Major depressive disorder (MDD) has prevailed as a leading cause of non-fatal health loss for nearly three decades according to the World Health Organization and is predicted to be the leading cause of disability worldwide by 2030 (Vos et al., 2017). It is a common psychiatric disorder, affecting about 30.3 million people each year in Europe alone (Wittchen et al., 2011). Approximately one out of five adults will suffer from depressive disorder during their lifetime.

Adding to the burden, depression is also characterized by high relapse rates (Mueller et al., 1999): About 75% of all patients experience more than one episode during their lifetime (Boland et al., 2002). In a study examining up to 15 years of prospective follow-up data on the course of MDD, Hardeveld et al. (2010) found that a cumulative proportion of 85% of the 380 recovered subjects experienced a recurrence versus only 35% in the general population. The number of depressive episodes was shown to be a consistent predictor: The risk of recurrence rose with each new episode. Presence of subclinical residual symptoms, however, appeared to be the strongest predictor. Therefore, in addition to treating acute depression, further treatment of remaining residual symptoms is highly important (Hardeveld et al., 2010).

Currently, both psychological and pharmacological treatments are relatively well established as having at least moderate efficacy for the treatment of acute major depression (Baldessarini et al., 2015; Bauer et al., 2002; Undurraga & Baldessarini, 2012). An equally common course of action for moderately to severely depressed patients is a combination of psychological treatment and pharmacotherapy, which has been shown to be more effective than both psychotherapy or pharmacotherapy alone (Cuijpers et al., 2009a, b). Despite these therapeutic options, in their review on unipolar depression, Forte et al. (2015) found that during an average follow-up period of 12 to 13 years depressive patients reported being ill nearly 50% of the time. Considering the presence of substantial residual symptoms in depressive patients who have been successfully treated, there is a clear need for improving the treatment of depression as well as the prevention of relapse. Research recommendations include further exploration of specific, improved individual and combination therapies (pharmacological and psychological), as well as identification of clinical predictors of treatment responses in MDD (Sim et al., 2016).

To address the need for psychological interventions targeting relapse prevention, Segal et al. (2012) developed mindfulness-based cognitive therapy (MBCT). A meta-analysis by Kuyken et al. (2016) showed that MBCT for patients with recurrent depression in remission resulted in a 31% risk reduction of a relapse/recurrence in a 60-week follow-up period compared with all control conditions. A growing number of studies have indicated that MBCT may also be effective in decreasing depressive symptoms in patients with current depression (Strauss et al., 2014).

However, several studies have shown that even after MBCT residual symptoms remained considerable (Piet & Hougaard, 2011; Van Aalderen et al., 2012). Focusing on known working mechanisms of MBCT in the development of sequential treatment may lead to further benefits. One of the most established working mechanisms of MBCT is reduction in rumination. A meta-analysis by Van der Velden et al. (2015) including 23 studies on MBCT reported that alterations in rumination, worry, and meta-awareness were associated with predicted or mediated treatment outcome. Besides this, mindfulness and compassion were also found to mediate treatment outcome: Kuyken et al. (2010) found a decoupling of the relationship between cognitive reactivity and depressive symptoms after MBCT, which was associated with the cultivation of self-compassion. Normally, cognitive reactivity predicts depressive relapse, but in participants who developed self-compassion, this was no longer the case.

Gilbert and Procter (2006) found that one of the possible underlying mechanisms for the chronic and recurrent nature of depressive symptoms is low self-esteem or self-denigration. These findings are in line with research by Beck (1967) and Beck et al. (1979), which indicated that recurrently depressed patients suffer from severely self-denigrating core beliefs. It may be that the relation between cognitive reactivity and relapse is connected to the activation of such self-denigrating core beliefs and that compassion mitigates the effect of this activation, or even prevents it. Being able to adopt a caring attitude towards the self might be a skill that could help reduce the undermining mechanisms of self-criticism and hence reduce the vulnerability to recurrence of depressive symptoms. As (self-)compassion is taught mostly implicitly in MBCT (Segal et al., 2012), explicitly cultivating self-compassion may offer a complementary contribution to reducing rumination and increasing mindfulness skills for individuals who are prone to depressive relapse or recurrence.

Contemplative literature emphasizes the importance of developing mindfulness and (self-)compassion simultaneously: In their most recent publication, Feldman and Kuyken (2019; p. 180) emphasized that “the key insight is that mindfulness reminds us to return to the actuality of the present with an attitude of befriending, to establish a body of stillness and calm where we can find the strength to meet the difficult with balance and compassion.” Mindfulness is defined as “the practice of purposely bringing one's attention in the present moment without judgment” (Kabat-Zinn, 1994; p. 4). Compassion is defined as “the capacity to open to the reality of suffering and to aspire to its healing,” presenting a “multi-textured response to pain, sorrow and anguish including kindness, empathy, generosity and acceptance” (Feldman & Kuyken, 2011; p. 144). In line with this definition, Strauss et al. (2016) concluded that “compassion consists of five elements: recognizing suffering, understanding the universality of human suffering, feeling for the person suffering, tolerating uncomfortable feelings, and motivation to act/acting to alleviate suffering” (p. 15).

Although still in its infancy, the evidence base for the efficacy of compassion programs is growing. A meta-analysis on compassion-based interventions in non-clinical populations including 21 RCTs showed moderate effect sizes for improvements of compassion, self-compassion, mindfulness skills, depression, and well-being (Kirby et al., 2015). Ferrari et al. (2019) found in their meta-analysis that (self-)compassion interventions led to a significant improvement. In the overall population, a hedge’s g of 0.66 was found for depression and 0.75 for (self-)compassion. In the clinical subpopulation, a hedge’s g of 0.82 was found for self-compassion.

Van den Brink and Koster (2015) developed mindfulness-based compassionate living (MBCL), partly based on previous work in compassion research as conducted by Neff and Germer (2013), Germer (2009), and Gilbert (2009). MBCL was designed as a follow-up training to MBCT or mindfulness-based stress reduction (MBSR), to be used in both clinical and non-clinical settings. For a more detailed description of MBCL, we refer to Schuling et al. (2016). Recently, Schuling et al. (2020) conducted an RCT to assess efficacy of MBCL in a population of recurrently depressed adults (N = 122) which showed MBCL resulted in a reduction of depressive symptoms and increases in self-compassion, mindfulness, and quality of life. The increase in (self-)compassion appeared to mediate the improvements of depressive symptoms.

Despite these positive findings, little is known about participants’ perspectives on the experienced added value of compassion interventions such as MBCL. Although kindness and compassion are key attitudes underlying MBCT, in its pedagogy, they are conveyed more implicitly. Particularly, patients with recurrent depression may need more explicit instructions and additional support to develop a compassionate attitude to both self and others, as they often are highly self-critical and plagued by feelings of shame, guilt, and inferiority (Gilbert & Procter, 2006; Judge et al., 2012).

The qualitative study was embedded in an RCT comparing MBCL added to treatment as usual (TAU) to TAU alone in patients with recurrent depression who had participated in MBCT in the past (Schuling et al., 2020). Following our hypothesis that explicitly cultivating (self-)compassion may offer a complementary contribution to reducing depressive symptoms and increasing mindfulness and (self-)compassion skills for individuals who are prone to depressive relapse or recurrence, the aim of the current study was to gain insight into participants’ experiences with this more explicit teaching of (self-)compassion in MBCL: Would MBCL, with its (partial) overlap in format and content, lead to similar experiences as MBCT, or would participants gain new, complementary skills? And would their experiences with MBCL lead to new insights into clinical predictors of treatment responses in MDD?

Method

Participants

At least one year prior to the trial, the participants had all participated in MBCT at our center. As the aim of the study was to explore the potential added value of MBCL, we only contacted participants who had participated in a minimum of four sessions. We selected participants with a sampling strategy aimed at having as much diversity as possible, taking account of gender, age, current status of depression, number of previous depressive episodes, age of onset of depression, childhood trauma, and satisfaction with MBCL (as perceived by the teacher). Of the 23 patients contacted by telephone, one person indicated that he did not want to participate in an interview. The remaining 22 participants were able and willing to participate. After 22 interviews, saturation was reached. All participants provided written informed consent before participation in the study. Participants’ demographic and clinical characteristics are shown in Table 1.

Table 1 Demographic and clinical characteristics of the participants (subsample of the randomized controlled trial into the efficacy of mindfulness-based compassionate living added to treatment-as-usual, compared to only treatment-as-usual; N = 22). *Out of 18 instead of 22
Full size table

Procedure

The MBCL courses were delivered at the Radboudumc Centre for Mindfulness, Nijmegen, the Netherlands. Groups of about 8–10 participants were taught by one of two teachers, who both met the Good Practice Guidelines for teaching mindfulness-based courses by the UK Network for Mindfulness-Based Teacher Training Organizations (https://app.ukmindfulnessnetwork.co.uk/guidelines/) and were trained to teach MBCL by its developers Frits Koster and Erik van den Brink (HJ and RM). The MBCL training consisted of 100 h of training plus 80 h of supervised teaching. Treatment integrity and therapist competence were assessed as competent based on two randomly selected videotapes of each teacher using the adapted mindfulness-based intervention teacher assessment criteria (MBI:TAC) (Crane et al., 2013), rated by two experienced mindfulness teachers.

Within three months after completing MBCL, participants were invited by telephone to take part in an individual interview about their experiences with MBCL. The interviews were held with participants who consented to participate. The interviewers had not been involved in the clinical care or MBCL training of the participants. Before starting the interviews, interviewers introduced themselves, their occupation, and their role in the study to the participants. After this, the interviewers asked permission to record the interview.

The interview started with one open-ended question: “Do you feel MBCL has had an added value over MBCT alone, and if so, what is this added value according to you?” If participants responded positively on the first question, the added value was explored using additional, open-ended questions. If answered negatively, overall experiences with the MBCL were explored. The interviewers pursued unexpected areas of interest to collect as wide a variety in responses as possible. The interviews were recorded with a Dictaphone and transcribed verbatim. No one but the participant and interviewer were present during the interview.

Data Analyses

The interviews were coded and analyzed by the interviewers in a multistage process, using the methodological orientation of a grounded theory (Boeije, 2014), which is aimed at theory development and provides explicit, sequential guidelines for conducting qualitative research, most importantly in offering specific strategies for handling the analytic phases of inquiry and integrating data collection and analysis. The research team consisted of six people: two psychiatrists (AS and HR), one research psychologist (MH), one Mindful Self-Compassion teacher (NS), and one linguist (RS). One psychiatrist had special expertise in diagnosing and treating adults with recurrent depression (AS), and both had expertise in the application of mindfulness-based interventions. Three members of the team were familiar with qualitative research (AS, HR, and RS). One member had an experience as a compassion intervention teacher but had not been one of the trainers of this population (NS). At the time of data collection, first interviewer RS was a 37-year-old female linguist, MBSR teacher, and PhD student. She had practiced mindfulness for approximately eight years, had graduated from the MBSR/MBCT teacher-training program, and had taught six MBSR classes. The second interviewer (HR) was a 33-year-old female psychiatrist. She had practiced mindfulness for approximately eight years and had taught five MBCT courses for patients with psychiatric disorders. RS held a prior relationship with the participants, whom she had done assessments with pre- and post-intervention for the RCT. HR had no prior relationship with the participants.

The qualitative analysis consisted of three coding phases, namely, open coding, axial coding, and selective coding (Boeije, 2014). In this way, coding was data driven instead of theory driven. A repeating cycle was used, which consisted of conducting four interviews, transcribing and coding these four interviews, and comparing notes (HR and RS). After three such cycles, i.e., twelve interviews in total, the data was analyzed with the whole research team. After this, it was determined which information remained underexposed: It was found that more information was required on which elements of the didactic and practice content of MBCL led to (which) additional value. As a result, additional questions were added to the interviews. Then, the next six interviews were conducted, after which the team came together again to discuss the data. At this stage, i.e., after 18 interviews, it was found a few of the themes that were emerging needed more exploration. Additional interviews included in-depth probing on these themes.

Open coding consisted of reading and re-reading the interviews and developing a code tree. During the first interviews, this was conducted by both interviewers together. To ensure reliability, transcripts were coded independently by the two interviewers, and the codes were compared and discussed until agreement was reached. Field notes and memos were used as sources of information throughout this process. Codes were also discussed in the research team. The list of codes was used during the coding of the next interviews. This was followed by the next phase of axial coding in which the interviewers made a list of categories and (sub) themes by the list of codes, which was discussed with the research team. After this phase, the findings were integrated within the different categories and checked and re-integrated by the research team. For reasons of triangulation, the two teachers who provided the training were interviewed as well, and notes from these interviews were integrated in the overall analysis. After this phase, a more detailed analysis of each theme was developed, and compelling quotes were selected. During this process, Atlas.Ti software (2017) was used for analyzing and classifying qualitative data.

The first 18 interviews were held face-to-face at the Radboudumc, the remaining four were held by telephone. The interviews lasted between 20 and 60 min. In total, 37.4 h of audiotaped interview data were collected (mean duration was 36.5 min per interview).

Ethical Considerations

The qualitative study was exempt from ethical approval by the Medical Ethical Committee Arnhem/Nijmegen, the Netherlands, by whom the RCT was approved (2013/220). All participants were informed about the study procedures and the fact that participation was entirely voluntary. Participants consented to audio recordings of the interviews and to the use of data for research purposes before the interview commenced.

Results

Two of the 23 participants interviewed indicated that they had not experienced additional value: One was neutral, stating that the value was the same as that of MBCT and the other found that the value was less: MBCL exercises, with their focus on imagery and recalling difficult experiences, drew him into thinking instead of feeling. For all participants, MBCT had been of sufficient value to be interested in a follow-up program, though nearly all had dropped practice routine.

Four themes and ten subthemes were identified in the data; these are represented in Table 2.

Table 2 Themes and subthemes with example quotes that emerged from qualitative analysis
Full size table

Theme 1: the Container of Kindness

Participants mentioned that from the start, the program was infused with explicit kindness, evident in all elements: the practices, theory, guidance of the teacher, and group inquiry. The container was sometimes described as a very basic quality of “being okay” with and welcoming all that participants experienced and brought into the sessions. Participants noted that practicing within this container changed their experience of phenomena”.

In one word: allowing. Yes, that’s right. Allowing, but especially allowing in so much love. (Male, age 70)

Not only the content and guidance of practices was experienced as explicitly geared towards kindness; participants also reported how the teacher, during inquiry, encouraged participants to be flexible and lenient with home practice:

Well, I simply couldn’t do it, and normally I would see that as a huge failure, but because there was space for: “if it doesn’t work, it simply doesn’t work”. I didn’t get that feeling of failure as much and I felt a sense of mildness come over me. (Female, age 37)

Subtheme: Coping with the “Backdraft” Effect

Not all participants found it easy to connect with kindness though. Some noted that they were not used to being gentle with themselves, which made it hard and confronting. Many participants identified the resistance as a signaller of deeply painful experiences and stated that with hindsight, not giving in to the resistance, and subsequently allowing gentle exposure to what had been feared, was what propelled transformation of the pain the most.

Theme 2: Exposure to the Difficult

Throughout the interviews, it became evident that being able to expose oneself to a fearful or painful experience required a willingness to allow that experience. Participants reported the active approach to the difficult as an essential difference between MBCT and MBCL. Within the container of kindness and self-care as described in theme 1, participants were explicitly invited and supported in approaching fearful or painful emotions and thoughts that sometimes had been repressed for a long time.

Subtheme A: Overcoming Resistance—Approaching Rather than Avoiding the Difficult

This active approach was reported by many participants; therefore, we identified it as an important subtheme within exposure to the difficult. Participants stated that during MBCT, in contrast, they had been able to remain in more “neutral” observing, essentially avoiding what scared them the most.

Quite a few participants were able to identify tendencies of passive or active avoidance behaviors that normally prevented approaching the difficult:

I notice now it doesn’t help me, talking about things. I made a very conscious choice not to do that. It can trigger me even more. Really staying aware of yourself and letting yourself feel the emotions is a lot more helpful. (Female, age 61)

Participants noticed the difference between avoiding emotions by thinking about them, and actually feeling them. As one participant remarked:

It’s allowed to touch you, it’s okay that it hurts. Those things are really helpful when you let yourself experience it and don’t suppress it. That’s the difference, I feel it, it’s not just theoretical anymore. If it were just theory, you’d think: “sure, it’s okay to feel pain. Whatever.” Then you’re just using tricks. So, the authenticity [of the experience] is what’s most important to me. (Female, age 61)

In many practices, this approaching was promoted by asking participants to recall painful memories and allowing them to come to the foreground:

I only know what it’s called in German. “Zuwendung” [turning towards]. It’s a particular mode of observing. (Female, age 46)

Subtheme B: Allowing the Difficult Without Trying to Fix It

Several participants mentioned that aside approaching the difficult instead of avoiding it, a non-fixing attitude was required to expose themselves to deeper suffering. It seemed easier to expose themselves to pain if they let go of their desire to change it.

Definitely. That’s it. Surrendering to what’s here. And then comes the peace. (Female, age 48)

Subtheme C: Effects of Feeling the Difficult

Participants reported that once pain had been approached and allowed, there was room for exposure, experiencing the felt sense of difficult emotions. This was mentioned as a specific aspect that MBCL addressed more than MBCT did and it seemed to result in a diminishing of suffering:

For me, it was a revelation… now I dare to feel that emotion using those CDs and for the first time in my life I’m letting myself feel all that and I know: well, I’m not going to die, it’s not going to end horribly. (Female, age 48)

Participants reported that having a felt experience of emotion in the body especially seemed to make it accessible, as something concrete and actual that they could “work with”:

So, when the recording asked, “where is the pain?” I was taken aback, because it was actually somewhere [points to chest] and I would feel my tension. That was more specific for me, much more, it was more present. (Female, age 61)

This more actually felt experience of emotions made people take their emotions more seriously:

I started to take my [emotional] pain much more seriously, I took it seriously before, but it would be caught up in phrases and thoughts and now it was here [points to chest], I focused on it for a while and noticed much more clearly: what is pain? (Female, age 61)

In a more general sense, participants stated that after the program compassion was “just there in their system”:

It seems like the sting is gone. My body was full of splinters, and they’ve been taken out one by one. (Male, age 70)

Theme 3: Empowerment

When participants had practiced cultivating the kind container and been able to expose themselves to deeper layers of suffering, they seemed to experience a new empowerment in life:

I used to think I was very dysfunctional, but I notice more and more: actually, I function quite well. (Male, age 50)

This empowerment seemed to be the result of three different developments: an increase in self-compassion, confidence in the self and the future, and self-care.

Subtheme A: Compassion to Self

Addressing the “inner critic” is an important part of MBCL. The interview data indicate that the influence of this critical inner voice was being diminished, making room for self-compassion:

I think that trust comes from lowering the bar. From not continually being so harsh on myself, allowing myself to make mistakes. “Everything has to go well, otherwise I’m worthless.” There’s much more room now. (Female, age 37)

Subtheme B: Confidence in Self and the Future

As already alluded to in the last quote, for many participants, approaching and feeling the difficult within the container of kindness led to a sense of confidence. They seemed impressed by their own ability to bear difficult feelings:

I don’t want to idealize it, but that compassion training has certainly meant so much for my life. I was just thinking, jeez I’ve become so much softer towards myself. It’s all still very new and shaky… but the fact [that] I can see myself this way and I’m able to bear it all. Yes. (Female, age 48)

Subtheme C: Self-Care

The patients indicated that MBCL had particular added value in terms of its applicability in acute situations of deep suffering. They used compassion often as immediate self-care, taking agency in that situation. This increased trust and empowerment in turn seemed to affect self-care. A few participants decided to go for (longer) walks outside, one started decided to work fewer hours:

At the beginning of the year, I said I wouldn’t give up a second of work. And now I think, well 28 hours a week is just not doable for me. I’ve tried that five times and I’ve relapsed five times. If I want to take good care of myself, then I just have to stop that. (Female, age 51)

Theme 4: Common Humanity

For quite a few participants, it was helpful to hear others struggling with the same issues and hear about their insights. Three subthemes emerged within this theme.

Subtheme A: Relating to Others

Participants seemed to gain additional insight into how suffering is part of the human condition. Naturally, this common struggling with life comes up in MBCT as well, but as participants were explicitly geared towards deeper suffering in MBCL, the commonalities seemed to gain even more weight.

Subtheme B: Modelling by Others

Other participants modelling what compassion feels or looks like when applied successfully was revealing as well. For many, compassion was such an alien concept that they really needed this modelling by others.

Subtheme C: Compassion for Others

Additionally, participants reported an increased sense of compassion for others. Though the MBCL is geared towards self-compassion solely in the first half of the program, the second half includes practices on compassion for others.

Discussion

This study qualitatively explored the experienced additional value of compassion training (MBCL) after MBCT in a group of recurrently depressed patients. Participants seemed able to incorporate the program and gain substantial benefits in terms of (self-)compassion and self-care. Specifically, MBCL seemed to enable approaching and being with difficult emotions and thoughts with kindness. The program also seemed to cultivate skills to immediately apply kindness in difficult situations.

It seems that the crucial element is the practice in (self-)compassion itself: Through various practices, step-by-step, gentle guidance is offered through difficult experiences, creating a bedding of safety and kindness to facilitate exposure. This bedding of safety and kindness and being encouraged to set their own pace seemed to support participants in actively responding to suffering, applying that response in difficult or painful daily life situations. The wish to respond to the suffering is in accordance with the definition of compassion, which emphasizes recognizing suffering combined with a wish to alleviate it (Strauss et al., 2016).

In MBCL, the practice is mostly experiential and not cognitive. Participants are stimulated to primarily attend to their direct experience, be it pleasant or unpleasant, rather than think about it. We may view this as a type of exposure. In MBCT, exposure to the difficult is also practiced. When comparing the two however, some participants stated that during MBCT, they had been able to remain in more “neutral” observing, as it lacked the active invitation to approach the difficult. This made it possible to continue not engaging with strongly rooted avoidance patterns. This seems in congruence with a basic assumption underlying compassion focused therapy, namely, that the target population is not only highly self-critical, but also highly fearful of both negative and positive emotions, i.e., that resistance to working with difficult experiences is partly rooted in fear of positive emotions surfacing as well as negative ones (Gilbert, 2009; Gilbert et al., 2013). This resistance is in line with what participants describe in our qualitative study: They seemed to meet a lot of resistance in the first couple of sessions. They stated that exposure to experiences and emotions that they had suppressed for long times while being kind to themselves felt alien for them to do. In this light, we might formulate the added value of MBCL as giving participants the immediate tools and bedding in safety and self-compassion to facilitate exposure to deeper levels of suffering as well as “staying with” this suffering, which then leads to desensitization to fear. In sum, there seems to be a process of (a) a familiarization with kindness and safety, (b) an approaching and allowing the difficult, (c) a soothing and self-compassionate response, (d) more trust in being able to handle suffering, and (e) a growing experience of safety and kindness. These steps seem to mutually reinforce each other rather than forming a linear process, which fits in with Garland et al.’s (2015) mindful positive emotion regulation theory. This mutual reinforcing or upward spiral may account for the continued and further improvements observed over time, in particular the further improvement in quality of life (Schuling et al., 2020).

Limitations and Future Research

The study took place in the context of a large, well-powered randomized controlled trial (Schuling et al., 2020). The original developers of MBCL were involved in the teaching of the teachers, who were already highly experienced in teaching MBCT, and who were interviewed as well to ensure triangulation of the data. Participants were selected by means of purposive sampling. The research team was varied and consisted of six people with different levels of experience with mindfulness-based interventions and recurrent depression, ensuring different perspectives on the data. Interviewing and coding was carried out by two independent persons, ensuring the consistency of coding. Data collection ended only when saturation was reached, and no new codes or information emerged from the data.

One of the limitations of recruiting participants from our RCT sample is that it potentially caused selection bias: It was a clinical population, suffering from relatively severe depressive symptoms that had completed and possibly benefited from MBCT in earlier treatment. As a result, there may have been an underrepresentation of negative experiences. We used purposive sampling to minimize the potential bias however and asked about experiences with MBCT. Additionally, the primary researcher was also the interviewer, possibly resulting in social desirability. We also did not perform member checks. Furthermore, though the teachers were highly qualified, only two were involved, and they were both from the same center. As the researchers were also employed by this center, they may have been biased. Lastly, it is difficult to differentiate the effects of MBCT and its repeated application from the effects of MBCL. This is complicated further by the overlap in the two programs.

It seems that MBCL is a useful addition to MBCT in enhancing sustained recovery in adults suffering from recurrent depression. The four themes that emerged, (1) the container of kindness, (2) exposure to the difficult, (3) empowerment, and (4) common humanity, lead to interesting hypotheses. Explicit cultivation of compassion, i.e., establishing a container of kindness, seemed crucial in this population. This qualitative study focused on the potential of MBCL to add value to what was previously learned in MBCT in a sequential approach, i.e., offering MBCL after MBCT. There may be multiple ways of explicitly cultivating compassion however: Other ways could be to integrate explicit practices of compassion in MBCT or even to offer MBCL as a stand-alone intervention. Though offering compassion without first establishing a mindfulness practice may be overwhelming for this population, as Segal et al. (2012) expected when developing MBCT, it would reduce the trajectory for patients on their way to mental resilience and higher quality of life, if shown effective. As indicated by the second theme of actively approaching the difficult, in particular, populations with stronger or deeper-rooted suffering in terms of trauma or core beliefs regarding shame and guilt may benefit from MBCL first. Populations with such trauma or core beliefs often have accompanying passive and active avoidance behaviors, and for them, the cultivation of kindness in MBCT might be too implicit to facilitate opening to difficult emotions and learning to meet them with a compassionate attitude rather than with self-criticism and rejection. After assessing whether participants who suffer from trauma or who are highly self-critical indeed the ones to benefit most from MBCL, further exploration of predictors for the application of MBCL in recurrent depression could include the extent to which a participant is willing or aversive to working with emotional avoidance, related to the underlying trauma.

In addition, future research would benefit from exploring potential mediators drawn from the interviews and themes, such as the level of resistance to negative emotions, the fear of self-compassion or positive emotions, changes in agency, trust and self-confidence, and the extent to which common humanity was experienced throughout the intervention.

A fruitful approach to examining the various options in treatment trajectories may be to list empirically supported principles of change (and prediction) emerging from the research recommendations mentioned above, instead of focusing on trademarked therapies or other treatment packages, in line with Rosen and Davison (2003). This approach enables the creation of new, perhaps personalized trajectories by adding evidence-based components to already listed effective treatments.