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

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