Postnatal psychiatric morbidity: a validation study of the GHQ-12 and the EPDS as screening tools

Abstract

Objective

To assess the validity of the 12-Item General Health Questionnaire (GHQ-12) and the Edinburgh Postnatal Depression Scale (EPDS) in screening for the most common postnatal psychiatric morbidities (mood, anxiety and adjustment disorders).

Method

A two-phase cross-sectional study was designed. First, a sample of 1453 women visiting at 6 weeks postpartum completed the GHQ-12 and the EPDS questionnaires. Second, based upon EPDS outcomes, participants were stratified and randomly selected within each stratum for clinical evaluation [Structured Clinical Interview for DSM-IV (SCID)]. Receiver operating characteristic (ROC) analysis was used.

Results

The concurrent validity was satisfactory (0.80). At optimum cut-off scores, both GHQ-12 and EPDS yielded very good sensitivity (80; 85.5) and specificity (80.4; 85.3), respectively. ROC curves showed that the performance of the EPDS (AUC=0.933) is slightly superior to that of GHQ-12 (AUC=0.904).

Conclusion

Both GHQ-12 and EPDS are valid instruments to detect postnatal depression as well as postnatal anxiety and adjustment disorders.

Introduction

Postnatal mental disorders are common, disabling and frequently undiagnosed in primary care, with prevalence rates ranging from 8.7% to 18% [1], [2]. Most research has been focused on postnatal depression (PND) which affects 10–13% of mothers after delivery [3], [4]; although recent evidence suggests that postnatal anxiety disorders may be more common than PND [5].

Postnatal anxiety has been considered a feature of postpartum depression [6] or a specific anxiety disorder, such as generalized anxiety, panic, obsessive compulsive disorder or posttraumatic stress disorder [5]. In a recent meta-analysis, anxiety during pregnancy was found to be a strong to moderate risk factor for postnatal depression [7]. Growing evidence indicates that maternal anxiety and depression during pregnancy [8], [9] and the postpartum period [10], [11] can negatively affect infant welfare. As Stewart [12] pointed out, “what is special about maternal mood disorders in the perinatal period is the impact they may have on the health and well-being of the woman, foetus and the family.” In this context, screening patients for early postnatal detection of both anxiety and depression is necessary. Although there is an ongoing debate on the usefulness and psychometric properties of screening measures of perinatal psychopathology [13], more research is clearly needed in this field.

The General Health Questionnaire (GHQ) [14] and the Edinburgh Postnatal Depression Scale (EPDS) [15] are two well-known rating scales potentially applicable for the detection of postnatal psychiatric morbidity. The GHQ is a self-administered questionnaire originally designed to detect common mental disorders, such as depression and anxiety, among general medical outpatients and in the community. Multiple-item versions have been shown to have good reliability and validity in general populations, including the 12-, 28-, 30- and 60-item versions. However, few studies have carried out a postpartum validation of the GHQ (12-, 28- and 30-item versions) and in these studies only a case criterion for depression was used [2], [16], [17]. In a recent review of screening instruments for PND, Boyd et al. [18] recommended the 12-item GHQ because this version is the shortest and has a slightly higher sensitivity and positive predictive value. To our knowledge, no version of the GHQ has been validated in postpartum women using a case criterion for anxiety and depression.

In contrast to the GHQ, the EPDS is a 10-item self-administered scale specifically designed for screening PND [15]. Extensive data about the validity and reliability of this instrument have been generated for different languages and cultures [19]. Cox et al. [15] stated that the EPDS is a one-dimensional instrument, but confirmatory factor analysis carried out by Pop et al. [20] found two distinct factors called “depressive feelings” and “cognitive anxiety.” The presence of two factors has been replicated by others using exploratory factor analyses [21], [22]. Furthermore, the total score of EPDS rather than its anxiety subscale [23] has shown a higher correlation with other anxiety measures, such as the State-Trait Anxiety Inventory [24]. In another longitudinal study, the anxiety subscale of the EPDS explained 38% of the total EPDS score at 6 weeks postpartum [6]. These findings indicate that EPDS has a strong anxiety component and is not a direct measure of depression. Therefore, it could be used as a measure of dysphoria or distress as hypothesized by Green [21]. A single validation study to identify postpartum women diagnosed with major and minor depression, panic disorder and acute adjustment disorder with anxiety showed good sensitivity and specificity [25]. The authors recommended a lower EPDS cut-off (7/8) than that suggested by most researchers who used the EPDS to screen for major and minor depression. However, this study explored a limited range of anxiety disorders and was carried out in primiparous who were recruited from antenatal parenthood classes, with a subsequent selection bias. Furthermore, concurrent validity of the EPDS was tested among other measures of depression, but a reliable measure of distress such as GHQ was not included.

The present study was designed with two objectives: (1) to validate the GHQ-12 and the EPDS questionnaires for the screening of postnatal psychiatric morbidity in a large sample of childbearing women using the Structured Clinical Interview for DSM-IV (SCID) as the gold standard for diagnosing anxiety, depression and adjustment disorders; and (2) to compare the performance of the GHQ-12 and the EPDS instruments using a case criterion for anxiety and depression.