We included the following intervention meta-analyses from the Cochrane Database of Systematic Reviews (March 31 , 2013 [ 11, 12]. 1.
For people with early psychosis, in four trials of early intervention services, four trials of CBT, and three trials of family intervention, meta-analysis demonstrated advantages over standard care.
Yet despite suggestions of the suitability of multiple interventions meta-analysis for this task, and an existing literature describing the multiple sources of complexity and heterogeneity in systematic reviews of complex interventions [ 5, 15, 16], little attention has been given as to how to model this complexity and heterogeneity in multiple intervention meta-analyses.
In fact, summative and interactive component models are conceptually equivalent between meta-regression and multiple interventions meta-analysis, though multiple interventions meta-analysis offers additional benefits in terms of flexibility of the modelling approach.
Yet in all of this work, multiple interventions meta-analysis was primarily considered for use with pharmacological, 'non-complex' interventions.
In particular, multiple interventions meta-analyses may be helpful both when a broad class of interventions is synthesised and when broadly similar interventions have different combinations of components.
A more recent iteration of this approach was offered by Madan and colleagues in a multiple interventions meta-analysis of smoking cessation interventions [ 31].
However, a multiple interventions meta-analysis conducted in this framework may mask substantial heterogeneity within relative treatment effects that may make intervention comparisons unhelpful due to imprecision.
For example, Barth and colleagues [ 20] conducted a multiple interventions meta-analysis of 198 randomised trials of psychotherapeutic interventions for depressive disorders.
Multiple interventions meta-analysis is a method that holds promise for bringing increased relevance, utility and coherence to systematic reviews of heterogeneous complex interventions.
Multiple interventions meta-analysis may also be of particular use when included interventions are not only heterogeneous, but complex in nature [ 4, 5].
