AIDSImpact.com:: - Santa Fe, USA 2011

Abstract #245 - Screening for depression during pregnancy among women testing for HIV in rural South Africa
Authors: Presenting Author: Dr Tamsen Rochat - University of KwaZulu Natal
Additional Authors: Prof. Mark Tomlinson, Prof. Marie-Louise Newell, Prof. Alan Stein,
Aim: Prevalence of perinatal depression in Southern Africa is concerningly high and associated with poor foetal and obstetric outcomes, poorer uptake of antenatal care and could impact on uptake of PMTCT and adherence to HIV treatment. Shortages of professional health care workers result in reliance on lay counsellors and community health care worker (CHW) resources in service delivery. Brief, user friendly screening methods are critical to optimising prevention and treatment. The objective of this research was to examine whether shorter versions of the Edinburgh Postnatal Depression Scale (EPDS) were as effective as the commonly used 10-item version to screen for depression.
Method / Issue: Cross-sectional study at a large primary health care facility in rural South Africa in which a consecutive series of 109 women in antenatal care during the second half of pregnancy were enrolled 2-4 weeks after HIV testing. Women completed the EPDS and a structured clinical interview for depression administered by a trained interviewer. Inter-rater agreement was assessed using Cohen’s kappa coefficient and reliability using Cronbach’s alpha. Multiple regression techniques examined EPDS items to identify those significantly associated with depression. Using the recommended cut off for clinical depression (≥13) and previously established methodology for scoring brief versions of the EPDS the performance of four versions were evaluated using ROC analysis.
Results / Comments: The majority (85.3%) of women reported the pregnancy to be unplanned, 44.9% were HIV positive. Point prevalence of depression determined by clinical interview (46.79% CI 37.2-56.3) and EPDS (44.04% CI 34.57-53.50) was similar. The kappa statistic suggested moderate agreement when accounting for chance (K=0.4638, agreement 73.39%; expected agreement 50.38% SE 0.095). EPDS showed sensitivity of 69% and specificity of 78%. Cronbach’s alpha for the EPDS10 was fair (ά=0.6130) but fell short of guidelines (ά ≥0.70) for a stand-alone screening measure. Regression analysis of EPDS items against the depression outcome showed five significant items in univariate analysis at p<0.001 level. The alpha statistic (ά=0.7501) was improved when using only these five items to calculate Cronbach’s alpha. In multivariable analysis, three items remained significant. In ROC analysis this novel 3-item version was shown to have improved effectiveness (AUC =0.83) compared to the 10-item EPDS (AUC =0.73). Neither depression as measured by the clinical interview (OR 1.84 [0.86-3.95] p=.117) or as measured by the novel 3-item version was significantly associated with HIV status (OR 1.74 [0.78-3.88] p=0.173).
Discussion: Both HIV positive and HIV negative women are at high risk of depression during pregnancy, suggesting universal screening in primary health care is critical for prevention and treatment. The brevity and sensitivity of this novel 3-item EPDS could facilitate screening for antenatal and postpartum depression by lay health counsellors and community health care workers in primary health care and PMTCT. In Southern Africa, where resources are scarce, this could significantly improve detection of women in need of further assessment and treatment, thus improving treatment entry and maternal and child outcomes
Go Back