Construction and validation a risk prediction model for precipitate delivery in multiparous women based on LASSO regression
HAN Lu
WANG Xue
XUE Jun
YANG Yixuan
FU Li
Abstract：Objective:To explore the risk factors for precipitate delivery in multiparous women undergoing vaginal delivery,to construct a risk prediction model,and to test the predictive effectiveness of the model.Methods:A retrospective data collection method was used.A total of 360 multiparous women admitted to a tertiary grade A hospital in Tianjin from October 1,2022 to December 31,2023.The related factors were selected and compared between the group of precipitate delivery.And the group of non-precipitate delivery by using Logistic regression based on the LASSO selection for construct a predictive model.The discrimination was evaluated by the area under the ROC curve(AUC).The goodness of fit of the model was verified by Hosmer-Lemeshow test.And the DCA curve was used for verify the net benefit in clinical.The Bootstrap resampling was used for internal validation.Results:There were 141 cases occur precipitate delivery among the 360 multiparous women with the incidence rate 39.2%.By using multivariate Logistic regression analysis,the results showed that the age(OR=1.633,95%CI 1.130-2.361),the gestational age≥40 week(OR=0.524,95%CI 0.309-0.891),hypertensive disorders of pregnancy(OR=2.079,95%CI 1.055-4.099),labor induction by oxytocin(OR=0.503,95%CI 0.280-0.903),the interpregnancy interval(OR=0.652,95%CI 0.453-0.939),use labor analgesia(OR=0.137,95%CI 0.046-0.405),history of precipitate delivery(OR=4.438,95%CI 1.314-14.984),premature rupture of membranes(OR=2.124,95%CI 1.235-3.651)were independently associated with precipitate delivery.The area under ROC of the training set was 0.757(95%CI 0.706-0.808).The Hosmer-Lemeshow test showed P=0.620.The best value of cutoff was 0.39.The Youden index was 0.40 with the sensitivity of 73.8%and the specificity of 66.2%.Meanwhile the probability of the display threshold of the decision making curve at 0.1-0.7 benefits high.The area under ROC of the internal validation by bootstrap was 0.713(95%CI 0.641-0.785).Conclusion:The predictive model constructed in this research has performed a good predictive power.The Nomogram could be used as a visualization tool for clinical to identify and to select the high risk patients.
Keywords：precipitate deliveryparitymultiparous womenLASSO regression modelNomograminfluencing factors
Publication Date：2025-06-10
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：9( 1785-1793 )