Remission or Persistence? A Prediction Tool to Identify Women at Risk for Long-Term Depressive Symptoms Postpartum

Abstract

Background. Peripartum depression is a common complication with potential long-term adverse effects on the woman and her family. Approximately 30%–50% of newly delivered women experience prolonged depressive symptoms at 6–12 months postpartum. Early detection may facilitate preventive and treatment interventions. Aim. To investigate correlates for and create a tool for predicting long-term symptomatology in women experiencing depressive symptoms at 6 weeks postpartum. Materials and Methods. Data from the Biology, Affect, Stress, Imaging, and Cognition study was used, to identify women who scored high (≥12) on the Edinburgh Postnatal Depression Scale (EPDS) at 6 weeks postpartum (n = 697). Further, we collected data from medical records and included 40 variables based on earlier studies and clinical experience. A total of 654 women were included. Elastic net linear regression analysis was performed to identify predictors of continued symptoms at 6 months postpartum. An equation predicting the EPDS score at 6 months postpartum based on weighted variables was developed. Results. High education level and sleep for more than 6 hr per night in pregnancy week 17 were protective factors. Parity, pregnancy complications, stressful events, attention deficit hyperactivity disorder/attention deficit disorder, history of depression, depressive symptoms, and anxiety during pregnancy were predictive factors of prolonged depressive symptoms. A prediction tool with area under curve 0.73 and positive predictive value of 79%–83% depending on chosen EPDS cutoff was developed for clinical use. Conclusions. Our prediction tool offers a method to identify women at risk for persisting depressive symptoms postnatally, based on their significant depressive symptoms during the first weeks after delivery. Screening in order to identify these women can already start in the antenatal setting.