Comparative analysis of the performance of the postpartum depression screening scale with two other depression instruments

Background: Postpartum depression affects approximately 13% of mothers but up to 50% of all cases of this tragic illness can go undetected. Objective: The purpose of this study was to compare the performance of a ne wly created instrument, the Postpartum Depression Screening Scale (PDSS), w ith the Edinburgh Postnatal Depression Scale (EPOS) and a general depressio n scale, the Beck Depression Inventory-II (BDI-11). Method: In this methodological design a total of 150 new mothers completed these three instruments in random order, followed immediately by a DSM-IV d iagnostic interview. Using the LABROC I program, the areas under each of th e instrument's Receiver Operator Characteristic (ROC) curves were compared to determine if they were significantly different. Results: Eighteen (12%) of the women were diagnosed with major postpartum d epression, 28 women (19%) with minor postpartum depression, and 104 women ( 69%) with no depression. Compared to the EPOS, the PDSS had a significantly larger area under the ROC curve when screening for major or minor postpart um depression. When using the published recommended cut-off scores for majo r depression for the three instruments, the PDSS achieved the highest combi nation of sensitivity, 94%, and specificity, 98%. When detecting women with major or minor postpartum depression, the PDSS again yielded the highest c ombination of sensitivity (91%) and specificity (72%) of the three instrume nts. The PDSS identified 17 (94%) of the women diagnosed with major postpar tum depression, the EPOS identified 14 of these women (78%), and the BDI-II identified 10 of the 18 women (56%). Conclusion: If mothers identified as "most depressed" are substantially det ermined by the instrument used, the implications for both research and clin ical practice are significant. Researchers and clinicians need to be aware of the differential sensitivity of depression instruments which, while supp osedly measuring the same construct, are focused on different components of this mood disorder.