Final Report Summary - IMMA (Integrating the Multiple Meta-Analysis: a framework for evaluating and ranking multiple health care technologies.)
The great potential of network meta-analysis (NMA) to answer complex questions relevant to contemporary decision-making problems has been highlighted already in 2009 in a highly cited and discussed research article where twelve new generation antidepressants have been ranked according to their efficacy and acceptability (Cipriani et al. Lancet 2009 28;373(9665):746). However, skepticism and the lack of user-friendly methodology and software as well as gaps in the relevant statistical methodology hindered wide uptake of this evidence synthesis tool. In the last few years the project Integrating the Multiple Meta-Analysis (IMMA) has successfully tackled the obstacles towards establishing NMA as the new state of the art in evidence synthesis. We undertook extensive empirical research that led into a better understanding of the performance of the methodology and we expended on the statistical methods. We introduced several alternative models that account for the threat of publications bias and the fact that small studies might yield results that are different to those in large studies. We developed innovative statistical approaches that account for the fact that treatments in a network of interventions might be heterogeneously defined. We showed how knowledge about treatment effects could be enhanced when data from multiple treatments and multiple outcomes could be summarized in one step. We proposed a framework that can be used to evaluate the credibility of the conclusions from a NMA. Finally, we showed how future research could be informed and guided by identified gaps in a network of evidence. Our methodological developments were coupled with some groundbreaking clinical applications; among others we provided the hierarchy of drug interventions for acute mania, we ranked eighteen antipsychotics for schizophrenia, we ranked pharmacological treatments in the maintenance treatment of bipolar disorder, we estimated the hierarchy of percutaneous coronary interventions in-stent restenosis, we compared all blood pressure-lowering agents in patients with diabetes and kidney disease. We have produce a suite of free routines in user-friendly software to facilitate uptake of the methodology by non-statisticians.
As a team we have delivered more than a hundred presentations in international meetings, courses and workshops that familiarized statistical and clinical audiences with NMA. We have published about 50 articles, 5 chapters in books and four PhD theses while two more PhD theses are ongoing. We have established early on and cure a website that provides theoretical but also practical information to interested researchers.
As a team we have delivered more than a hundred presentations in international meetings, courses and workshops that familiarized statistical and clinical audiences with NMA. We have published about 50 articles, 5 chapters in books and four PhD theses while two more PhD theses are ongoing. We have established early on and cure a website that provides theoretical but also practical information to interested researchers.