Abstract
<p>Depression that occurs during pregnancy and postpartum (collectively, “peripartum”) is a major public health concern. The goal of this study was to examine how individual depressive symptoms and affective, cognitive, and somatic subscales change and are related to one another over peripartum time. 418 women completed the Beck Depression Inventory at each of five phases: three pregnancy trimesters, early (<9 >weeks), and late (9-61 weeks) postpartum. ANOVA was used to examine differences in symptom and subscale severity and trajectory across the phases. Network analysis was used to model the causal mechanisms underlying these changes. Severity of most symptoms and all three subscales varied across the phases, but did not follow the same trajectory as overall severity. “Indecisiveness” and “libido” did not vary significantly as a function of peripartum time. Network analysis demonstrated that, on average, symptoms clustered within their respective subscales, but the first trimester and early postpartum networks were characterized by greater dimensionality and disorganization, while second trimester and late postpartum networks were characterized by tight clusters. Somatic symptoms, overall, were the least predictable by the network, pointing to causal mechanisms outside of the networks. A difference in symptom centrality was observed in the symptoms representing negative affect: “sadness” was most important in pregnancy, “pessimism” in early postpartum”, and “past failure” in late postpartum. This study was limited by low power, and a sample of women receiving medication treatment. Yes, taken together, these results suggest that peripartum depression symptom severity, and the mechanisms that maintain these symptoms are heterogeneous. Constellations of similarities and differences across phases were observed.</p>