Large trials or meta-analysis? That is not the question

Selection and evaluation of evidence to identify the most effective treatment modalities is a difficult process. Randomized controlled trials are well accepted as the least biased means of evaluating medical, surgical, screening or preventive manoeuvres. The most commonly used statistical strategy to pool results from trials identified during a systematic review is the meta-analysis. Heterogeneity of trial results should be evaluated, interactions and specific populations identified for the planning of the new large trial.