The diagnostic accuracy of the Edinburgh Postnatal Depression Scale without the self-harm item

Abstract

The Edinburgh Postnatal Depression Scale (EPDS) is the most commonly used screening tool for perinatal depression. However, the 10th item of the EPDS intended to evaluate suicide ideation does not accurately capture suicide ideation and may cause psychological distress. Whereas endorsement of the strongest agreement “yes, quite often” on this item may be associated with suicide ideation, the response on this frequency is perfectly predicted by the EPDS full score. The discard of this item, therefore, is preferred. In the current study, using data from a nation-wide internet survey conducted in Japanese postpartum (n = 5688) and pregnant women (n = 1639), we show that the EPDS-9 without the 10th item correlates perfectly with the full EPDS or EPDS-10. Furthermore, the EPDS-9 and EPDS-10 have equivalent performance in differentiating participants' self-reported depression diagnosis as well as Kessler Psychological Distress Scale (K6)-based screening of depression. Lastly, at the cutoff of 9, the EPDS-9 performs well in predicting participants’ response on the 10th item of EPDS; specifically, the sensitivity of differentiating frequent thoughts of self-harm is 0.968 and 1 in postpartum and pregnant women, respectively. Therefore, the EPDS-9 performs equivalently to EPDS-10 and can be considered to replace EPDS-10 in future use.