Abstract Objective: To assess differences in two different approaches to early inpatient Edinburgh Postnatal Depression Scale (EPDS) screening (targeted risk factor screening versus universal screening). Methods: Retrospective study of two cohorts delivering from January 2010 to June 2010. Cohort 1 included 5-weeks of consecutive deliveries. Risk factors for a positive EPDS screen were determined by multivariable logistic regression. Cohort 2 included all delivering women in the subsequent 5 months who had the most common risk factor (history of mental illness) for postpartum depression identified from Cohort 1. Results: For Cohort 1 of universal screening (N = 322), 11 (3.4%) had positive screens with seven (63.6%) having a history of any mental illness yielding the number needed to screen to identify one EPDS positive mother of 29. For Cohort 2 (N = 215), there were 33 (15.3%) positive screens and the number needed to have a positive screen was 6. Without a history of mental illness, predictive models for a positive screen were poor with the number needed to screen of 80. Conclusions: Early EPDS screening targeted for those with a history of mental illness is an efficient way to identify postpartum women with active mental health problems, but may miss up to 36.4% of potential screen positive women.