
2012). Future research should also examine the effectiveness and potential working mechanisms of MBIs in other types of cancer. Another limitation is the relatively high intervention drop-out rates. It might bias the sample and decrease generalizability of the results. As most participants did not provide a reason for dropping out, we do not know whether attrition is related to practical reasons, such as scheduling conflicts, or due to the high level of burden data collection presented to them. In addition, the sample of follow-up data was limited, preventing us from including it in the mediation analysis. Consequently, both the mediator as well as the outcome variables were assessed before and after the intervention, which limits conclusions about what changed first, social support or mood/stress symptoms. Future studies should take the temporal order of the mediator and outcome variable into account, exploring whether early changes in the mediator predict later changes in the outcome (Labelle et al.,
2014). In previous work, for example, we demonstrated that early increase in observing (i.e., change during first half of MBCR) predicted later increase in awareness of daily activities (i.e., change during second half of MBCR) (Labelle et al.,
2014). Another limitation is that we relied on a self-report questionnaire for assessing mindfulness measuring only one facet of mindfulness (i.e. attention to and awareness of daily experiences). While the validity of the measure is under debate (Grossman & Van Dam,
2011), the scale has high internal consistency and has been adopted successfully in studies on the effects of mindfulness (Brown & Ryan,
2003), and for use in people with cancer (Carlson & Brown,
2005). In addition, the MAAS was filled out by a smaller sample of participants (n = 110) than the other questionnaires (n = 139) due to procedural changes in the protocol partway through the study, potentially decreasing statistical power to detect group differences.
The present study implies that the group-based character of MBCR is of added value to breast cancer survivors’ mood and stress. This implication should be seen in the light of an increase in MBIs that are adapted to the individual in both clinical practice as well as in research settings (Compen et al.,
2015; Tovote et al.,
2014; Wahbeh et al.,
2014; Schroevers et al.,
2016). For patients who are unwilling or unable to participate in a group due to disabilities or constrained time schedules, an individual MBI program appears to be a good solution. However, these patients will miss out on the group support and observational learning that typically occurs in group-based settings. Future non-inferiority trials should examine whether individual-based and group-based MBIs are equally effective in improving mood and stress in cancer patients. In addition, it would be interesting to examine whether the central role of social support only holds for (breast) cancer or also generalizes to other MBI target groups. When the main motivation for participants is learning to cope with day-to-day stressors rather than coping with a life-threatening diagnosis, actual mindfulness practice might be more important than social support. Testing social support as a mediator across populations and MBIs will inform our understanding of this intervention and may lead to program modifications that might maximize the effectiveness of MBIs.