[A study of the Edinburgh Postnatal Depression Scale (EPDS) on 859 mothers: detection of mothers at risk for postpartum depression].

The postpartum is a high-risk period for the occurrence of anxious and depressive episodes. Indeed, during the first few days after delivery, mothers can present postpartum blues symptomatology: fatigue, anxiety, disordered sleeping and a changing mood. Postpartum depression is characterised by a changing mood, anxiety, irritability, depression, panic and obsessional phenomena. It occurs in approximately 10 to 20% mothers. The exact prevalence depending on the criteria used for detection. The first symptoms usually appear between the fourth and sixth week postpartum. However, postpartum depression can start from the moment of birth, or may result from depression evolving continuously since pregnancy. We can add that the intensity of postpartum blues is a risk factor that can perturb maternal development. So it is important for health professionals to dispose of predictive tools. This study is a validation of the French version of the EPDS. The aims of the study were to evaluate the postpartum depression predictive value at 3 days postpartum and to determine a cut-off score for major depression. Subjects participating in this study were met in 3 obstetrical clinics in, or in the vicinity of, Toulouse. Mothers with psychological problems, under treatment for psychological problems or mothers whose babies present serious health problems were excluded from the study. The EPDS was presented to 859 mothers (mean age=30.3; SD=4.5) met at one of the clinics at 3 days postpartum (period 1). They had an EPDS mean score of 6.4 (SD=4.6); 258 (30%) mothers had an EPDS score 9. 82.6% of these mothers experienced a natural childbirth and 17.3% a caesarean section; 51.5% gave birth to their first child, 36.2% to their second child and 12.3% to their third or more. All subjects were given a second EPDS with written instructions to complete the scale during the period 4 to 6 weeks postpartum and return it for analysis (period 2). Between the 4 to 6 weeks postpartum period, 722 mothers replied again to the EPDS. 131 mothers had an EPDS score 11 (mean age=30.3; SD=4.8). They had an EPDS mean score of 13.6 (SD=3.3). Mothers with probable depression were interviewed and assessed, using the Mini (Mini Neuropsychiatric Interview, Lecrubier et al. 1997), the SIGH-D (Structured Interview Guide for the Hamilton Depression Scale) and the BDI (Beck Depression Inventory) in order to diagnose a major depressive episode. They had a HDRS mean score of 13.7 (SD=5.1) and a BDI mean score of 13.6 (SD=5). At 3 days postpartum, we observed that 258 mothers (30%) had an EPDS scores 9 and 164 mothers (19%) had an EPDS scores 11. Between 4 and 6 weeks postpartum, we observed 18.1% of postpartum depression (EPDS 11) and 16.8% (EPDS 12) of major postpartum depression. The analysis of the sensitivity and the specificity at 3 days postpartum provides a cut-off score of 9 (Sensibility: 0.88) (Specificity: 0.50) as predictive of postpartum depression, for this cut-off score, the type I error is low (5.8%) but the type II error is more higher (18.9%). The analysis of the sensitivity and the specificity between 4 and 6 weeks postpartum provides a cut-off score of 12 (Sensibility: 0.91) (Sensibility: 0.74) for the detection of major postpartum depression. Factor analysis shows at 3 days postpartum that the internal structure of the scale is composed of two subscales. The first factor F1 "anxiety" accounts 28% of the variance and the second factor F2 "depression" accounts 20% of the variance. Between 4 and 6 weeks postpartum, factor analysis suggests an unidimensional model in the evaluation of postpartum depression which is better than a two factor model. This factor accounts 40% of the variance. The scale has a good predictive value, and we can observe a significant correlation with the EPDS periods 1 and 2 (r=0.56; p<0.05). This result shows that the depressive mothers mood intensity predicts a future depressive risk. Furthermore, correlations between EPDS and BDI (r=0.68; p<0.05) and EPDS and HDRS (r=0.67; p<0.05) show a good convergent validity. The reliability study confirms the good internal consistency of the EPDS, at 3 days postpartum and in the postpartum depression -symptomatology evaluation (Cronbach's Alpha>0.80). In conclusion, this scale demonstrates good validity and is fast and easy use in obstetrical services, allowing early detection of women who risk to develop postpartum depression and, in the first week of postpartum, of mothers who suffer from a major postpartum depression. The use of the EPDS for an early screening of the risk of postnatal depression which is essential considering the consequences that postnatal depression can have on the development of the infant, on the quality of the relationship within the couple and on other social relationships. Mothers at risk for postnatal depression should be controlled and surveyed by the health professionals in obstetrical clinics.