目的 本研究旨在探讨河北省孕晚期胎盘早剥的发病率及其相关危险因素，并构建相应的列线图预测模型。 方法 收集2013—2023年间河北省22所监测医院的孕产妇资料，分析胎盘早剥的发病率、趋势及相关危险因素。通过R语言绘制预测模型的列线图，并利用受试者工作特征曲线下面积及校准曲线指标评估其预测效能。 结果 在2013—2023年期间，共纳入480 690名孕妇进行研究，孕晚期胎盘早剥的患病率为0.29%（1 395/480 690），平均年百分比变化为2.73%，且保持稳定趋势（P = 0.34）。多因素回归分析显示，妊娠合并贫血、子痫前期、前置胎盘、28周以下阴道出血以及妊娠28 ~ 31⁺⁶周均为胎盘早剥的高危因素（P < 0.05）；而规律产检（7 ~ 11次）和多胎妊娠是胎盘早剥保护因素（P < 0.05）。基于上述独立影响因素构建的列线图模型对胎盘早剥的预测曲线下面积为0.79，校准曲线显示预测值与实测值基本相符。 结论 应重视产前检查，特别关注子痫前期、前置胎盘及妊娠28周以下阴道出血的孕妇，同时加强对妊娠28 ~ 31⁺⁶周期间的管理，以降低胎盘早剥的发生率。基于此构建的列线图模型预测效能好，为胎盘早剥的预防与临床管理提供理论依据。

Objective To investigate the incidence of placental abruption in the third trimester of pregnancy in Hebei Province， identify its associated risk factors， and develop a corresponding nomogram prediction model. Methods Data from pregnant women at 22 monitored hospitals in Hebei Province， collected between 2013 and 2023， were analyzed to assess the incidence， trends， and associated risk factors of placental abruption. A prediction model was developed and visualized using R programming to generate the receiver operating characteristic （ROC） curve. The model's predictive performance was evaluated using the area under the curve （AUC） and calibration curve parameters. Results A total of 480 690 pregnant women were included in this study over the specified period. The incidence of placental abruption was 0.29% （1 395 out of 480 690）， with an average annual percentage change of 2.73%， indicating a stable trend （P = 0.34）. Multivariate logistic regression analysis revealed that pregnancy-related anemia， preeclampsia， placenta previa， vaginal bleeding before 28 weeks' gestation， and between 28 and 31⁺⁶ weeks' gestation were significant risk factors for placental abruption （all P < 0.05）. In contrast， regular prenatal examinations （7 ~ 11 visits） and multiple pregnancies were identified as protective factors against placental abruption （both P < 0.05）. The area under the curve （AUC） for the nomogram model constructed based on independent risk factors for placental abruption was 0.79， and the calibration curve demonstrated that the predicted values closely aligned with the observed values. Conclusions Prenatal examinations should be prioritized， especially for women with preeclampsia， placenta previa， or a history of vaginal bleeding before 28 weeks of gestation. Management of pregnancies between 28 and 31⁺⁶ weeks is also crucial to reduce the incidence of placental abruption and mitigate adverse maternal and neonatal outcomes. The nomogram model constructed based on these factors exhibits excellent predictive performance， providing a solid theoretical foundation for the prevention and clinical management of placental abruption.