Can non-pharmacological interventions prevent ...

Clinical Psychology Review 39 (2015) 58–70

Contents lists available at ScienceDirect

Clinical Psychology Review

Can non-pharmacological interventions prevent relapse in adults who have recovered from depression? A systematic review and meta-analysis of randomised controlled trials Katherine Clarke ⁎, Evan Mayo-Wilson, Jocelyne Kenny, Stephen Pilling Centre for Outcomes Research and Effectiveness (CORE), Research Department of Clinical, Educational & Health Psychology, UCL, 1-19 Torrington Place, London WC1E 7HB, United Kingdom

H I G H L I G H T S • • • • •

We evaluated non-pharmacological interventions given after recovery from depression. The main meta-analysis incorporated 27 two-way comparisons and 2742 participants. CBT, MBCT and IPT reduced risk of relapse 12 months after recovery. The efficacy of MBCT following acute psychotherapy was largely untested. Longer term outcomes and the impact of acute treatment need further exploration.

a r t i c l e

i n f o

Article history: Received 6 August 2014 Received in revised form 8 April 2015 Accepted 15 April 2015 Available online 20 April 2015 Keywords: Depression Relapse Prevent Psychological therapy Long-term

a b s t r a c t Objective: To identify studies of non-pharmacological interventions provided following recovery from depression, and to evaluate their efficacy in preventing further episodes. Method: We identified relevant randomised controlled trials from searching MEDLINE, Embase, PsycINFO, CENTRAL, and ProQuest, searching reference and citation lists, and contacting study authors. We conducted a meta-analysis of relapse outcomes. Results: There were 29 eligible trials. 27 two-way comparisons including 2742 participants were included in the primary analysis. At 12 months cognitive–behavioural therapy (CBT), mindfulness-based cognitive therapy (MBCT), and interpersonal psychotherapy (IPT) were associated with a 22% reduction in relapse compared with controls (95% CI 15% to 29%). The effect was maintained at 24 months for CBT, but not for IPT despite ongoing sessions. There were no 24-month MBCT data. A key area of heterogeneity differentiating these groups was prior acute treatment. Other psychological therapies and service-level programmes varied in efficacy. Conclusion and implications: Psychological interventions may prolong the recovery a person has achieved through use of medication or acute psychological therapy. Although there was evidence that MBCT is effective, it was largely tested following medication, so its efficacy following psychological interventions is less clear. IPT was only tested following acute IPT. Further exploration of sequencing of interventions is needed. Systematic review registration number: PROSPERO 2011:CRD42011001646 © 2015 Elsevier Ltd. All rights reserved.

Contents 1.

2.

Introduction . . . . . . . . . . . . . . . . . 1.1. Characterising relapse . . . . . . . . . 1.2. Current approaches to preventing relapse 1.3. Intervention after recovery . . . . . . . 1.4. Aims of this review . . . . . . . . . . Method . . . . . . . . . . . . . . . . . . . 2.1. Eligibility criteria . . . . . . . . . . . 2.2. Search strategy . . . . . . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

⁎ Corresponding author. Tel.: +44 20 3108 3260; fax: +44 20 7916 8511. E-mail address: [email protected] (K. Clarke).

http://dx.doi.org/10.1016/j.cpr.2015.04.002 0272-7358/© 2015 Elsevier Ltd. All rights reserved.

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

59 59 60 60 60 60 60 61

K. Clarke et al. / Clinical Psychology Review 39 (2015) 58–70

2.3. Data extraction . . . . . . . . . . . . . . . . 2.4. Analysis . . . . . . . . . . . . . . . . . . . 3. Results . . . . . . . . . . . . . . . . . . . . . . . 3.1. Trial flow . . . . . . . . . . . . . . . . . . . 3.2. Study characteristics . . . . . . . . . . . . . . 3.3. Intervention groups . . . . . . . . . . . . . . 3.4. Meta-analysis . . . . . . . . . . . . . . . . . 3.5. 12 months . . . . . . . . . . . . . . . . . . 3.6. 24 months . . . . . . . . . . . . . . . . . . 3.7. Attrition . . . . . . . . . . . . . . . . . . . 3.8. Comparators . . . . . . . . . . . . . . . . . 3.9. Risk of bias assessment . . . . . . . . . . . . 3.10. Reporting bias . . . . . . . . . . . . . . . . 3.11. Quality of the evidence . . . . . . . . . . . . 4. Discussion . . . . . . . . . . . . . . . . . . . . . 4.1. Findings . . . . . . . . . . . . . . . . . . . 4.2. MBCT . . . . . . . . . . . . . . . . . . . . 4.3. CBT . . . . . . . . . . . . . . . . . . . . . 4.4. IPT . . . . . . . . . . . . . . . . . . . . . . 4.5. Service level programmes and other interventions 4.6. Who might benefit? . . . . . . . . . . . . . . 4.7. Strengths and limitations of this review . . . . . 4.8. Implications for future research . . . . . . . . . 5. Conclusion . . . . . . . . . . . . . . . . . . . . . Role of funding sources . . . . . . . . . . . . . . . . . . Contributors . . . . . . . . . . . . . . . . . . . . . . . Conflict of interest statement . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

59

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

61 61 61 61 61 62 62 63 63 63 63 63 63 63 63 63 63 66 67 67 67 68 68 68 68 68 68 68 68

1. Introduction

1.1. Characterising relapse

People who have recovered from an episode of depression (whether spontaneously or following the provision of treatment) are at an increased risk of becoming depressed again in the future. This risk can be as high as 60% for people who have experienced one episode, 70% in those who have had two episodes, and 90% in those who have had three previous episodes (American Psychiatric Association, 2000). These repeated episodes and the associated long-term suffering mean depression produces a large disease burden: the World Health Organisation estimates it to be the leading cause of disability worldwide (WHO, 2009) and is projected to rise to be the second biggest cause of burden of disease by the year 2020 (Murray & Lopez, 1997). Given the extent of the problem it is unsurprising that there are calls for the prevention and treatment of depression to be made a global public health priority (Whiteford et al., 2013). Intervening after recovery from depression can prevent relapse, but has been the subject of comparatively little research. A meta-analysis of maintenance medication following recovery (Geddes et al., 2003) included only 31 trials, and there has not yet been, as far as we know, a comparable review in the non-pharmacological domain (one that considers randomised trials of all non-pharmacological options following recovery). A recent review looked at psychological therapies delivered to adults aged 18–54 (Biesheuvel-Leliefeld et al., 2015) and included 25 studies. Most reviews prior to this have been limited to one psychological intervention, for example mindfulness-based cognitive therapy (Piet & Hougaard, 2011), continuation cognitive behavioural therapy (, Clark, & Jarrett, 2009) or brief psychological interventions (Rodgers et al., 2012). In comparison, a series of recent reviews undertaken for a major national clinical guideline for depression (NICE, 2009), identified over 1500 trials of acute treatment using antidepressant medication and over 200 trials of acute psychological interventions. This serves to highlight the importance of further work in the area, particularly research that complements the existing evidence for acute treatments, and that extends our focus beyond ‘recovery’, towards prevention of relapse.

Clarity in the characterisation of relapse in depression is a difficult task given the nature of the disorder. There is (as yet) no biomarker to indicate presence of an episode, and symptoms tend to fluctuate over time rather than having dimensional shifts where there is a clear demarcation between illness and normal functioning (Clark & Beck, 1999; Hersen & Beidel, 2011; Ingram, 1998). For this reason, the criteria and level of symptoms needed to signify an episode of depression can vary. Diagnosis is important in indicating when a person may require treatment, but equally a person with symptoms that do not meet criteria for an episode may still have impaired functioning, and is at a greater risk of illness than someone with low or no symptoms (Cuijpers & Smit, 2004). Taking into account this increased risk of further episodes of depression in those who have brief or incomplete recovery, Frank et al. (1991) identified four clinically important change points. They used the word ‘remission’ to indicate reduction of symptoms to below the diagnostic threshold for depression, and ‘recovery’ to indicate remission that lasted more than 6 months. They then differentiated between ‘relapse’, which was a diagnosable episode of depression occurring during remission, and ‘recurrence’ an episode that followed recovery. Although developed over twenty years ago with the aim of clarifying important points in the course of depression, these terms have not been universally adopted. This may be due to limited dissemination, but some have argued that the distinction between ‘relapse’ and a ‘recurrence’ is not clearly supported by evidence from intervention trials (Richards & Perri, 2010). Inconsistencies in the terminology used to describe the course of depression present a challenge when attempting to combine data from multiple research studies. Indeed, a recent review of the lifetime approach to depressive disorder promotes the use of distinct terms for relapse and recurrence, yet these seem to be of limited utility, with the authors reverting to the broader ‘relapse/recurrence’ summary term to describe the majority of research (Bockting, Hollon, Jarrett, Kuyken, & Dobson, 2015). In this paper we use the term ‘recovery’ to denote any period where a person no longer meets diagnostic criteria

60


for depression, the term ‘relapse’ to denote “any kind of significant deterioration in depression, which was preceded by clear improvement” (pp. 5 Richards & Perri, 2010). Without strong evidence of a dimensional shift between, say, a relapse at 5 months, and a recurrence at 7 months, it seems prudent to limit the number of categories we use. This also allows the inclusion of a broad range of studies, which may use varying terminology, in which we are able to consistently distinguish between the presence and absence of a depressive episode. 1.2. Current approaches to preventing relapse In contrast to medication, which seems to lose efficacy when discontinued (Anderson & Tomenson, 1995), acute phase psychological interventions may prevent relapse even when they are terminated at the point where a person has recovered (Gortner, Gollan, Dobson, & Jacobson, 1998). This is apparent with CBT (Hollon, Stewart, & Strunk, 2006), and there is a growing body of literature supporting the longterm impact of problem solving therapy given during the acute-phase (Nezu & Nezu, 2010). Perhaps these enduring effects are achieved through the amelioration of the causes of risk of relapse, or through the development of knowledge and skills that equip people to offset their risk (Hollon et al., 2006). After the acute phase of treatment there has been some move to divide the post recovery period into ‘stabilisation’ or ‘continuation’ and ‘maintenance’ phases. In practice however, there has often been little to distinguish between these post-recovery phases, with patients identified at an elevated risk of relapse typically staying on medication for at least a two year period (Geddes et al., 2003; Thase, 2012). It is interventions delivered after the acute phase, when patients are no longer depressed, that are the focus of this paper. 1.3. Intervention after recovery Continued pharmacological treatment reduces relapse following successful pharmacological treatment during an acute episode (Geddes et al., 2003). This is the most common form of post-recovery intervention, but not all patients benefit, side-effects may increase over time, and people who are unwilling to stay on medication forever are likely to relapse when they discontinue it (Anderson & Tomenson, 1995), even if discontinuation is gradual (Viguera, Baldessarini, & Friedberg, 1998). Psychological approaches are often preferred by patients (Priest, Vize, Roberts, Roberts, & Tylee, 1996), and tend to be timelimited. Psychological interventions have been designed for, or in some cases been adapted for use after the acute phase. One such intervention is mindfulness-based cognitive therapy (MBCT, Segal, Williams, & Teasdale, 2012), perhaps the most widely known psychological intervention developed specifically to prevent relapse. It was designed to be delivered after a person has recovered from depression, and can follow spontaneous recovery or acute treatment. It typically delivered in groups of 12 to 15 participants with two professionals in fortnightly sessions over three months. MBCT encourages people to process experience without judgement through mindfulness and meditation techniques, thereby helping to change their relationship with troubling thoughts and feelings (Segal et al., 2012). A meta-analysis of MBCT trials (Piet & Hougaard, 2011) suggests that it is effective (with a 34% reduction in risk of relapse compared to controls), particularly for patients who have had three of more previous episodes of depression (where the reduction was 43%). Cognitive behavioural therapy (CBT) may also work through equipping people with specific skills to reduce their likelihood of relapse. As an acute treatment it aims to modify thoughts and behaviours in a way that alleviates key features of depression, such as identifying and challenging negative automatic thoughts and overcoming avoidance and reduced activity (Beck, 2011). There have been variations of CBT adapted to target relapse prevention, either as a continuation of prior acute treatment or as a standalone intervention. In contrast to MBCT

where the delivery of the intervention follows a common format, the duration, mode of delivery (individual, group, self-help), and frequency of CBT has varied. Vittengl, Clark, Dunn, and Jarrett (2007) reviewed the effectiveness of CBT to prevent relapse, both from studies which terminate treatment at recovery, and studies which continued CBT after this point. They estimate that continuing with CBT reduces risk of relapse by 14% compared with discontinuing. Although this estimate of effect is lower than in the aforementioned MBCT meta-analysis, there are key differences in the acute treatment received and the length of included trials, as well as the treatment of missing data by reviewers. Interpersonal psychotherapy (IPT) (Klerman & Weissman, 1994) can also be used after recovery with reduced frequency of sessions. The model of delivery and objectives of IPT remain essentially the same as in the acute phase, focussing on strategies to deal with interpersonal and social role problems which are held to be casual in the development and maintenance of depression (Klerman & Weissman, 1994). IPT ‘booster sessions’ may reduce relapse by providing continued support after the period where contact with a psychologist would typically end. As part of a wider review, Cuijpers et al. (2011) identified four studies which compared combination of maintenance IPT and pill placebo with pill placebo alone. Maintenance IPT was more protective against relapse than pill placebo. Other non-pharmacological approaches have sought to prevent relapse by continuing to provide enhanced monitoring and careplanning regardless of whether patients continue to be symptomatic (Gilbody, Bower, Fletcher, Richards, & Sutton, 2006; Smit, Tiemens, & Ormel, 2007). Similarly there has been some research into the impact of continued monitoring of symptoms and coordination of care postrecovery (Unützer et al., 2002). Sequencing of interventions and support, as well as long-term management of care from different providers may have important effects (Wells et al., 2000) as well as interventions used in routine clinical practice such as exercise, symptom monitoring, and building social support. There are a broad range of intervention options for those who have recovered from depression, and a review of all randomised trials of these interventions may help to highlight what these are and whether they are effective. 1.4. Aims of this review This review will gather evidence about non-pharmacological interventions to prevent relapses in those who have recovered from depression. We firstly aim to describe the research that exists in this field, and highlight any potential gaps in evidence. We then aim to examine the efficacy of these interventions. 2. Method We conducted a systematic review and meta-analysis following a pre-specified protocol (Clarke, Pilling, Kenny, & Mayo-Wilson, 2011). 2.1. Eligibility criteria We included trials where adults in full or partial remission from depression were randomised to either • A non-pharmacological intervention aimed at reducing relapse of depression (this aim could be implied through the inclusion of depression relapse in the study's outcome measures), or • Any comparator condition, • And were followed up for a minimum of 1 year after randomisation.

‘Remission’ was considered to be any reduction in symptoms to below a diagnostic threshold or other threshold on a validated assessment tool. We excluded studies with fewer than 5 people in each arm of the study. We only included studies which randomised participants


who had recovered. This meant we excluded studies that randomised to acute treatment even if they then followed them up after recovery. 2.2. Search strategy We searched Embase, MEDLINE, and PsycINFO (using the Ovid interface), CENTRAL, and ProQuest from inception to 04 December 2013 using a sensitive RCT filter without language or publication status restrictions. The search was focussed on all terms associated with two concepts: ‘depression’ and for ‘relapse/recurrence/long-term’ (see Appendices A, B, and C for full search strategies). We did not search for particular types of intervention, so as to identify all randomised trials which looked at a depression intervention without assuming, a priori, what research would exist. Tests of pharmacological interventions were removed during screening. Citations were screened by two authors (KC and JK) independently and disagreements about inclusion were resolved by discussion with a third author (SP or EMW). We also checked reference lists and citations (using SCOPUS) of included studies and some excluded studies, and contacted study authors of all included studies. 2.3. Data extraction Data from each study were extracted by one review author and checked for errors by a second. We extracted information about the content and delivery of the intervention being tested, as well as trial start dates and locations. We extracted information about the clinical history of the population where available, as this is linked to baseline risk of relapse. Where possible we extracted information about how many people had relapsed, not relapsed, or not provided sufficient information, when this was measured relative to randomisation, and how it was defined and measured. Additional data or information about study methods were provided by study authors for 45% of the included studies (Bondolfi et al., 2010; Godfrin & van Heeringen, 2010; Holländare et al., 2013; Jarrett et al., 2000, 2001; Katon et al., 2001; Klein et al., 2004; Kuyken et al., 2008; Ma & Teasdale, 2004; Reynolds et al., 1999, 2006; Wilkinson et al., 2009; Williams et al., 2013).

61

medication or non-specific support, as well as evaluation-only controls. A broad range of controls were acceptable as, despite having different descriptions and setting for delivery, there could be significant overlap in the care received between control group types (for instance evaluation-only controls could seek treatment from primary care professionals, which could include medication/clinical management/ non-specific support). We conducted meta-analyses using Review Manager (RevMan) Version 5.2 (http://ims.cochrane.org/revman) for all of the prespecified time points where sufficient study data were available. Relative risks and 95% confidence intervals (CIs) were calculated and combined using Mantel–Haenszel methods. We assumed that study drop-outs had relapsed in order to obtain conservative measures of treatment effect and to minimise methodological differences in the meta-analysis. We adopted this approach only as some studies reported their outcomes on an intention to treat basis, with drop-outs assumed to be relapsed, and we could not obtain data from studies to form a completer analysis. This approach was the only one which allowed a consistent approach across all studies. We assessed risk of bias for included studies, and the quality of the evidence available. Trials were assessed using the Cochrane Risk of Bias Tool (Higgins, Green, & Collaboration, 2011b) with regard to our outcome of interest, relapse of depression. Confidence in the outcome for each subgroup was rated according to GRADE criteria (Balshem et al., 2011) and incorporated into the interpretation of results. 3. Results 3.1. Trial flow We identified 20,531 potentially relevant citations and retrieved 389 papers for full-text review (Fig. 1). Of these, 29 trials were eligible for this review, and 22 trials were included in the meta-analysis. Seven of the trials either did not report or provide usable data on request, or did not have a control group that fitted into our definition. 3.2. Study characteristics

2.4. Analysis We grouped trials if they tested the same intervention according to the study authors' descriptions of interventions and according to discussion within the review group. Where there were several trials of the same type of intervention, we conducted a meta-analysis of numbers who had relapsed in intervention versus control groups. Studies with more than two relevant arms were split into two-way comparisons as follows: • When there was more than one eligible intervention, the control group was divided for analysis. • Where there was more than one eligible control group they were combined for analysis. • Factorial trials were treated as separate two-way comparisons.

Relapse was defined a change in the status of a patient, from recovered (fully or partially) to depressed. This had to be measured through according to meeting accepted diagnostic criteria on the basis of a clinical interview, or by meeting a diagnostic threshold on a validated measure. If study authors used an unvalidated or self-report measure of relapse alone, this was not included in the meta-analysis. Service utilisation alone (e.g., admission to hospital) was not taken as an indicator or relapse. We defined a control arm as any intervention that did not fall under our focus (non-pharmacological relapse prevention). This could include active interventions such as clinical management, antidepressant

The 29 eligible trials were conducted in the USA (n = 12), the UK (n = 6), the Netherlands (n = 3), Italy (n = 2), Canada, Denmark, Germany, Sweden and Switzerland (all n = 1). They included 4216 participants; the size of individual trials varied (range 31 to 485 participants, median 124), with a ‘mood management’ trial of 485 participants, (Van der Meer, Willemsen, Smit, Cuijpers, & Schippers, 2010) and four CBT trials with less than 50 participants. Most trials included a majority of female participants, and adults of any age, although four trials limited inclusion to adults aged 55 years or over. Key characteristics of the studies are summarised in Table 1. All our included trials randomised participants who were not currently depressed to the intervention or control arms. The acute treatment that preceded this point is detailed in Table 1. Some trials recruited people who had previously recovered from depression outside of the research context. Other trials delivered acute treatment in linked research trials and then randomised people to receive relapse prevention or a comparator. To examine these studies uniformly, we have used the randomisation to relapse prevention as our baseline. Control conditions included ‘evaluation-only control’, and ‘treatment as usual’, as well as ‘active’ interventions (which we defined as interventions specifically administered for the trial). However, there was considerable overlap in the content of these controls; for example, antidepressant medication was used in control groups described as ‘treatment as usual’ and as ‘clinical management’, as well as those described as ‘medication’ arms. There was a large difference in control event rate for relapse across the trials (for 12-month data, range = 0.2–0.79, IQR = 0.41–0.63).

62


20, 531 records identified from electronic database searching Medline/PsycInfo/EMBASE= 16, 264 ProQuest= 1197 CENTRAL=3070 4, 762 duplicates removed 15, 769 records screened by title and abstract

15, 180 records excluded: the trial in the record was not eligible, or was eligible but was included under a different record

389 full-texts assessed for eligibility 360 full-texts excluded because they: -included participants who had never before been depressed, or who were acutely depressed with no recovery, -were included under another reference, -were not adequately randomised/controlled, -had less than 12 months follow-up/were incomplete.

29 trials included 24 records identified through correspondence with authors of included trials.

347 records identified through hand-searching references and citations of included trials.

0 additional trials identified

29 trials included in the narrative review 22 trials included in the meta-analysis Fig. 1. PRISMA flowchart of trial flow.

3.3 . Intervention groups The majority of included studies evaluated psychological therapies: cognitive behavioural therapy (CBT) (studies 1–10 in Table 1); mindfulness-based cognitive therapy (MBCT) (studies 11–17); and interpersonal psychotherapy (IPT) (studies 18–21). All were structured, time-limited interventions, except for IPT which was delivered monthly throughout the entire duration of follow-up. The next largest group of interventions was of service-level programmes (studies 22–25 in Table 1). One study tested a primary care relapse prevention programme that consisted of 2 visits and 3 phone calls with a depression specialist aimed at enhancing medication adherence, monitoring potential prodromal symptoms and developing a relapse plan (Katon et al., 2001). Other trials evaluated a stepped care programme (Apil, Hoencamp, Judith Haffmans, & Spinhoven, 2012), referral to a specialised outpatient intervention programme (Hansen et al., 2012), and a web-delivered care-management and selfmanagement programme (Hunkeler et al., 2012).

The remaining trials tested psychotherapies that did not fall under the scope of the groups already mentioned (studies 26–29 in Table 1). One trial tested ‘cognitive–behavioural analysis system of psychotherapy’ (CBASP), a psychological intervention usually used to treat chronic depression (Klein et al., 2004). This has similarities to CBT but did not fulfil our criteria to be grouped with CBT. Similarly a trial of an early variant of IPT was not included in the IPT subgroup (Weissman, Kasl, & Klerman, 1976). Two trials looked at mood management additions to smoking cessation treatment (Brown et al., 2001; Van der Meer et al., 2010). 3.4. Meta-analysis Relative rather than absolute risk reduction was used in the metaanalyses because of the variation in control event rate (for 12-month data, range = 0.2–0.79, IQR = 0.41–0.63). All comparators were included within the same meta-analysis and the effect of the ‘control type’ explored post-hoc.


3.5. 12 months For the four intervention groups (CBT, MBCT, IPT and servicelevel programmes) data were available for 26 two-way comparisons at 12 months. Compared with all controls, the average risk of developing a new episode of depression by 12 months was reduced by 25% for CBT (RR = 0.75; 95% CI 0.64 to 0.89, I 2 = 8%), 21% for MBCT (RR = 0.79, 95% CI 0.69 to 0.91, I 2 = 0%) and 22% for IPT (RR = 0.78, 95% CI 0.65 to 0.95, I2 = 0%). These subgroups were not statistically different at 12 month follow-up (chi 2 = 0.21, df = 2 (P = 0.90), I2 = 0%) (Fig. 2). Service-level programmes were clinically heterogeneous and did not reduce the risk of developing a new episode of depression by 12 months (RR = 1.00, 95% CI 0.81 to 1.23, chi2 = 1.28, df = 2 (P = 0.53); I2 = 0%). Results for specific interventions within this group included a stepped care programme (RR = 1.26, 95% CI 0.74 to 2.15), a specialist outpatient programme (RR = 0.85, 95% CI 0.57 to 1.28) and a ‘relapse prevention programme’ (RR = 1.01 95% CI 0.77 to 1.33). Data were not available for the subset of relevant participants within a study of a web-delivered care-management and self-management programme (Hunkeler et al., 2012). One trial, of continuation CBASP, had a large effect but small number of participants (N = 82) and a large confidence interval (RR = 0.12, 95% CI 0.02 to 0.91). Data were not available for the remaining trials.

63

in studies that included drop-outs as relapsed. If they did not provide separate numbers for participants who identified as relapsed, or for whom there was no information, this meant we could not assess whether there was a differential drop-out between intervention and comparison groups. 3.10. Reporting bias Reporting bias could not be assessed statistically because there were too few studies and insufficient variation in study size. This review is at risk of reporting bias because most included trials were not prospectively registered, so it is possible there are other trials that have been conducted but remain unpublished, and a small number of medium-to-large studies could change the conclusions. Included studies, particularly in the IPT group, were conducted by only a few researchers. 3.11. Quality of the evidence The GRADE rating criteria (Balshem et al., 2011) were applied for relapse for each type of intervention. Evidence was generally of low quality, which indicates that results may change following further research (Guyatt et al., 2008). Reasons for downgrading were inconsistency of results, indirectness of evidence, imprecision and reporting bias.

3.6. 24 months 4. Discussion Data for relapse at 24 months were available for seven of the ten CBT studies, all six of the IPT comparisons, but none of the other trials. The effect for CBT at 24 months (Fig. 3) was similar to the effect at 12 months but with greater heterogeneity (RR = 0.72, 95% CI is 0.57– 0.91, I2 = 63%), but the effect of IPT was not sustained (RR = 0.92, 95% CI is 0.81–1.05, I2 = 0%). 3.7. Attrition Fifteen trials (1421 participants) reported drop-out at 12 months, which was the time-point with the most available data. There was no evidence of an important difference between intervention and comparison groups (RR = 1.15, 95% CI 0.85 to 1.52, I2 = 0%). 3.8. Comparators We also used 12-month data to assess the impact of treatmentas-usual or active comparators due to the limited data at 24-months. At this time point, interventions reduced risk of relapse compared with treatment as usual (RR = 0.79, 95% CI 0.70 to 0.91) and active comparators (RR = 0.77, 95% CI 0.68 to 0.87). There was no evidence of a difference between these groups (chi2 = 0.12, df = 1 (P = 0.73), I2 = 0%).

4.1. Findings This review suggests that psychological interventions (specifically MBCT, CBT and continuation IPT) delivered when an adult has recovered from an episode of depression may reduce the risk of experiencing another episode within the following year by 22%. The baseline risk of relapse varied greatly between studies (for 12-month data, range = 0.2–0.79, IQR = 0.41–0.63), but as an indication of the clinical utility of these interventions, we estimate a number-needed-to-treat (NNT) of 8.33 for the study population with the lowest baseline risk (Paykel et al., 1999), and an NNT of 4.76 for the study population with the highest baseline risk (Reynolds et al., 1999). This estimate may be conservative given that the comparator groups were often active interventions, and given our imputation of drop-outs as relapsed. Considering the suffering associated with an episode of depression, and the potential for an extended period of recovery to itself have a protective impact, this represents an important effect. However, we found that longer-term evidence for the efficacy of these psychological therapies was incomplete: there was evidence for a sustained effect for CBT at two years, but not for IPT, and no evidence for MBCT. In contrast the service-level interventions we identified did not seem to have a reliable impact on relapse at 12 months.

3.9. Risk of bias assessment 4.2. MBCT To assess risk of bias in a standardised and systematic way across studies, we used the Cochrane Collaboration Risk of Bias Tool (Higgins et al., 2011a). Full risk of bias assessments for each study is summarised in Appendix D. Random sequence generation and allocation concealment were either at low risk of bias or we do not have sufficient information to assess risk. It was not possible to conceal the intervention from participants or providers those delivering it, so all studies were at high risk of bias for masking per se. The majority of studies used masked outcome assessors. Half of the studies had a low risk of bias from attrition; however half were unclear or high risk of bias because of uncertainty about drop-outs, or because of high levels of attrition. There was a high risk of bias for selective outcome reporting

There was considerable homogeneity in the structure and content of MBCT with all trials following the same manual (Segal, Williams, & Teasdale, 2002) and offering eight weekly two-hour sessions to groups of between 9 and 15 people. However, there was some variation in the number of additional sessions offered, from none (Godfrin & van Heeringen, 2010) to monthly sessions for the trial duration (Segal et al., 2010). There was also a fairly high level of consistency in the way MBCT trials were structured and conducted, with two studies explicitly replicating Teasdale et. al's (Teasdale et al., 2000) methods (Bondolfi et al., 2010; Ma & Teasdale, 2004). However this consistency in methods has meant that, in contrast to the CBT studies, MBCT has largely been tested in people who have

64

Table 1 Study characteristics. Study Lead author, number year CBT 1

N

Mean age (SD)

Mean previous episodes (SD)

Acute treatment

Pre randomisaton/ eligibility

% on ADM at baseline

Relapse prevention intervention

Intervention length

Comparison

Criteria used to measure relapse

Weeks of follow-up

*3–4 (IQR ±

Any

2+ recent episodes, remission for 10

51

Group CT + TAU

2 h/week for 8 weeks

TAU

DSM-IV diagnosis of MDE (SCID)

104

100

CBT + discontinuing ADM

40 min/fortnight for 20 weeks

Clinical management +

RDC-defined MDE

104

CBT + ADM

30 min/fortnight for 20 weeks

discontinuing ADM Clinical management +

RDC-defined MDE (CID)

104

187 44.7 (9.5)

Fava (1994)

43 46.1 (2.8)

----

45 46.0 (10.8)

3.6 (0.8)

Holländare et al. (2013)

84 45.3 (12.8)

5.96 (9.1)

5

Jarrett et al. (2000)

31 41.2 (10.5)

6


84 42.7 (10.5)

7


2 3

ADM

100

Psycotherapy + /ADM

Partial remission

50

----

^^^CT

3.4 (0.2)

^^^CT

241 42.7 (11.8)

*4

^^^CT

Atypical features and treatment response Responders who completed 20+ CT sessions Responders who were high risk

Paykel et al. (1999)

158 43.27 (10.9)

*37% had 3+

9

Stangier et al. (2013)

180 48.6 (11.6)

10

Wilkinson et al. (2009)

Bondolfi et al. (2010)

8

MBCT 11

12

13 14

15

Godfrin and van Heeringen (2010) Kuyken et al. (2008) Ma and Teasdale (2004) Segal et al. (2010)

ADM

weeks–2 years Residual symptoms 3–5 months remission

4

Fava (1998)

3.8)

Internet-based CBT

Internet modules & 10

0

Continuation CT

weeks of email support 10 sessions over 35 weeks

0

Continuation CT

4× 1 h/fortnight then 6× 1 h/month

Evaluation only

0

Continuation CT

4× 1 h/fortnight then 6× 1 h/month

Clinical management +

100

CT + clinical management

2–18 months remission, high residual symptoms

7.4 (8.3)

Any, no CBT in the past year

Maintenance CBT

45 74.0 (7.3)

3.0 (1.9)

ADM

75 3+ previous episodes, 8+ weeks in remission 100 Remission for 2+ months and remain on ADM

60 ----

*4 (range 3–18)

ADM, CBT b 5 times

MBCT + TAU

106 45.7 (10.6)

----

Any

0 3-months withdrawal from ADM 68 3+ previous episodes, 8+ weeks in remission

123 49.2 (11.2) 75 44.5 (8.9)

6.4 (3.0)

In full or partial remission 12+ weeks remission with no ADM 7 months clinical remission

MBCT + tapered ADM 8× 2 h/week + 4 sessions each 3 months MBCT + TAU 8× 2 h/week + 2 sessions at 1 and 6 months

84 44.1 (10.9)

3 (2)

4.7 (2.3)

^^ÂDM

100 0

100

Group CBT

Evaluation only

fluoxetine or placebo 16 sessions over 20 weeks + Clinical management

8+ weeks of ADM

ADM for 6+ months ADM CBT b 5 times

ADM Non-specific email support

2 sessions 6 and 14 weeks later 16 50-minute sessions over Manualized 8 months psychoeducation 8× 1 h 30 min over 10

TAU

weeks

MBCT + TAU

Diagnosis of depression (SCID) 114 RDC-defined MDD OR Patient had more treatment DSM-IV diagnosis of MDD (LIFE 25)

104

104

DSM-IV criteria for MDD (SCID)

140

DSM-III-R criteria for MDD for at least 1 month AND Minimum symptom score

68

DSM-IV diagnosis of MDE (LIFE)

87

MADRS score of 10+ at 6/12

52

months

8× 2 h/week + 4 sessions each 3 months

TAU

DSM-IV diagnosis of MDE (SCID)

60

2 h 45 min/week for 8

TAU

DSM-IV diagnosis of MDE

56

Maintenance ADM

DSM-IV diagnosis of MDE (SCID) DSM-IV diagnosis of MDE (assessments “modelled on the SCID”) 16+ on the HRSD-17 assessed twice then criteria for MDD measured with SCID, OR LIFE

65

weeks

MBCT + tapered ADM 8× 2 h/week + 1 day + optional 1 h/month for trial duration

TAU

ADM /Clinical management + placebo

52

86


Bockting et al. (2005)

16

Teasdale et al. (2000)

145 43.3 (10.0)

*median is 3 or 3.5

ADM CBT b 5 times

17

Williams et al. (2013)

274 43 (12)

80% had ≥5

Any, exclude successful CBT

0

MBCT + TAU

8 X 2 h/week + 4 more sessions over trial duration

TAU

DSM-III-R criteria for MDE (SCID)

60

≈45

MBCT + TAU

Individual pre-class, then 8× 2 h/week

Cognitive psychological education + TAU or TAU

Criteria for MDE ≥ 2 weeks (SCID)

61

20+ weeks relatively symptom-free ^^ÎPT (+ ADM 12–24 weeks if needed) treatment ^^^Nortriptyline Clinical response and IPT

100

mIPT (+ placebo/imipramine) IPT

Monthly sessions

RDC-defined MDD AND Medication clinic (+placebo/imipramine) independently confirmed minimum symptom score None DSM-IV diagnosis of major depression Medication clinic (+ RDC-defined MDE placebo/nortriptyline) (structured psychiatric interview) Medication clinic (+ DSM-IV diagnosis of paroxetine/placebo) depression (SCID)

12+ weeks remission with no ADM 3+ recent episodes, remission for 8 weeks

IPT 18

19 20

21

128 40.2 (10.9)

----

Frank et al. (2007) Reynolds et al. (1999)

131 30.6 (10.3) 124 ----

----

Reynolds et al. (2006)

116 76.8 (5.7)

*59% had 1

^^^Paroxetine and IPT

Clinical response

100

136 65.6 (8.3)

4.4 (7.0)

Any

No current psychotherapy

46

Stepped care programme

Symptom dependent for trial duration

TAU

Specialised outpatient programme Web delivered careand self-management programme Relapse prevention programme

Symptom dependent for 52 weeks ----

TAU

Varied for 52 weeks

TAU

LIFE

52

CBT+ smoking cessation treatment CBASP

8× 2 h over 6 weeks

Smoking cessation treatment Assessment only

Not measured

52

DSM-IV diagnosis (checklust) AND Minimum symptom score AND Clinical interview Not measured

52

Not measured

87

23 24

25

Others 26 27

28

29

24 100

Hansen et al. (2012) Hunkeler et al. (2012)

268 *med is ---38.4-38.7 66 -------

Hospitalisation

Discharged

----

----

----

----

Katon et al. (2001)

386 46.0 (12.6)

0.8 (0.4)

ADM

High risk of relapse 81

Brown et al. (2001) Klein et al. (2004)

179 45.1 (9.3) 82 45.0 (11.4)

----

Smoker

0

2.5 (1.6)

Any (no current treatment) ^^^CBASP

Incomplete remission

0

Van der Meer et al. (2010)

485 43.8 (10.0)

*47% had 1

Any

Smokers with depression N1

0

Weissman et al. (1976)

150 39 (----)

----

^^ÂDM

month ago Clinical response to 6-8 weeks acute treatment

100

Weekly/twice monthly/monthly sessions 50 min/month

IPT (+ placebo/nortriptyline) 45 min/month IPT (+ paroxetine/placebo)

Monthly sessions for 52 weeks

Mood management + 1× 30 min and 9× 15 min over 91 days

smoking cessation counselling Therapy (+ amitriptyline/ placebo/no drug)

1–2 sessions/ week for 8 months

Usual specialty mental health care

Smoking cessation counselling Minimal contact (+ amitriptyline/ placebo/no drug)

104 156

104

52 DSM-IV diagnosis of MDE (MINI 5.0) OR Medical records show any additional treatment Not measured (they looked 4-5 years at hospitalisation) PSR every 2 weeks using LIFE ----


Care programmes 22 Apil et al. (2012)

----

^^ÂDM and decreasing IPT

156

Frank et al. (1991)

52

*= Not ‘mean (SD)’, —— = not measured/reported, and ^^^ = in earlier stage of the trial. NB The acute treatment column shows any treatment received before randomisation to relapse prevention or comparison. If this was delivered within an earlier stage of the same relapse prevention trial it is marked ^^^. ADM = antidepressant medication, CBASP = cognitive behavioural analysis system of psychotherapy, CBT = cognitive behavioural therapy, CT = cognitive therapy, MBCT = mindfulness-based cognitive therapy, TAU = treatment as usual. Criteria used to measure relapse: CID = Clinical Interview for Depression, DSM = Diagnostic Statistical Manual, LIFE = Longitudinal Interval Follow-up Evaluation, MDD = Major Depressive Disorder, MDE = Major Depressive Episode, RDC = Research Diagnostic Criteria, SCID = Structured Clinical Interview for DSM Disorders, MADRS = Montgomery–Asberg Depression Rating Scale, HRSD, and PSR = Psychiatric Status Rating.

65

66


responded to medication, and not in those who have experienced or respond to psychological intervention. Four trials explicitly exclude those who have experienced CBT, and of those that did not, two required participants to have remitted on antidepressant medication (see Table 1). There are several issues with this. Firstly, effectiveness may be limited to participants who recovered by means of pharmacological treatment. Secondly, the efficacy of MBCT may be a nonspecific effect of psychological treatment. Some support for this can be found in a recent study (Williams et al., 2013) which found no difference in the efficacy of MBCT compared to a version of MBCT with the meditation component removed. 4.3 . CBT There was considerably more variation in the CBT relapse prevention trials in terms of the nature and duration of the intervention, as well as in the inclusion of participants who had previous psychological or pharmacological treatments. Generally the authors of CBT trials had adapted CBT for relapse, produced their own manuals, and used this in their own trials. Sessions were between 30 and 90 min and spanned

periods of 8 to 35 weeks. Most studies treated people individually, but two trials tested the intervention delivered to groups of 8 people (Bockting et al., 2005) or 4 people (Wilkinson et al., 2009). Assuming one therapist delivered each group, the therapist contact hours per person in these studies was 2 and 3 h respectively. Group formats are often chosen in an attempt to reduce costs, and may particularly suitable within relapse prevention where patients have no current pathology (Bockting et al., 2005). One study tested a CBT intervention that placed relatively little demand on therapist time. Participants were provided with a series of internet-based self-learning modules and given email support over a period of 10 weeks (Holländare et al., 2013). Although a formal cost analysis is beyond the scope of this review, it seems likely that this would be cheaper than, say, the 10 sessions of individual face-to-face CBT offered in the three Jarrett's studies (5, 6 and 7 in Table 1), although reduction in cost does not necessarily mean an intervention is more cost-effective. The internet based CBT was effective in this individual trial, with a risk reduction of 58%, but there was a large level of uncertainty around this estimate (95% CI 15% to 79%] and replication would be needed in further studies.

Fig. 2. Relapses in first 12 months in intervention vs. control groups for CBT, MBCT and IPT.


67

Fig. 3. Relapses in first 24 months in intervention vs. control groups for CBT and IPT.

4.4 . IPT IPT trials adopted a common approach to the prevention of relapse by offering maintenance sessions of IPT, with little in the way of adaptation for relapse. All trials provided monthly sessions for the duration of follow-up (either two or three years), except for one (Frank et al., 2007) that randomised participants to different frequencies of IPT: weekly, twice monthly, or monthly for the two-year trial. Two included trials focussed on older adults (age 60 + and 70 + in Reynolds et al., 1999, 2006 respectively). As older adults are at an increased risk of relapse (Mitchell & Subramaniam, 2005, Arean & Ayalon, 2010), the majority of patients within this meta-analysis could be considered high risk. Whilst the reduction in relapse was similar to that seen for CBT and MBCT at one year follow-up any benefit associated with IPT was not present at two years despite the intervention being maintained over the two year period. To give a numerical indication of the clinical impact of IPT in this group at 2 years, there is an NNT of 14.29, but with a large degree of uncertainty around this estimate (95% CI 6.25 to 33.33). All IPT trials were conducted by related research groups from the University of Pittsburgh (USA). With an absence of consistent registration of protocols, there may be a risk that reporting bias is having an impact on the results of this group. 4.5. Service level programmes and other interventions The fourth group of interventions reviewed were service-level programmes, for which we did not find evidence of efficacy at one year. This group consisted of only four trials, each of a different approach, for only three of which were able to identify efficacy data. The remaining trials were of interventions not tested elsewhere (see studies 22 to 29 in Table 1). They did not appear to be effective, other than in one trial of CBASP (Klein et al., 2004). Of 42 participants given CBASP, one relapsed, compared to eight of 40 control group participants. Although independently developed, some consider CBASP to have key

similarities to CBT, and that it should perhaps be viewed as a new form of CBT (Power, 2013). 4.6 . Who might benefit? There was some systematic variation in the intake populations of these trials, particularly in the acute treatment that had been received, which we may be able to harness to gain more of an idea of who specifically might benefit from each of these interventions. Information about acute treatment prior to randomisation was widely available due to being part of linked trials, or being part of the studies' eligibility criteria. There was evidence that CBT may be effective for people who have received CBT to treat their acute episode, but also evidence that both CBT and MBCT may be effective as standalone interventions (that is, where participants were not treated with the same intervention in the acute phase). Further exploration of whether MBCT is effective in people who have received CBT or other psychological interventions is recommended. We only found evidence for continuation IPT among people who had received IPT combined with medication for the treatment of an acute episode. Clinical factors that have been identified as potentially important in relapse such as number of previous episodes (Bulloch, Williams, Lavorato, & Patten, 2013), levels of residual symptoms (Monroe, 2010), and psychiatric and physical health comorbidities (Hardeveld, Spijker, De Graaf, Nolen, & Beekman, 2010), did vary between trials. For example two of the MBCT trials required participants to have experienced three or more episodes of depression, whereas others, such as Vandermeer et al.'s mood management study (Van der Meer et al., 2010), included participants who were recovering from their first episode. However, the absence of consistent summary statistics across trials meant we could not make a reliable statistical comparison despite recommendations on the reporting of randomised controlled trials that should improve consistency of reporting (Moher, Schulz, Altman, & CONSORT Group (Consolidated Standards of Reporting Trials), 2001;

68


Schulz, Altman, Moher, et al., 2010). We are not alone in continuing to experience this problem (Biesheuvel-Leliefeld et al., 2015), and access to individual patient data from trials may be the way forward (Lo, 2015). 4.7 . Strengths and limitations of this review We have reported the review according to PRISMA recommendations (Moher, Liberati, Tetzlaff, & Altman, 2009), and believe it has several methodological strengths. Firstly, we pre-registered a protocol (Clarke et al., 2011), which is recommended to promote consistent conduct by the review team, and to ensure accountability, research integrity, and transparency of the eventual completed review (Moher et al., 2015). To ensure coverage of the relevant research, we conducted a comprehensive search of literature databases without specifying which interventions we were looking for, as well as contacting experts in the field. We used two independent researchers to sift titles and abstracts of studies for inclusion, and checked data extraction for errors (Day, Fayers, & Harvey, 1998). We assessed risk of bias and rated strength of evidence using recommended and widely used assessment tools (Guyatt et al., 2008; Higgins et al., 2011a). We used patient numbers rather than reported effect sizes in our analysis, attempting to obtain these from study authors when they were not reported. We then applied a consistent method of analysis to all included studies, meaning our estimates of effect were not biased by variations in the design of included trials. We were consistent in our treatment of drop-outs, used a clinically consistent point of randomisation, and compared data at the same point in time rather than using the last follow-up available. This leads us to a conservative estimate of treatment benefit and importantly allowed for a more accurate assessment of the relative evidence for CBT, IPT and MBCT. This allowed us to draw out potentially important differences in these interventions and trials which had not been possible in previous reviews (e.g. Biesheuvel-Leliefeld et al., 2015). The limited number of trials and the difference in populations and trial designs meant it was not possible to statistically explore patient or treatment factors that might be associated with variation in outcomes. We also did not make any direct comparison against medication. This was because we were not confident that the use of medication was limited to only to control arms that explicitly tested medication. To illustrate the extent of the issue: some trials required participants to have had medication, some required them not to have had medication, some delivered medication themselves, some didn't control medication use but did measure it, and some had no information about medication use whatsoever (see Table 1). There was a similar picture within the relapse prevention stage, with some trials requiring participants to continue with, or discontinue their medication, not allowing participants to start medication, having no medication requirements but measuring their use, or not controlling or measuring medication use. 4.8 . Implications for future research This review identifies a number of important issues for consideration in future research. Data from individual patients within trials in this review may allow for further exploration of relevant factors, including previous treatment (drugs or psychological) and specific psychological relapse interventions, the chronicity of the disorder and other possible indicators of relapse. An understanding of the role of these factors could lead to more targeted, appropriately sequenced (in relation to both psychological and pharmacological interventions), and effective interventions. From a psychological perspective the focus should be on understanding the differential benefits of different treatments and the mechanism by which they exert their effect. Recent studies (Rodgers et al., 2012) suggest that brief self-help interventions may be as effective as formal psychological interventions for some people, thereby reducing both the burden and cost of treatment, and are worthy

of further investigation. Larger trials of psychological therapies, particularly ones that directly compare different interventions within one trial, are also warranted. Although the review was not limited to psychological interventions, the majority of trials tested these. Current evidence suggests that, particularly given the fact that there are barriers to accessing healthcare services with depression (Hirschfeld et al., 1997), physical exercise, building networks of social support, psychologically informed selfhelp and engaging in meaningful activities could be a useful route to staying well (NICE, 2009). Given the need to target people who are currently at least partially recovered, and to reduce the possibility of future episodes of illness, these interventions may be more flexible and more readily adapted for widespread use in relapse prevention. Seeing how they measure up to more formal interventions may well be an important avenue to pursue in the development of new effective ways to limit the suffering associated with this widespread and damaging disorder. 5 . Conclusion Psychological therapies delivered following recovery from depression, including CBT, MBCT, and IPT, may reduce risk of relapse after one year by 22%. For CBT this effect is maintained after two years. This is similar to that previously reported for continuing medication, but without the need for continued treatment throughout the follow-up period. The difference in effectiveness between interventions and the mechanisms by which these interventions exert their effect are not well understood, nor is the most effective and efficient way to deliver them. Further research into the sequencing of interventions is recommended, particularly exploring how the efficacy of post-recovery interventions relates to acute treatment. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.cpr.2015.04.002. Role of funding sources This study was supported by a grant from the British Psychological Society to SP (grant number 509517). The funding source had no involvement in the study design, collection, analysis, or interpretation of data, writing the manuscript, and the decision to submit the manuscript for publication. Contributors KC, SP and EMW designed the study and wrote the protocol. KC and JK conducted literature searches and extracted data. KC conducted the statistical analysis and all authors contributed to interpretation of results. KC and SP wrote the manuscript and all authors have revised and approved the final manuscript. Conflict of interest statement SP also receives funding from NICE for the production of clinical guidelines. KC, EM-W and JK have no financial or other conflicts of interest to declare. Acknowledgements The authors thank Katie Greenfield for assisting with sifting references, data extraction and management, Laura Gibbon for assessing risk of bias, and Rob Saunders and Ellie Bishop for commenting on revisions of the manuscript.

References American Psychiatric Association (2000). DSM-IV-TR: Diagnostic and statistical manual of mental disorders (diagnostic & statistical manual of mental disorders (4th ed.). American Psychiatric Press Inc (text revision). Anderson, I.M., & Tomenson, B.M. (1995). Treatment discontinuation with selective serotonin reuptake inhibitors compared with tricyclic antidepressants: A metaanalysis. BMJ, 310(6992), 1433–1438. http://dx.doi.org/10.1136/bmj.310.6992.1433. Apil, S.R.A., Hoencamp, E., Judith Haffmans, P.M., & Spinhoven, P. (2012). A stepped care relapse prevention program for depression in older people: A randomized controlled trial. International Journal of Geriatric Psychiatry, 27(6) (June–591). Arean, P.A., & Ayalon, L. (2010). Preventing depression relapse in older adults. In C.S. Richards, & M.G. Perri (Eds.), Relapse prevention for depression (pp. 157–176). Washington, DC, US: American Psychological Association. Balshem, H., Helfand, M., Schünemann, H.J., Oxman, A.D., Kunz, R., Brozek, J., et al. (2011). GRADE guidelines: 3. Rating the quality of evidence. Journal of Clinical Epidemiology, 64(4), 401–406. Beck, J.S. (2011). Cognitive behavior therapy, second edition: Basics and beyond. Guilford Press.

K. Clarke et al. / Clinical Psychology Review 39 (2015) 58–70 Biesheuvel-Leliefeld, K.E.M., Kok, G.D., Bockting, C.L.H., Cuijpers, P., Hollon, S.D., van Marwijk, H.W.J., et al. (2015). Effectiveness of psychological interventions in preventing recurrence of depressive disorder: Meta-analysis and meta-regression. Journal of Affective Disorders, 174, 400–410. http://dx.doi.org/10.1016/j.jad.2014.12.016. Bockting, C., Hollon, S.D., Jarrett, R.B., Kuyken, W., & Dobson, K. (2015). A lifetime approach to major depressive disorder: The contributions of psychological interventions in preventing relapse and recurrence. Clinical Psychology Review. http://dx.doi. org/10.1016/j.cpr.2015.02.003. Bockting, C.L., Schene, A.H., Spinhoven, P., Koeter, M.W., Wouters, L.F., Huyser, J., et al. (2005). Preventing relapse/recurrence in recurrent depression with cognitive therapy: A randomized controlled trial. Journal of Consulting and Clinical Psychology, 73(4), 647. Bondolfi, G., Jermann, F., Der Linden, M.V., Gex-Fabry, M., Bizzini, L., Rouget, B.W., et al. (2010). Depression relapse prophylaxis with mindfulness-based cognitive therapy: Replication and extension in the Swiss health care system. Journal of Affective Disorders, 122(3), 224–231. Brown, R.A., Kahler, C.W., Niaura, R., Abrams, D.B., Sales, S.D., Ramsey, S.E., et al. (2001). Cognitive–behavioral treatment for depression in smoking cessation. Journal of Consulting and Clinical Psychology, 69(3), 471. Bulloch, A., Williams, J., Lavorato, D., & Patten, S. (2013). Recurrence of major depressive episodes is strongly dependent on the number of previous episodes. Depression and Anxiety, 31(1), 72–76. Clark, D.A., & Beck, A.T.B. (1999). Scientific foundations of cognitive theory and therapy of depression. John Wiley & Sons. Clarke, K., Pilling, S., Kenny, J., & Mayo-Wilson, E. (2011). Non-pharmacological relapse prevention for depression: systematic review. Retrieved April 4, 2014, from http://www.crd.york.ac.uk/PROSPERO/display_record.asp?ID=CRD42011001646#. Uz8skfZAte4 Cuijpers, P., Geraedts, A.S., van Oppen, P., Andersson, G., Markowitz, J.C., & van Straten, A. (2011). Interpersonal psychotherapy for depression: A meta-analysis. American Journal of Psychiatry, 168(6), 581–592. Cuijpers, P., & Smit, F. (2004). Subthreshold depression as a risk indicator for major depressive disorder: A systematic review of prospective studies. Acta Psychiatrica Scandinavica, 109(5), 325–331. http://dx.doi.org/10.1111/j.1600-0447.2004.00301.x. Day, B.S., Fayers, B.S., & Harvey, B.S. (1998). Double data entry: What value, what price? Controlled Clinical Trials, 19(1), 15–24. Fava, G. A., Grandi, S., Zielezny, M., & Canestrari, R. (1994). Cognitive behavioral treatment of residual symptoms in primary major depressive disorder. The American Journal of Psychiatry, 9, 1295–1299. Fava, G. A., Rafanelli, C., Grandi, S., Conti, S., & Belluardo, P. (1998). Prevention of recurrent depression with cognitive behavioral therapy: Preliminary findings. Archives of General Psychiatry, 9, 816–820. Frank, E., Kupfer, D.J., Buysse, D.J., Swartz, H.A., Pilkonis, P.A., Houck, P.R., et al. (2007). Randomized trial of weekly, twice-monthly, and monthly interpersonal psychotherapy as maintenance treatment for women with recurrent depression. The American Journal of Psychiatry(5), 761–767. Frank, E., Prien, R.F., Jarrett, R.B., Keller, M.B., Kupfer, D.J., Lavori, P.W., et al. (1991). Conceptualization and rationale for consensus definitions of terms in major depressive disorder: Remission, recovery, relapse, and recurrence. Archives of General Psychiatry, 48(9), 851–855. http://dx.doi.org/10.1001/archpsyc.1991.01810330075011. Geddes, J.R., Carney, S.M., Davies, C., Furukawa, T.A., Kupfer, D.J., Frank, E., et al. (2003). Relapse prevention with antidepressant drug treatment in depressive disorders: A systematic review. Lancet, 361(9358), 653–661. Gilbody, S., Bower, P., Fletcher, J., Richards, D., & Sutton, A.J. (2006). Collaborative care for depression: A cumulative meta-analysis and review of longer-term outcomes. Archives of Internal Medicine, 166(21), 2314–2321. Godfrin, K.A., & van Heeringen, C. (2010). The effects of mindfulness-based cognitive therapy on recurrence of depressive episodes, mental health and quality of life: A randomized controlled study. Behaviour Research and Therapy, 8, 738–746. Gortner, E.T., Gollan, J.K., Dobson, K.S., & Jacobson, N.S. (1998). Cognitive–behavioral treatment for depression: Relapse prevention. Journal of Consulting and Clinical Psychology, 66(2), 377–384. http://dx.doi.org/10.1037/0022-006X.66.2.377. Guyatt, G.H., Oxman, A.D., Vist, G.E., Kunz, R., Falck-Ytter, Y., Alonso-Coello, P., et al. (2008). GRADE: An emerging consensus on rating quality of evidence and strength of recommendations. BMJ, 336(7650), 924–926. http://dx.doi.org/10.1136/bmj. 39489.470347.AD. Hansen, H.V., Christensen, E.M., Dam, H., Gluud, C., Wetterslev, J., Kessing, L.V., et al. (2012). The effects of centralised and specialised intervention in the early course of severe unipolar depressive disorder: A randomised clinical trial. PLoS ONE, 7(3), e32950. Hardeveld, F., Spijker, J., De Graaf, R., Nolen, W. A., & Beekman, A. T. F. (2010). Prevalence and predictors of recurrence of major depressive disorder in the adult population. Acta Psychiatrica Scandinavica, 122(3), 184–191. Hersen, M., & Beidel, D.C. (2011). Adult psychopathology and diagnosis. John Wiley & Sons. Higgins, J.P.T., Altman, D.G., GU00F8tzsche, P.C., Jüni, P., Moher, D., Oxman, A.D., et al. (2011a). The Cochrane collaboration's tool for assessing risk of bias in randomised trials. BMJ [British Medical Journal], 343. Higgins, J.P.T., Green, S., & Collaboration, C. (2011b). Cochrane handbook for systematic reviews of interventions. Vol. 5.1.0, Wiley Online Library. Hirschfeld, R.M., Keller, M.B., Panico, S., Arons, B.S., Barlow, D., Davidoff, F., et al. (1997). The National Depressive and Manic–Depressive Association consensus statement on the undertreatment of depression. JAMA, 277(4), 333–340. Holländare, F., Anthony, S.A., Randestad, M., Tillfors, M., Carlbring, P., Andersson, G., et al. (2013). Two-year outcome of internet-based relapse prevention for partially remitted depression. Behaviour Research and Therapy, 51(11), 719–722. Hollon, S.D., Stewart, M.O., & Strunk, D. (2006). Enduring effects for cognitive behavior therapy in the treatment of depression and anxiety. Annual Review of Psychology, 57(1), 285–315. http://dx.doi.org/10.1146/annurev.psych.57.102904.190044.

69

Hunkeler, E.M., Hargreaves, W.A., Fireman, B., Terdiman, J., Meresman, J.F., Porterfield, Y., et al. (2012). A Web-delivered care management and patient self-management program for recurrent depression: A randomized trial. Psychiatric Services, 63(11), 1063–1071. Ingram, R.E. (1998). Cognitive vulnerability to depression. Guilford Press. Jarrett, R.B., Kraft, D., Doyle, J., Foster, B.M., Eaves, G.G., & Silver, P.C. (2001). Preventing recurrent depression using cognitive therapy with and without a continuation phase. Archives of General Psychiatry, 4, 381–388. Jarrett, R.B., Kraft, D., Schaffer, M., Witt-Browder, A., Risser, R., Atkins, D.H., et al. (2000). Reducing relapse in depressed outpatients with atypical features: A pilot study. Psychotherapy and Psychosomatics, 5, 232–239. Jarrett, R. B., Minhajuddin, A., Gershenfeld, H., Friedman, E. S., & Thase, M. E. (2013). Preventing depressive relapse and recurrence in higher-risk cognitive therapy responders: a randomized trial of continuation phase cognitive therapy, fluoxetine, or matched pill placebo. JAMA Psychiatry, 70(11), 1152–1160. Katon, W., Rutter, C., Ludman, E.J., Von Korff, M., Lin, E., Simon, G., et al. (2001). A randomized trial of relapse prevention of depression in primary care. Archives of General Psychiatry, 58(3), 241. Klein, D.N., Santiago, N.J., Vivian, D., Blalock, J.A., Kocsis, J.H., Markowitz, J.C., et al. (2004). Cognitive–behavioral analysis system of psychotherapy as a maintenance treatment for chronic depression. Journal of Consulting and Clinical Psychology(4), 681–688. Klerman, G.L., & Weissman, M.M. (1994). Interpersonal psychotherapy of depression: A brief, focused, specific strategy. Jason Aronson. Kuyken, W., Byford, S., Taylor, R.S., Watkins, E., Holden, E., White, K., et al. (2008). Mindfulness-based cognitive therapy to prevent relapse in recurrent depression. Journal of Consulting and Clinical Psychology(6), 966–978. Lo, B. (2015). Sharing clinical trial data: Maximizing benefits, minimizing risk. JAMA, 313(8), 793–794. Ma, S.H., & Teasdale, J.D. (2004). Mindfulness-based cognitive therapy for depression: Replication and exploration of differential relapse prevention effects. Journal of Consulting and Clinical Psychology, 1, 31–40. Mitchell, A. J., & Subramaniam, H. (2005). Prognosis of depression in old age compared to middle age: a systematic review of comparative studies. The American Journal of Psychiatry, 162(9), 1588–1601. Moher, Liberati, A., Tetzlaff, J., & Altman, D.G. (2009). Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement. Annals of Internal Medicine, 151(4), 264–269. Moher, D., Schulz, K.F., Altman, D., & CONSORT Group (Consolidated Standards of Reporting Trials) (2001). The CONSORT statement: revised recommendations for improving the quality of reports of parallel-group randomized trials. JAMA, 285(15), 1987–1991. Moher, Shamseer, L., Clarke, M., Ghersi, D., Liberati, A., Petticrew, M., et al. (2015). Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015 statement. Systematic Reviews, 4(1), 1. http://dx.doi.org/10.1186/ 2046-4053-4-1. Monroe (2010). Recurrence in major depression: Assessing risk indicators in the context of risk estimates. Retrieved from http://psycnet.apa.org/psycinfo/2009-18374-001 Murray, C.J., & Lopez, A.D. (1997). Alternative projections of mortality and disability by cause 1990–2020: Global Burden of Disease Study. The Lancet, 349(9064), 1498–1504. http://dx.doi.org/10.1016/S0140-6736(96)07492-2. Nezu, A.M., & Nezu, C.M. (2010). Problem-solving therapy for relapse prevention in depression. In C.S. Richards, & M.G. Perri (Eds.), Relapse prevention for depression (pp. 99–130). Washington, DC, US: American Psychological Association. NICE (2009, October 28). Depression in adults (update) [Guidance/clinical guidelines]. Retrieved April 16, 2012, from http://publications.nice.org.uk/depression-in-adults-cg90. Paykel, E.S., Scott, J., Teasdale, J.D., Johnson, A.L., Garland, A., Moore, R., et al. (1999). Prevention of relapse in residual depression by cognitive therapy: A controlled trial. Archives of General Psychiatry(9), 829–835. Piet, J., & Hougaard, E. (2011). The effect of mindfulness-based cognitive therapy for prevention of relapse in recurrent major depressive disorder: A systematic review and meta-analysis. Clinical Psychology Review, 31(6), 1032–1040. http://dx.doi.org/ 10.1016/j.cpr.2011.05.002. Power, M. (2013). CBT for depression. In M. Power (Ed.), The Wiley-Blackwell Handbook of Mood Disorders (pp. 173–191). John Wiley & Sons Ltd (Retrieved from http://onlinelibrary.wiley.com/doi/10.1002/9781118316153.ch7/summary). Priest, R.G., Vize, C., Roberts, A., Roberts, M., & Tylee, A. (1996). Lay people's attitudes to treatment of depression: Results of opinion poll for Defeat Depression Campaign just before its launch. BMJ, 313(7061), 858–859. Reynolds, C.F.I., Dew, M.A., Pollock, B.G., Mulsant, B.H., Frank, E., Miller, M.D., et al. (2006). Maintenance treatment of major depression in old age. The New England Journal of Medicine(11), 1130–1138. Reynolds, C.F.I., Frank, E., Dew, M.A., Houck, P.R., Miller, M., Mazumdar, S., et al. (1999). Treatment of 70+-year-olds with recurrent major depression: Excellent short-term but brittle long-term response. The American Journal of Geriatric Psychiatry(1), 64–69. Richards, C.S., & Perri, M.G. (Eds.). (2010). Relapse prevention for depression (1st ed.). American Psychological Association (APA). Rodgers, M., Asaria, M., Walker, S., McMillan, D., Lucock, M., Harden, M., et al. (2012). The clinical effectiveness and cost-effectiveness of low-intensity psychological interventions for the secondary prevention of relapse after depression: A systematic review. Retrieved from http://www.ncbi.nlm.nih.gov/pubmedhealth/PMH0047049/ Schulz, K.F., Altman, D.G., & Moher, D. (2010). The CONSORT Group statement: Updated guidelines for reporting parallel group randomised trials. BMC Medicine, 8(1), 18. Segal, Z.V., Bieling, P., Young, T., MacQueen, G., Cooke, R., Martin, L., et al. (2010). Antidepressant monotherapy vs sequential pharmacotherapy and mindfulnessbased cognitive therapy, or placebo, for relapse prophylaxis in recurrent depression. Archives of General Psychiatry(12), 1256–1264.

70


Segal, Z.V., Williams, J.M.G., & Teasdale, J.D. (2002). Mindfulness-based cognitive therapy for depression: A new approach to preventing relapse. Guilford Press. Segal, Z.V., Williams, J.M.G., & Teasdale, J.D. (2012). Mindfulness-based cognitive therapy for depression (2nd ed.). Guilford Press. Smit, A., Tiemens, B.G., & Ormel, J. (2007). Improving long-term outcome of depression in primary care: A review of RCTs with psychological and supportive interventions. The European Journal of Psychiatry, 21(1), 37–48. http://dx.doi.org/10.4321/S021361632007000100005. Stangier, U., Hilling, C., Heidenreich, T., Risch, A. K., Barocka, A., Schlösser, R., et al. (2013). Maintenance cognitive-behavioral therapy and manualized psychoeducation in the treatment of recurrent depression: A multicenter prospective randomized controlled trial. American Journal of Psychiatry, 170(6), 624–632. Teasdale, J.D., Segal, Z.V., Williams, J.M., Ridgeway, V.A., Soulsby, J.M., & Lau, M.A. (2000). Prevention of relapse/recurrence in major depression by mindfulness-based cognitive therapy. Journal of Consulting and Clinical Psychology, 4, 615–623. Thase, M.E. (2012). Relapse and recurrence of depression: An updated practical approach for prevention. Unützer, J., Katon, W., Callahan, C.M., Williams, J.W., Jr., Hunkeler, E., Harpole, L., et al. (2002). Collaborative care management of late-life depression in the primary care setting: A randomized controlled trial. JAMA, 288(22), 2836–2845. Van der Meer, R.M., Willemsen, M.C., Smit, F., Cuijpers, P., & Schippers, G.M. (2010). Effectiveness of a mood management component as an adjunct to a telephone counselling smoking cessation intervention for smokers with a past major depression: A pragmatic randomized controlled trial. Addiction, 11, 1991–1999. Viguera, A.C., Baldessarini, R.J., & Friedberg, J. (1998). Discontinuing antidepressant treatment in major depression. Harvard Review of Psychiatry, 5(6), 293–306. http://dx.doi. org/10.3109/10673229809003578. Vittengl, J. R., Clark, L. A., Dunn, T. W., & Jarrett, R. B. (2007). Reducing Relapse and Recurrence in Unipolar Depression: A Comparative Meta-Analysis of Cognitive–Behavioral Therapy’s Effects. Journal of Consulting and Clinical Psychology, 75(3), 475–488.

Vittengl, J.R., Clark, L.A., & Jarrett, R.B. (2009). Continuation-phase cognitive therapy's effects on remission and recovery from depression. Journal of Consulting and Clinical Psychology, 2, 367–371. Weissman, M.M., Kasl, S.V., & Klerman, G.L. (1976). Follow-up of depressed women after maintenance treatment. The American Journal of Psychiatry, 7, 757–760. Wells, K.B., Sherbourne, C., Schoenbaum, M., Duan, N., Meredith, L., Unützer, J., et al. (2000). Impact of disseminating quality improvement programs for depression in managed primary care: A randomized controlled trial. JAMA, 283(2), 212–220. Whiteford, H.A., Degenhardt, L., Rehm, J., Baxter, A.J., Ferrari, A.J., Erskine, H.E., et al. (2013). Global burden of disease attributable to mental and substance use disorders: Findings from the Global Burden of Disease Study 2010. Lancet, 382(9904), 1575–1586. http://dx.doi.org/10.1016/S0140-6736(13)61611-6. WHO (2009). Global health risks: mortality and burden of disease attributable to selected major risks. World Health Organization (Retrieved from http://books.google.co.uk/ books?hl=en&lr=&id=Ycbr2e2WPdcC&oi=fnd&pg=PR5&dq=world+health+ organisation+depression+leading+cause+of+diasbility&ots=afz-ZuYb4t&sig= xMsZYud1NGfHQsp9SgZz2-5yUxg). Wilkinson, P., Alder, N., Juszczak, E., Matthews, H., Merritt, C., Montgomery, H., et al. (2009). A pilot randomised controlled trial of a brief cognitive behavioral group intervention to reduce recurrence rates in late life depression. International Journal of Geriatric Psychiatry(1), 68–75. Williams, J.M.G., Crane, C., Barnhofer, T., Brennan, K., Duggan, D.S., Fennell, M.J., et al. (2013). Mindfulness-based cognitive therapy for preventing relapse in recurrent depression: A randomized dismantling trial. Retrieved from http://psycnet.apa.org/ psycinfo/2013-42381-001/