CITRIS-ALI Results & Discussion Thread

i finally found some time to run data simulations to examine power for a composite of the co-primary outcomes in CITRIS-ALI. Can be found here: https://pmbrown.net/post/composites-sepsis/

i can make my sas code available to anyone who requests it (random samples from a multivariate Normal before transforming the correlated variates to the appropriate scale eg exponential survival; correlations achieved through iteration). Naturally i’m quite sceptical about the power gain, as i explain, and worried composites will gain popularity in phase II trials in sepsis with the promise of accumulated power