Binary logistic regression is a model for prediction of pregnancy by the in vitro fertilization in women with tuboperitoneal infertility

Abstract

Predicting the onset of pregnancy by in vitro fertilization (IVF) in women with tuboperitoneal infertility is a pressing challenge of modern medicine. Aim: to develop a method for determining the chance of pregnancy by IVF in women with tuboperitoneal infertility, based on the analysis of physiological parameters, using the binary logistic regression. Materials and methods. The object of the study is clinically healthy women before pregnancy and in the first trimester after conception, women with tuboperitoneal infertility before entering the IVF and in the first trimester of gestation, and women with tuboperitoneal infertility and an unsuccessful IVF trial. We analyzed the state of the cardiovascular system and autonomic regulation, adaptive capabilities and height-weight indicators at the first stage of the study. The data were statistically processed using the Microsoft Office and Statistica10 software package. We used the binary logistic regression method to create a model for predicting pregnancy in the second stage. Results. Based on the physiological indicators obtained using logistic regression analysis a mathematical model was built – a formula for assessing the probability of non-pregnancy. Conclusion. The quality of prognosis of the binary logistic regression model was 97.2%. The model can serve as a method for predicting the onset of pregnancy by IVF in women with tuboperitoneal infertility. We recommend this method for use in the work of a practicing physician.