Variability in use of cut-off scores and formats on the Edinburgh Postnatal Depression Scale: implications for clinical and research practice.

TLDR: The increasing use in the literature of unvalidated cut-off scores on the Edinburgh Depression Scale (EDS/EPDS), as well as different wording and formatting in the scale, is highlighted to highlight possible reasons for these ‘errors’.

Abstract: Objectives: i) To highlight the increasing use in the literature of unvalidated cut-off scores on the Edinburgh Depression Scale (EDS/EPDS), as well as different wording and formatting in the scale; ii) to investigate and discuss the possible impact of using an unvalidated cut-off score; iii) to highlight possible reasons for these ‘errors’; and iv) to make recommendations to clinicians and researchers who use the EDS/EPDS. Method: A convenience sample of studies that have used unvalidated cut-off scores, or different formatting, are cited as evidence that these types of ‘errors’ are occurring fairly frequently. Examination of previous data from one of the authors is undertaken to determine the effect of using an unvalidated cut-off score. Summary: Many studies report rates of high scorers on the EDS/EPDS using different cut-off scores to the validated ones. The effect of doing this on the overall rate can be substantial. The effect of using different formatting is not known, though excluding items from the EDS/EPDS must also make a substantial difference. Recommendations: We recommend that i) the validated score of 13 or more is used when reporting on probable major depression in postnatal English-speaking women, and 15 or more when reporting on antenatal English-speaking women; ii) that the wording used is “13 or more” (or equivalent), and not other terms that may cause confusion (e.g., ‘>12’; ‘more than 12’; ‘13’ etc), iii) if a different cut-off score to the validated one is used, a clear explanation is given as to why this has been done; and iv) that the scale should be worded and formatted as originally described by its authors.