Objective To analyze the influencing factors of postpartum post-traumatic stress disorder (PP-PTSD) in pregnant women, and construct a predictive nomogram model for occurrence of PP-PTSD to verify its efficacy.
Methods A total of 1 027 pregnant women after delivery of 42 days were included and randomly divided into model group (719 cases) and validation group (308 cases) at a ratio of 7∶3. They were divided into PP-PTSD group and non-PP-PTSD group based on whether PP-PTSD occurred. The maternal status was conducted using a case study questionnaire, a Self-rating Anxiety Scale (SAS), a PTSD Checklist-Civilian Version (PCL-C), a Perceived Social Support Scale (PSSS), and the Edinburgh Postnatal Depression Scale (EPDS). Logistic regression model was applied to analyze the influencing factors of PP-PTSD, and R software was applied to draw a nomogram model, and the predictive efficiency of the model was verified by receiver operating characteristic (ROC) curve, calibration curve and H-L test.
Results Within 42 days after delivery, 136 (13.24%) of 1 027 pregnant women had positive PP-PTSD symptoms, including 94 cases in the model group and 42 cases in the validation group. The cesarean section rate, EPDS score, proportion of patients with gestational diabetes mellitus, 5 min Apgar score of newborns (0 to 7), SAS score in the PP-PTSD group were higher, and PSSS score of was lower than that of the non-PP-PTSD group. Logistic regression model showed that EPDS score, delivery mode, gestational diabetes mellitus, SAS score, 5 min Apgar score for neonates and PSSS score were all independent influencing factors for the occurrence of PP-PTSD(P < 0.05). A nomogram prediction model was constructed based on the above 6 influencing factors. The area under the curve (AUC) of the model group to predict PP-PTSD was 0.930(95%CI, 0.905 to 0.955), the sensitivity and specificity were 89.40% and 87.40%, respectively. The slope of calibration curve was close to 1. H-L test showed that chi-square value was 8.873 and P value was 0.409. In the verification group, the model predicted that the AUC of PP-PTSD was 0.900 (95%CI, 0.850 to 0.950), the specificity and sensitivity were 86.90% and 84.50%, respectively, and the slope of the calibration curve was close to 1. The H-L test showed that chi-square value was 6.089, and P value was 0.457.
Conclusion The influencing factors of PP-PTSD in parturients include social support status, cesarean section, pregnancy complicating with diabetes, postpartum depression, Apgar score at 5 minutes, and anxiety state. The nomogram model constructed on above factors has a good predictive value for the occurrence of PP-PTSD in parturients.
DownLoad: