, Volume 16, Issue 5, pp 401-410,
Open Access This content is freely available online to anyone, anywhere at any time.
Date: 25 Apr 2013

Detection of antenatal depression in rural HIV-affected populations with short and ultrashort versions of the Edinburgh Postnatal Depression Scale (EPDS)

Abstract

Risk of antenatal depression has been shown to be elevated in Southern Africa and can impact maternal and child outcomes, especially in the context of the Human Immunodeficiency Virus (HIV). Brief screening methods may optimize access to care during pregnancy, particularly where resources are scarce. This research evaluated shorter versions of the Edinburgh Postnatal Depression Scale (EPDS) to detect antenatal depression. This cross-sectional study at a large primary health care (PHC) facility recruited a consecutive series of 109 antenatal attendees in rural South Africa. Women were in the second half of pregnancy and completed the EPDS and Structured Clinical Interview for Depression (SCID). The recommended EPDS cutoff (≥13) was used to determine probable depression. Four versions, including the 10-item scale, seven-item depression, and novel three- and five-item versions developed through regression analysis, were evaluated using receiver operating characteristic (ROC) analysis. High numbers of women 51/109 (47 %) were depressed, most depression was chronic, and nearly half of the women were HIV positive 49/109 (45 %). The novel three-item version had improved positive predictive value (PPV) over the 10-item version and equivalent specificity to the seven-item depression subscale; the novel five-item provided the best overall performance in terms of ROC and Cronbach's reliability statistics and had improved specificity. The brevity, sensitivity, and reliability of the short and ultrashort versions could facilitate widespread community screening. The usefulness of the novel three- and five-item versions are underscored by the fact that sensitivity is important at first screening, while specificity becomes more important at higher levels of care. Replication in larger samples is required.