BackgroundThe Edinburgh Postnatal Depression Scale (EPDS) has been used successfully across diverse cultural settings. However, a recent study found poor validity in detecting postnatal common mental disorders (CMD) in rural Ethiopia. Using similar methodology, the study was replicated in the capital, Addis Ababa.MethodsSemantic, content and criterion validity of EPDS, Kessler scale-6 (K6) and Kessler scale-10 (K10) were assessed in postnatal women attending vaccination clinics. Criterion validation was undertaken on 100 postnatal women, with local psychiatrist diagnosis of CMD using the Comprehensive Psychopathological Rating Scale (CPRS) as the criterion measure.ResultsThe areas under the Receiver Operating Characteristic (AUROC) curve for the EPDS, K6 and K10 were 0.85 (95%CI 0.77–0.92), 0.86 (95%CI 0.76–0.97) and 0.87 (95%CI 0.78–0.97), respectively. The EPDS generated sensitivity, specificity and misclassification rates of 78.9%, 75.3% and 24.0%, respectively at an optimal cut-off point of 6/7. The corresponding values for the K6 were 84.2%, 82.7% and 17.0% at a cut-off point of 4/5, and for K10 were 84.2%, 77.8% and 21.0% at a cut-off point of 6/7, respectively. The internal reliability Cronbach's alpha for the EPDS, K6 and K10 were 0.71, 0.86 and 0.90, respectively.LimitationsNot all postnatal women bring their infants to vaccination clinics which may limit generalisability.ConclusionThe EPDS, K6 and K10 all demonstrated acceptable clinical utility as screening scales for postnatal CMD in an urban setting in Ethiopia. The marked urban-rural difference in EPDS performance within Ethiopia highlights the difficulty of applying urban-validated instruments to rural settings in LAMIC.