The detection and treatment of mental illness during pregnancy and postpartum are essential for the well-being of both mother and child. For this study, we set out to estimate the prevalence of perinatal depression among Bedouin mothers in southern Israel and determine if the latent structure of responses to the Edinburgh Postpartum Depression Scale (EPDS)-including depression, anxiety, and anhedonia-previously reported also applies to this population. A total of 332 women completed the EPDS during pregnancy (26-38 weeks) and again 2-4 months postpartum. Confirmatory factor analyses were performed to determine if first-order factors were correlated, or instead measure a second- or higher-order latent construct. We next performed temporal invariance analyses to compare the latent structure of EPDS responses over time. When pregnant, 35% of women provided EPDS responses suggestive of elevated depressive symptomology; this decreased to 23% postpartum. At both points, each EPDS factor significantly measures a higher-order, latent construct. The EPDS appears to measure three factors, labeled sadness, anxiety, and anhedonia. This latent structure appears stable (i.e., during pregnancy and postpartum). Further research is needed to validate EPDS responses versus structured clinical interviews. The construct validity of EPDS factors should be examined across other at-risk groups and over time.
Keywords: Bedouin; Edinburgh Postpartum Depression Scale; depression; postpartum; pregnancy.