This study highlights the essential concept in the statistical treatment of medical and pharmaceutical data. To explain the concept, artificial data generated using random numbers was analyzed, and the importance of the confidence interval and distribution of data was shown. Sole use of the standard deviation (SD) is not considered appropriate from the viewpoint of clinical treatment, because many oversights occur. It was shown by considering the confidence interval that such oversights rarely occur, and the safety level is increased. Moreover, the probability of the occurrence of outliers from the average can be calculated by the distribution of data. Thus, pharmacokinetic data for humans were examined. The results for the human data also support the importance of the confidence interval and the distribution of differences between groups.