I think with all EBM, it is important to think about context. It is all well and good giving a patient a drug because the papaers say so, but if that drug won't be beneficial for that patient as they have to come regularly to get it or travel far, then it may the best drug but it isn't the best drug for that patient!
publisher bias is an interesting point by Sultana. Its true that people won't want to read Descriptions of failed trials or ones that only affect few people so these wont be highlighted in the same way that others will... But this doesn't make them any less relevant? All need to be considered for a clinical decision.
