P-442 Establishment of prediction model of pregnancy rate and live birth rate in vitro fertilization-embryo transfer

Abstract

Abstract Study question Establishment of prediction model of pregnancy rate and live birth rate in vitro fertilization-embryo transfer Summary answer the calibration degree of the prediction model of live birth rate is good, and it is a relatively ideal prediction model of live birth outcome. What is known already The prediction model of clinical pregnancy rate has some differences between the average prediction probability of clinical pregnancy and the actual probability of occurrence of the IVF-ET patients, but the calibration degree of the prediction model of live birth rate is good, and it is a relatively ideal prediction model of live birth outcome. Study design, size, duration statistically significant influencing factors were obtained by multivariate logistic regression analysis. The influencing factors were used as covariables, and the prediction models of clinical pregnancy rate and live birth rate were constructed by using whether clinical pregnancy was obtained and whether live birth was obtained as dependent variables. The study population data (n = 2021) were randomly divided into the training set and the verification set according to the ratio of 6:4. Participants/materials, setting, methods Independent sample t test and nonparametric test were used for univariate analysis, and χ2 test was used to compare rates.Logistic regression analysis was used for multivariate analysis. The fit degree of the model was evaluated from the two aspects of differentiation degree and calibration degree respectively. Hosmer-Lemeshow x2 was used to test the degree of agreement between the two Main results and the role of chance The prediction model of clinical pregnancy rate was established with female age, basal FSH, progesterone level on HCG day, endometrial thickness on HCG day and embryo transfer as predictive variables. P=exp(1.669-0.069xfemale age -0.056x basal FSH-0.545xprogesterone level on HCG day+ 0.063x endometrial thickness on HCG day +0.807x transfer two cleavage embryos +0.803x transfer one blastocyst embryo) /[1 + (1.669-0.069x female age -0.056x basal FSH-0.545xprogesterone level on HCG day+ 0.063x endometrial thickness on HCG day +0.807x transfer two cleavage embryos +0.803x transfer one blastocyst embryo)] The prediction model of live birth rate was obtained with female age, basal E2, protocol of controlled ovarian hyperstimulation, progesterone level on HCG day, endometrial thickness on HCG day and embryo transfer as predictive variables. P=exp(0.135-0.074x female age +0.003x basal E2 + 1.110x ultra long protocol +0.768x long protocol +0.623x antagonist protocol -0.544xprogesterone level on HCG day+0.075xendometrial thickness on HCG day +0.771x transfer two cleavage embryos+0.750x transfer one blastocyst embryo) /[1 + (0.135-0.074x female age +0.003x basal E2 + 1.110x ultra long protocol +0.768x long protocol +0.623x antagonist protocol -0.544xprogesterone level on HCG day+0.075xendometrial thickness on HCG day +0.771x transfer two cleavage embryos+0.750x transfer one blastocyst embryo)] Limitations, reasons for caution The predictive model has not yet been translated into clinical application software Wider implications of the findings we hope to use the prediction model to give patients an expected value of assisted pregnancy outcome before treatment, so that patients can have an objective and correct understanding of their own conditions and assisted pregnancy outcome, reduce their psychological burden, and increase the confidence and compliance of treatment. Trial registration number not applicable