To the Editor: Dr Timsit and colleagues1 reported the results of a 2-by-2 factorial design for the prevention of extraluminal catheter colonization and subsequent infections. As the authors acknowledged, the validity of such design lies in the absence of interaction between the 2 interventions. In the “Methods” section, they stated that this assumption was confirmed by testing for interaction in the Cox model. However, they did not report the power of this test to rule out any possible interaction.
Parienti J. Chlorhexidine-Impregnated Sponges and Prevention of Catheter-Related Infections. JAMA. 2009;302(4):379-380. doi:10.1001/jama.2009.1035