Martina, B.R. et al. Journal of Critical Care | Published online: 4 May 2017
- No systematic exploration of confounders in randomized controlled trials reporting a significant effect on mortality in critically ill patients has been performed.
- Almost half of these trials were single centre and one third were not analysed according to the intention to treat principle, inflating effect size.
- Blinded and/or multicentre design was associated with an increased number needed to treat/harm.
- Major systematic biases exist and affect trial findings irrespective of the intervention being studied.
Read the full abstract here