多类反应结果的logistic回归在产后出血影响因素分析中的应用

Application of polychotomous logistic regression in the study of risk factors predisposed to postpartum hemorrhage]

To determine the role of history of induced abortion and pregnancy induced hypertension in postpartum hemorrhage, a polychotomous logistic regression method was introduced to analyze the data gathered in an investigation of blood loss among 933 parturients in some counties of Sichuan and Ningxia provinces. The results show that postpartum hemorrhage caused by placenta factors is related to a history of induced abortion and pregnancy induced hypertension with relative risks 2.55 (95% confidence interval = 1.2-5.3) and 4.94 (95% confidence interval = 1.8-13.2), respectively. An increased risk was noted for postpartum hemorrhage associated with pregnancy induced hypertension (relative risk = 2.58, 95% confidence interval = 1.3-5.2), but not with induced abortion. Suggestions were made for prevention of postpartum hemorrhage.