A transdiagnostic approach to conceptualizing depression across the perinatal period in a high-risk sample

• Authors: Amy L Cochran, University of Wisconsin-Madison; Blaire C Pingeton, Emory University; Sherryl Goodman, Emory University; Heidemarie Laurent, The University of Texas at Austin; Paul J Rathouz, The University of Texas at Austin; Jeffrey Newport, The University of Texas at Austin; ZacharyN Stowe, University of Wisconsin-Madison
• Journal Title: Journal of Abnormal Psychology
• Volume: Volume 129, Number 7
• Publisher: American Psychological Association. | 2020-10-01, Pages 689-700
• Type of Work: Article | Post-print: After Peer Review
• Publisher DOI: http://doi.org/10.1037/abn0000612
• Subjects: Health Sciences, Mental Health
• Keywords: Adult; Anxiety; Concept Formation; Depression; Depression, Postpartum; Factor Analysis, Statistical; Female; Humans; Mood Disorders; Postpartum Period; Pregnancy; Pregnancy Complications; Prospective Studies; Sleep Wake Disorders

Abstract

Clinical guidelines recommend assessing depression during pregnancy and postpartum but often overlook potential changes in symptoms across this developmental period. Such changes contribute to difficulties in conceptualizing maternal depression. This study aimed to situate depressive symptoms and related concerns (anxiety, stress, sleep) across the perinatal period within a transdiagnostic framework and to use this framework to better understand how depressive symptoms change across the perinatal period. First, items from seven symptom scales were a priori categorized into six transdiagnostic factors: four based on Research Domain Criteria (loss, potential threat, frustrative nonreward, and sleep-wakefulness) and two based on the depression literature (somatic and coping symptoms). Second, using prospective data from women with a history of an affective disorder (n = 657) in an observational study of neuropsychiatric illness, factor analyses were performed in seven periods (three trimesters of pregnancy and four quarters of first year postpartum). For each period, a bifactor model with six transdiagnostic factors and a general factor fit data better than models that combined or dropped a factor (p <.003). Except around delivery, item loadings and intercepts could be fixed between consecutive periods and still adequately fit data from both periods. Means of sleep-wakefulness and somatic factors increased significantly from second to third trimester (p<.01), with trends reversing early postpartum. In conclusion, depressive symptoms and related concerns exhibit factor structures that are only partly congruent across the perinatal period. This conclusion suggests that greater attention to specific life phases is warranted in the conceptualization of depression during this time in women's lives.