A JAMA meta-analysis looked at higher vs lower PEEP. Overall, for patients with ALI and ARDS, there was no improvement in survival associated with higher PEEP. For the subgroup with ARDS there was a trend toward improved survival but the difference was of borderline statistical signifance (p=0.49, adjusted RR CI .81-1). If this difference is true the NNT would be 20. There was no difference in pneumothorax or vasopressor use. Is this enough to recommend change in practice?