Adaptation and validation of the Luganda version of the Edinburgh Postnatal Depression Scale for screening prenatal depression among mothers attending the Antenatal Clinic at Kawempe National Referral Hospital

Abstract

Introduction: Depression affects approximately 364 million people, globally. 7.8% of mothers in USA suffer from ante partum depression, 27.5% in Turkey while between 26.3 and 32.9% suffer from the debilitating condition in Africa. Opportunities for screening during the several ante natal contacts are missed. This may be attributed to the higher technical and requirements to use the known gold standard diagnostic tool, the DSM-5 criteria for Major Depressive Disorder (MDD). There is still a paucity of information on locally adapted and validated user-friendly screening tools for prenatal depression in Uganda. Untreated prenatal mental illness is associated with poor outcomes. In this study, we proposed to adapt and validate the Luganda version of the Edinburgh Postnatal Depression (EPDS-L) Scale for screening prenatal depression among mothers attending the antenatal care (ANC) clinic at Kawempe National Referral hospital (KNRH). Methods: This was a cross-sectional facility-based study in KNRH using International Society for Pharmacoeconomics and Outcomes Research (ISPOR) guidelines for the adapting the EPDS-L and quantitative approaches for the validation phase. Consecutive sampling was done until the desired sample size of 100 participants was achieved. The EPDS-L was administered by research assistants, and then the DSM-5 diagnostic criteria for Major depressive disorder was administered by trained Psychiatric nurse to all participants. Quantitative data was analyzed using STATA version 15. Reliability was demonstrated by the Cronbach’s alpha coefficient while validity was demonstrated by sensitivity, specificity, Negative Predictive Value, Positive Predictive Value and Area Under the Receiver Operator Characteristic curve. Results: The EPDS-L was adapted following the ISPOR taskforce recommendation (Appendix 5). The tool had a Cronbach’s Alpha coefficient of 0.8515 hence considered reliable. Using a cut off of 13 or more, sensitivity was 65%, specificity was 100%, Positive Predictive Value of 100% and Negative Predictive Value of 83.3%. The Area Under the Receiver Operator Curve was 0.99. Hence the EPDS-L can be used reliably at a cut off of 13. Additionally, the tool can even perform better at a cut off of 10, with sensitivity of 97.14% and specificity of 98.46%. Conclusion: The EPDS-L was adapted following the ISPOR taskforce recommendations and can be used reliably at a cut off of 13. The tool can even perform better at a cut off of 10.