FROM RISK FACTORS TO PREDICTION: BINARY LOGISTIC REGRESSION FOR ENDOMETRIAL CANCER IN A SIMULATED CLINICAL DATASET

Cristina Gena Dascălu, Doriana Agop-Forna, Cristina David, Magda Ecaterina Antohe

DOI : 10.62610/RJOR.2025.2.17.18

Abstract

In this paper we discuss about a well-known method for qualitative data analysis – the binary logistic regression. We present the theoretical foundations of this method, conditions for its use and implementation aspects in a statistical software (SPSS 29.0), including predictor selection and processing strategies, model validation via significance tests and model quality evaluation. These theoretical issues are illustrated using a realistic AI-generated dataset of 400 cases of endometrial cases, analysing eight risk factors widely reported in the literature. Two predictive models are compared: one using predictors in their original form, and another based solely on statistically significant binary predictors. The results interpretation is carried out, in order to illustrate in detail how to conduct such an analysis in clinical studies. The example provided aims to emphasize the need for careful interpretation of results and outlines the method’s limitations, highlighting also its practical value.