Identifying women at-risk for depression in the immediate postpartum period
Dennis CLE, Janssen PA and Singer J;
Commented by , 22 Oct 2004
Background
Depression is the most common form of postpartum morbidity, carrying serious and at times tragic consequences for the mother, child and family. While postpartum depression is highly treatable with early intervention, there is limited evidence to justify universal screening. Thus, targeted screening programs are warranted.
Predicitve models aimed at identifying high-risk mothers antenatally and after 6 weeks postpartum have been of limited predictive value. Recent research suggests maternal mood in the immediate postpartum period to be strongly predictive for postpartum depression.
Purpose
To develop a multifactorial predictive model of depressive symptomatology in the first postnatal week, to enable early targeted screening programs for the identification of postpartum depression.
Methods
A population-based sample of 594 mothers near Vancouver, British Columbia, completed a mailed questionnaire at 1 week postpartum as part of a longitudinal study. The questionnaire included the Edinburgh Postnatal Depression Scale (EPDS), a 10-item self-report measure, established as the preferred method for community based screening.
Postpartum depression risk factors were determined following literature review of several medical and psychiatric databases. They were characterized into sociodemographic, biological, pregnancy-related, life stressors, social support, obstetric and maternal adjustment domains; several were quantified using standardized measures.
Following univariate analysis, sequential regression analysis was performed to predict depressive symptoms in the first postpartum week.
Results
In the multivariate model, significant depressive symptomatology (EPDS > 9) at 1 week was predicted by the following factors
- immigration within the last 5 years
- history of depression independent of childbirth
- diagnosis of pregnancy-induced hypertension
- vulnerable personality style
- stressful life events
- lack of perceived support
- lack of readiness for hospital discharge
- and dissatisfaction with infant feeding method (p,0.001).
Dr Ghatavi's comments
Several limitations noted by the authors deserve mention. The generalizability of the study is limited by the low proportion of ethnic minorities and single women. The validity of the diagnosis of depression is compromised by reliance on a self-report instrument.
While the authors acknowledge postpartum blues may confound the presence of a true depressive syndrome in the early postpartum period, it is unclear why they chose not to evaluate a predictive model of risk factors at 1 week with depressive symptoms at 8 weeks, as this data was obtained as part of their longitudinal study.
Notwithstanding these limitations, this is an important study; the first to develop a predictive model for postpartum depression in the very early postnatal period, aiding in earlier targeted screening of high-risk mothers.
While the findings are consistent with previous identified factors (determined in the antenatal or > 6-week period), several new risk factors were identified, three of which (noted below) were the strongest predictors in the model.
Recent immigration carried a five-fold risk of depressive symptoms, and pregnancy-induced hypertension almost a four-fold risk. Both of these factors would appear to be associated with significant stress, thus consistent with the well-established importance of stressful life events contributing to postpartum depression.
Lack of readiness for hospital discharge also carried an almost four-fold risk, which reflects the importance of adequate social support, another well-established predictor of postpartum depression.
Finally, mothers with a one-week EPDS score > 9 were 19 times more likely to have significant persistent depressive symptoms at 8-weeks, suggesting this brief self-report measure may be a valuable targeted screening tool in high-risk mothers.