To the Editor:
Thank you to Drs Safiri and Ayubi for their response to our article on the topic of mortality associated with vancomycin after implementation of an antimicrobial stewardship program.
1
We agree that a conservative P value of <.05 can increase the risk of testimation bias for univariate analysis included into a multivariate analysis. However, we feel that you are confusing effect size with P value.By using a P value of <.05, 5 predictors were included into our multivariate model in accordance with the commonly accepted rule of 10 events per variable (EPV) in logistic regression models. We had 11 events for every variable included in the model. Furthermore, a study done by Vittinghoff and McCulloch concluded the rule of 10 EPV could be further relaxed because they found uncommon statistical issues with 5-9 EPV.
2
Bootstrapping is unnecessary in studies with a large sample size.References
- Decreased mortality in patients prescribed vancomycin after implementation of antimicrobial stewardship program.Am J Infect Control. 2017; 45: 1194-1197
- Relaxing the rule of ten events per variable in logistic and Cox regression.Am J Epidemiol. 2007; 165: 710-718
Article info
Footnotes
Am J Infect Control 2017;45(11):1194-1197.
Conflicts of interest: None to report.
Identification
Copyright
Published by Elsevier Inc. on behalf of Association for Professionals in Infection Control and Epidemiology, Inc.
