The CMIMG was established as an official Methods Group of Cochrane in September 2010. Cochrane Methods Groups develop methodology and advise Cochrane on how the validity and precision of systematic reviews can be improved. Methods Groups base their activities around six core functions; those relevant to our Methods Group are elaborated on the "Our Core Functions" page.Background
In recent decades, a proliferation of treatment options and an explosion in the number of randomized trials has created a need for systematic reviews of studies of intervention effectiveness. This has resulted in an explosion in the number of systematic reviews: there are currently over 3000 published reviews in the Cochrane Database of Systematic Reviews. Many of these reviews include comparisons of multiple treatment options. For example, there are over 200 Cochrane reviews with a broad title of the form "Interventions for ...". The division of reviews into those addressing specific questions (i.e. of one intervention comparison) and those addressing multiple comparisons is somewhat variable. For instance, across 22 reviews of interventions for adult smoking cessation, there are over 40 distinct treatment regimes analysed for the single outcome of abstinence at 6 months.
Cochrane Overview of reviews (Overviews) were recently developed with the main aim of summarising the results of multiple systematic reviews covering multiple interventions for the same clinical condition. The initial format for Overviews is that they should extract the results as reported in the component systematic reviews, and reformat them in tables or figures. Thus they do not aim to repeat or update the literature searches, eligibility assessment, bias assessment or evidence synthesis from the reviews that are summarised. Since an important audience for Overviews is health care decision makers approaching The Cochrane Library for an answer to a question such as "which intervention should I use for this condition?", this simple re-formatting of summary estimates into a table does not necessarily provide rigorous and up-to-date evidence to meet the needs of users.
Statistical methods. Both Cochrane Overviews and conventional Intervention reviews that cover multiple interventions produce multiple intervention evidence structures. Statistical methods for analysing multiple interventions simultaneously, in a single meta-analysis, have become available during the last 15 years and are variously known as multiple-treatments meta-analysis (MTM), mixed-treatment comparisons (MTC) and network meta-analysis. MTM is a multidimensional extension of meta-analysis, taking advantage of indirect evidence in a network of interventions to infer the relative effectiveness of many treatments for the same health condition. In important attribute of a MTM is the ability to rank the interventions. Since naive rankings are typically misleading, this is best achieved probabilistically, for example by calculating the probability that each treatment is the most effective (or harmful).
A growing number of review authors are keen to employ these new techniques, both in Overviews and conventional reviews. The first Cochrane Overview with an MTM analysis was recently published (Singh et al., 2009), and more are expected to appear in the coming years. Several MTM reviews undertaken by Cochrane authors have been published outside of The Cochrane Library. The reasons for publishing externally likely relate to a combination of limitations in the current format of Cochrane Overviews, the lack of guidance regarding use of MTM in conventional Cochrane reviews, and a lack of knowledge or expertise among personnel of Cochrane Review Groups.
Opponents and enthusiasts of MTM met to discuss the relevance of MTM in Overviews during the 17th Cochrane Colloquium (2009) in Singapore. Attendees agreed that MTM is an area of new methodological development that extends beyond the scope of any existing Methods Group and that may influence the way in which Cochrane overviews are undertaken in the future. The need for a specific Methods Group was agreed upon. An exploratory meeting was held in March 2010 in Bristol to set up the Comparing Multiple Interventions Methods Group (CMIMG).