An individual participant data meta-analysis of psychological interventions for preventing depression relapse

Major depressive disorder (MDD) is a leading cause of disability worldwide, characterized by high relapse rates. Relapse prevention is crucial to reduce the burden of MDD. Common approaches include antidepressant medication (ADM), psychological interventions, or a combination of both. While these interventions are used, determining the optimal intervention for each individual remains a challenge. A personalized strategy is needed to reduce trial and error. Individual participant data meta-analysis (IPDMA) offers a powerful approach to identify moderators and predictors of relapse more precisely than standard meta-analyses by pooling individual patient data (IPD). Previous IPDMAs have investigated specific interventions like mindfulness-based cognitive therapy (MBCT) or compared tapering ADM with ADM continuation; however, a broader exploration of various psychological interventions (cognitive behavioral therapy (CBT), continuation cognitive therapy (C-CT), preventive cognitive therapy (PCT), and MBCT) and their comparisons against different control conditions (active control, treatment as usual (TAU)) is needed to guide personalized treatment decisions. This study addresses this gap by performing an IPDMA of RCTs on psychological interventions for relapse prevention in previously depressed patients compared with various control conditions, including a broader range of moderators than previously explored.

Existing systematic reviews and meta-analyses have shown the effectiveness of psychological interventions, either alone or in combination with antidepressants, in preventing MDD relapse. However, these studies often relied on aggregate data, limiting the ability to identify individual-level predictors and moderators of treatment response. Two previous IPDMAs have addressed depression relapse prevention. Kuyken et al. (2016) compared MBCT with antidepressants or TAU, finding MBCT effective in reducing relapse risk. Breedvelt et al. (2021) examined the effects of a psychological intervention while tapering ADM compared with ADM continuation, finding no difference in relapse time. This current study extends previous work by including a broader range of psychological interventions and comparisons, allowing for a more comprehensive evaluation of their efficacy and the identification of key moderators of treatment effects.

This study employed a two-stage IPDMA. Initially, a systematic literature search across four databases (PubMed, PsychINFO, Embase, and Cochrane Central Register of Controlled Trials) was performed. 28 studies (n = 4,053) comparing psychological interventions for relapse prevention versus control were identified. Of these, 14 studies (n = 1,725) provided IPD, while 10 were included as aggregate data in a sensitivity analysis. The IPDMA focused on the time to relapse (in weeks), with relapse assessed via clinical diagnostic interviews (e.g., SCID). The primary analysis used follow-up data up to 12 months. Two pairwise comparisons were conducted: (1) psychological interventions (alone, with TAU, or with ADM) versus any non-psychological control group and (2) psychological interventions with TAU versus TAU only. One- and two-stage random-effects and fixed-effects analyses were conducted using Stata (v. 15.1 and v. 17). The hazard ratio was the primary effect size. Heterogeneity was assessed using the I² statistic. Predictor and moderator analyses were conducted using one-stage fixed-effects Cox proportional hazards models. A risk of bias assessment was done using the Cochrane Risk of Bias tool. Sensitivity analyses were performed to investigate inclusion bias, small-sample-size effects, and potential data availability bias.

The two-stage random-effects analysis revealed that psychological interventions significantly delayed the time to relapse compared to control conditions (HR 0.60, 95% CI 0.48–0.74, P ≤ 0.000, I² = 14.9%, n = 1,720). Adding psychological interventions to TAU also significantly reduced relapse risk compared to TAU alone (HR 0.62, 95% CI 0.47–0.82, P = 0.005, I² = 28.3%, n = 1,191). Subgroup analysis showed no significant difference in efficacy among the various psychological interventions (PCT, MBCT, CBT, and C-CT). Predictor analysis, using control group data, indicated that the number of previous depressive episodes and baseline depressive symptoms (HAM-D) were significant predictors of relapse, independent of treatment. Moderator analysis revealed a significant interaction effect for previous depressive episodes in the comparison of psychological interventions plus TAU versus TAU alone. Patients with three or more previous episodes benefited more from adding psychological interventions to TAU (HR 0.55, 95% CI 0.37–0.79, P = 0.006). Sensitivity analyses indicated no significant evidence of bias due to data availability or small study effects.

This IPDMA provides strong evidence for the efficacy of psychological interventions in preventing MDD relapse, particularly for individuals with a history of recurrent depression. The findings support the use of psychological interventions as an add-on to standard treatments, offering a personalized approach. The consistent lack of difference in efficacy across various psychological interventions may reflect genuine equivalence or limitations in sample size for certain interventions. The identification of previous episodes and residual depressive symptoms as predictors of relapse highlights the importance of considering these factors when stratifying patients for treatment. The interaction effect of previous episodes suggests a potential for tailoring interventions based on relapse risk. The lack of other significant moderators, like age of onset or marital status, in multivariable models suggests these factors might be less impactful when considering relapse risk relative to previous episodes and residual symptoms. However, more research is needed on the role of demographic factors in relapse prediction.

This large-scale IPDMA demonstrates the effectiveness of various psychological interventions in preventing MDD relapse, particularly for individuals with three or more previous episodes. The findings suggest that integrating psychological interventions into relapse prevention strategies should be considered standard practice. Future research should focus on longer follow-up periods, exploring the effectiveness of interventions for those with fewer previous episodes and investigating the interplay of other potential predictors and moderators to further refine personalized treatment strategies. Larger, multifactorial trials are needed to pinpoint the most effective intervention components for individuals with varying profiles.

Several limitations exist. The follow-up was censored at 12 months, potentially missing long-term effects. Only studies randomizing after remission were included, excluding studies on long-term effects of active interventions. Not all eligible studies provided IPD. Some relevant predictors (e.g., childhood trauma, socioeconomic status) were not consistently available across studies. Subgroup analyses for certain comparisons were limited by small sample sizes. Potential allegiance bias exists due to the involvement of intervention developers as co-authors, though efforts were made to mitigate this through an independent statistical analysis and diverse author group. Finally, interpersonal psychotherapy (IPT) was not assessed due to a lack of data.