How Multiple-Treatments Meta-Analysis Can Challenge and Advance the Existing Clinical Evidence
Multiple treatments meta-analysis is increasingly used to evaluate the relative effectiveness of several competing interventions. In some fields which evolve with the continuous introduction of new agents over time, it is possible that in trials comparing older to newer regimens the effectiveness of the latter is exaggerated. Optimism, conflicts of interest and other forces may be responsible for this exaggeration, but its magnitude and impact, if any, needs to be formally assessed in each case. To evaluate this potential ?novelty' bias and adjust for it, multiple-treatments meta-regression models fitted within a Bayesian framework are applied. When there are several networks of interventions for diverse conditions or outcomes within the same field/specialty, one may consider either different bias in each meta-analysis or may consider novelty bias to be exchangeable across the different conditions and outcomes. The methodology is exemplified using several data from meta-analyses of multiple treatments and various simple graphical and quantitative ways to assist interpretation and improve presentation of results are discussed.