Predictors of postnatal depression: using an antenatal needs assessment discussion tool

Abstract

Depression affects 13% of postnatal women and the long‐term repercussions affect women, children and families. Health visitors are being urged to increase antenatal support, including preparation for the impact of becoming a parent, and identification of vulnerable families. An antenatal tool incorporating validated measures was developed to facilitate discussion on childhood experiences, perceived physical and emotional social support and to screen for depression (using EPDS). Postnatal information was also collected. Health visitors in Bristol PCT used the tool when visiting primiparous women antenatally at home over two years. Twenty health visitors covering seven GP practices in South Bristol completed 187 antenatal and 142 postnatal records. There was a significant correlation between the 118 antenatal and postnatal EPDS scores recorded. An antenatal 13+ EPDS cut‐off gave a positive predictive value (PPV) of 35% for a high postnatal EPDS score and the 15+ cut off gave a PPV of 67%; women with high antenatal EPDS scores were significantly more likely to have postnatal depression symptomatology. Computed ‘At risk’ scores, combining high antenatal EPDS, poor antenatal emotional support and a high proportion of negative childhood adjectives were compared with postnatal EPDS scores. This score had a PPV of 86% using the antenatal EPDS 15+ score and a very high odds ratio for the likelihood of screening high postnatally. The high predictive value of the ‘at risk’ tool in identifying primiparous women who are likely to develop postnatal depression indicates that it is important to start these discussions in the antenatal period.

Acknowledgements

We would like to thank Julie Warner for her input and involvement with the early stages of the project, members of the working group who were involved in developing the assessment tool, all the health visitors who piloted the tool and participated in the study and Rosemary Greenwood for statistical advice.