Clin Infect Dis. 2021 Aug 11:ciab677. doi: 10.1093/cid/ciab677. Online ahead of print.
ABSTRACT
Clinical trials of severe sepsis that target crude (total) mortality as an end point do not have sufficient power to detect differences in mortality due to the intervention. We aim to discuss the importance of including the subcomponents of crude mortality in study design; estimate how sample size requirements change based on the proportion of attributable mortality; and how minor changes from predicted outcomes can affect results. We derived statistical curves to illustrate our points.
PMID:34379735 | DOI:10.1093/cid/ciab677