Sample size determination in the laser-Doppler measurements of skin blood flow

Abstract

Pre-study calculations of the required sample size are vital to a large majority of studies. Using the method based on the Monte-Carlo simulations, we have illustrated how the sample size is related to the statistic power value, the significance level, the variability of observations and the minor magnitude of the effect of interest under study. If the study has been already completed, one should not perform any 'post hoc' power calculations. In this case, calculation of confidence intervals is a better choice. We assessed the confidence intervals given in a number of publications where microcirculation is studied by LDF techniques using different protocols. We have found that types I and II errors are frequently encountered in the LDF studies, which is a consequence of an inappropriate sample size.

Keywords: LDF techniques; Mann-Whitney-Wilcoxon test; Monte-Carlo simulations; Sample-size planning.