Assessment of P values of demographic data in randomized controlled trials reported in Korean Journal of Anesthesiology between 2010 to 2017.

In large number of randomized controlled trials, many of researchers would like to provide P values of demographic data which commonly described in table 1 for the purpose of emphasizing no difference between groups. The authors would intend to demonstrate that statistically insignificant P value in demographic data implicate randomization has been well performed between groups. However, insignificant P value does not necessarily reflect the successful randomization. It is more important to establish a statistical analysis plan thoroughly in the stage of the designing and planning study and to consider if any of the variables included in the demographic data may affect the research results. If a researcher designed and planned a study well and performed it exactly and nicely, the study results would not be influenced by the P values no matter what the p-values are significant or not. The most important steps should be on the stage of designing and planning study. Although well planned and performed, controlled variables might affect results in some studies. In this situation, the researcher can provide both results with initial method and second stage of analysis including such variables. For the brief conclusion, it would be useless to exhibit P values in baseline data table.