Patients with major depressive disorder have a high risk of relapse/ recurrence. Maintenance antidepressant medication (mADM) is typically recommended to prevent relapse/recurrence. Many individuals, however, are unwilling to continue mADM for a long period, for example because of side effects. Besides, mADM is only effective as long as it is taken. So, a substantial proportion of patients prefer psychosocial interventions. Mindfulness-based cognitive therapy (MBCT) is an eight-week group intervention designed to help patients prevent depressive relapse/recurrence by developing nonjudgmental awareness of their automatic patterns of thinking and behaviour. Previous studies have shown that MBCT helps to prevent future depressive episodes in patients with recurrent depression, and that it can be a viable alternative to mADM. However, up till now, no studies have investigated whether the combination of MBCT and mADM is more effective than either of these treatments on its own. The current thesis was designed to assess the effectiveness of MBCT, mADM and its combination in patients with recurrent depression (3 or more previous episodes) in full or partial remission who had been using mADM for at least six months. Chapter 2 describes the design and protocol of the two RCTs that constitute the backbone of this thesis, together referred to as the ‘MOMENT’ study. Originally we intended to conduct a three-armed RCT of MBCT alone, mADM alone or MBCT+mADM, but this turned out not to be feasible since almost all patients who were interested in the study either wanted to participate in MBCT, or hold on to their medication. Therefore we conducted two parallel RCTs. Patients preferring MBCT participated in trial A comparing the combination of MBCT+mADM to MBCT alone, i.e. with discontinuation of mADM. Patients preferring mADM participated in trial B comparing the combination of MBCT+mADM to mADM alone. This enabled us to acknowledge patients’ preferences while maintaining the experimental rigour of randomisation. An advantage of this new approach was that the study was more in line with clinical practice, as particular treatment preferences are very common. In addition, acknowledging patient preference is widely recognized as an important aspect of evidence-based practice (American Psychiatric Association, 2010). The primary finding of the current thesis is that patients who discontinued their mADM after having participated in MBCT had nevertheless an increased risk of relapse/recurrence of depression, even having stated a preference for MBCT rather than mADM (chapter 3). In the intention-totreat sample, relapse rates were 54% for the discontinuation group and 39% for the combination group. In the per-protocol sample these numbers were even higher: 69% and 46%, respectively. Discontinuation seemed much more challenging than we had anticipated, resulting in a large percentage of patients (47%) who did not adhere to the study protocol, i.e. who did not proceed with the process of discontinuing their mADM to zero milligram. Thus, our hypothesis that MBCT followed by discontinuation of mADM would be non-inferior to the combination of MBCT and mADM was rejected. However, the overall course of depression severity during the 15-month follow-up period was similar in both groups, although a small but significant increase of depressive symptoms was observed in the discontinuation group at 3-month follow-up. At that point, patients in the discontinuation group were in the middle of tapering their medication, which may have led to elevated symptom levels. Probably related to this, we observed that besides the fact that almost half of the patients did not complete or even start the discontinuation process, those who did fully discontinue often restarted their mADM within the study period (56%). Other secondary outcomes, including the number, duration and severity of depressive episodes and quality of life during follow-up, did not differ between the groups. Another hypothesis was rejected in chapter 4, where we showed that adding MBCT to mADM did not further decrease the risk of relapse/recurrence compared with mADM alone (36% for MBCT+mADM and 37% for mADM alone). The groups also did not differ on any of the secondary outcomes. The relapse/recurrence rates in the per-protocol analysis (MBCT+mADM 39% and mADM alone 48%) suggested that there may be a small benefit for adding MBCT to mADM in patients adhering to both interventions. However, a sample size of about 500 participants per group would be needed to detect such a difference with 80% power. One of the possible explanations for the absence of an additional effect of MBCT was that in this trial only patients with a preference for mADM were included, whereas those preferring MBCT were included in the previous trial (chapter 3). The possible influence of preference was further explored in chapter 5, where we hypothesized that patients preferring MBCT would benefit more from MBCT than those preferring mADM, in terms of adherence to the combination of treatments (MBCT+mADM) and its outcomes (risk of relapse/recurrence, severity of (residual) depressive symptoms and quality of life). However, neither preference type (MBCT versus mADM) nor preference strength (the score on a questionnaire about expectations of both treatments) affected adherence or outcomes. Again, our findings were in contrast with our expectations. In chapter 6, we tested the hypothesis that higher levels of competence of the mindfulness teacher would be associated with better adherence (number of MBCT sessions attended), reduction of rumination and cognitive reactivity, improvement of mindfulness and self-compassion, and reduction of post treatment depressive symptoms and risk of relapse/recurrence in the year after MBCT. However, the results showed that despite a relatively wide range of competence levels in the sample of teachers in the MOMENT study (n = 15), teacher competence was not related to any of these outcomes. This study also showed that it is fairly difficult to rate teacher competence reliably. The inter-rater reliability was only moderate for most of the subscales of the instrument that we used, the Mindfulness-Based Interventions: Teaching Assessment Criteria (MBI:TAC; Crane et al., 2013). Nevertheless, our findings seemed rather robust as other possible indices of competence, including years of clinical practice, number of MBCT courses taught and amount of personal practice, did not predict patient outcomes either.