Antenatal screening for the prediction of postnatal depression: validation of a psychosocial Pregnancy Risk Questionnaire.

OBJECTIVE: To assess the predictive value of an antenatal index of risk for postnatal depression (PND).
METHOD: Participants returned the Pregnancy Risk Questionnaire (PRQ; 18 antenatal items) and the Edinburgh Depression Scale (EDS) at a mean of 32 weeks gestation; the EDS was then mailed out at 2 and 4 months postpartum to ascertain those women who were screened positive (score>2), i.e. probable cases of depression. Only those returning the EDS at 2 and/or 4 months were included in the study (n=1296). Women who screened positive (n=322; 24.8%) were contacted and of these, 245 completed the Auto-Composite International Diagnostic Interview (CIDI) and form the basis of our primary analyses.
RESULTS: A CIDI diagnosis of major depression was found in 5.3% women at either 2 or 4 months. In this population, the optimal PRQ cut-off was >or=46 at which point sensitivity was 44% and specificity 92%. At this cut-off 9.9% tested positive (for risk of PND) on the PRQ with a positive predictive value (PPV) of 23.5%. Of 1079 women who scored below 46 on the PRQ, 3.2% were cases of CIDI depression while of those 119 scoring 46 or above, 23.5% were CIDI cases at 2 or 4 months postpartum, yielding an odds ratio of 9.18.
CONCLUSION: Using a cut-off of >or=46, the PRQ is better than previously reported tools in the antenatal prediction of PND with respect to sensitivity and specificity, while like other studies PPV remains limited. The PRQ allows identification of high and low risk groups and thus has applicability in both the research and clinical settings.