Univariate and Multiple Regression Analyses in Medical Research

Abstract

Summary Regression analysis methods, such as linear regression for continuous outcomes and logistic regression for binary outcomes, have been widely used in medical research data analysis for many years. However, there have been instances of misconceptions and misinterpretations of regression results within the medical community. Univariate and multiple regression analyses are commonly used by medical publications to identify factors that are significantly correlated with the outcome. In this manuscript, we critically evaluate the validity of this approach. Our findings indicate that this method is invalid and should be completely disregarded by medical researchers.