Culturally sensitive translation of the Edinburgh Postnatal Depression Scale into Marathi language to assess perinatal depression in rural Maharashtra, India

Background: Tools to screen perinatal depression are developed in English and for their effective use in culturally and linguistically diverse settings it is important to translate them appropriately, contextually adapt them, and rigorously validate them in the local populations. Marathi is the third most spoken language in India, yet there is no culturally sensitive translation of the Edinburgh Postnatal Depression Scale (EPDS) into Marathi till date. In this paper, we describe the systematic process of translation and adaptation of the EPDS into Marathi. Methods: EPDS was translated into Marathi using the following steps: forward translation, bilingual committee discussion, cognitive interviewing and pre-testing, backward translation, and comparison of the original English version with the back-translated English version (linguistic equivalence). Ten pregnant women were initially involved in cognitive interviewing and provided feedback on the Marathi version. Further, conceptual equivalence was tested with 20 non-pregnant women with bilingual skills and operational equivalence to compare self-administered and interviewer-administered version was completed with 20 pregnant women. Scatter plot, Bland-Altman plot, and Pearson's correlation coefficient were estimated to assess the relationship between the two versions: English and Marathi in case of conceptual equivalence and self-administered and interviewer-administered in case of operational equivalence. Results: Pearson's correlation coefficient between the original English version of the EPDS and the Marathi version of the EPDS was 0.99 (p<0.001). The difference in scores between the two versions for all participants was within the limits of agreement as per the Bland-Altman approach. Similarly, for the self-administered and interviewer-administered Marathi version of the EPDS, the correlation coefficient was 0.96 (p<0.001). The difference between the two scores for nine out of ten participants was within the limits of agreement as per the Bland-Altman approach. Conclusion: This is only the second study to date to use a systematic process-based approach for culturally sensitive translation and adaptation of the EPDS into another language. The two Marathi versions of the EPDS (self-administered and interviewer-administered) can now be used for assessment of the psychometric validity of the tool.