Abstract
The fundamental statistical knowledge that is required in clinical studies is, first, being able to decide “Which test would be the best to use in which situation” when comparing basic attributes, and understanding logistic regression analysis as a form of multivariate analysis that is used to “adjust for confounding factors”. In addition, because many clinical studies use survival time as their outcome, it is also important to learn about survival analysis. When comparing basic attributes, after determining whether the variables are “continuous variables” or “categorical variables”, the decision as to which test to use depends on whether “2 groups” are going to be compared or “more than 3 groups” are going to be compared. Logistic regression analysis is often used when there are two outcome values, and it is possible to estimate the effect of exposure as an odds ratio. By contrast, with the Cox proportional hazard model, which is used in survival analyses, the effect of exposure can be estimated in the form of a hazard ratio. “Increasing the statistical knowledge required after studying epidemiology thoroughly” might be an efficient way to acquire knowledge about clinical studies in the midst of a busy clinical practice.
