Predicting postnatal depression

Abstract

Background: The contribution that recently identified risk factors for Postnatal Depression (PND) made to an existing prenatal screening tool (‘Predictive Index’ of PND; Cooper et al., 1996) was assessed in an attempt to improve the performance of this measure. Method: The Predictive Index and measures of appraisal and coping were administered during the final trimester of pregnancy (n=306). The Edinburgh Postnatal Depression Scale (EPDS) and Childcare Stress Inventory were completed at 6-weeks postpartum (n=223). Results: The predictive index identified 23% of PND cases. The addition of specific appraisal and coping factors increased predictive performance by 26%. Inclusion of maternal reports of childcare stress in the early postpartum period resulted in a further 15% of PND cases being identified. Limitations: Diagnostic criteria were not used to assess the incidence of PND; instead PND case status was assessed using an established cut-off criterion on the EPDS. In addition, the contribution that women’s prenatal appraisal and coping processes made to improving the predictive index was not assessed on a second independent sample. Conclusions: The inclusion of prenatal appraisal, coping and maternal reports of childcare stress in a screening tool for PND substantially increased predictive performance, resulting in 64% of PND cases being identified. These findings highlight the importance of negative prenatal appraisal and coping processes in the development of PND and have important implications for the prevention and treatment of this disorder.

Introduction

The reported incidence of Postnatal Depression (PND) differs as a function of the type of assessment method used, with questionnaire measures yielding a higher prevalence rate than diagnostic interviews. The results of a recent meta-analysis, however, revealed that PND affects 13% of childbearing women (O’Hara and Swain, 1996). Although episodes of PND are generally short-lived (Cooper et al., 1991), recent research suggests that its effects can be detrimental to the mother–child relationship and result in a range of emotional and cognitive problems for the child (e.g., Campbell et al., 1992, Cogill et al., 1986, Murray et al., 1996, Sharp et al., 1995, Stein et al., 1991). In the light of these findings, the development of a sensitive screening tool to identify women at risk of developing PND would be of significant use to clinical practice and research. Early identification of such women would enable preventative interventions to be targeted at vulnerable women.

There have been several attempts to develop and prospectively validate such a screening tool (Appleby et al., 1994, Cooper et al., 1996, Leverton and Elliott, 1989; see also Elliott et al., 2000, Posner et al., 1997, Stamp et al., 1996). The measure developed by Appleby et al. (1994) was largely unsuccessful. Similarly, the measures developed by Posner et al. and Stamp et al. were of limited use for detecting major PND¹. Conversely, the ‘Predictive Index’ of Cooper et al. (1996), which was developed using a sample of 4000 women and prospectively validated on a further 2000 women, has achieved the most success at predicting women at risk of major PND. However, this index only identified one-third of those women who were subsequently diagnosed with major PND. Cooper et al. (1996) suggest that their index could be improved with the addition of certain postpartum variables, such as severity of maternity blues symptoms and infant behaviour.

Recent research has reported a link between the onset of PND and the use of: (i) maladaptive coping strategies (Gotlib et al., 1991, Honey et al., 2001, Honey et al., 2002; Terry et al., 1996), and (ii) negative appraisals of an anticipated childcare stressor (appraisals of ‘high threat’ and a ‘poor access to help’ from other people; Honey et al., 2001, Honey et al., 2002). Accordingly, the present research aimed to both evaluate the predictive index of Cooper et al. (1996), and investigate whether the addition of appraisal and coping processes to this index would improve its performance. The contribution that reports of childcare stress make to improving the predictive index was also assessed (cf. Cooper et al., 1996).