Postnatal depression (PND), although sometimes difficult to define and recognise, was described in the classic study by Pitt in 1968 as: ‘what lies between the extreme of severe puerperal depression, with the risk of suicide and infanticide, and the trivial weepiness of “the blues”; something occurring frequently, much less dramatic than the former, yet decidedly more disabling than the latter’ (p. 1325). There is no precise definition of postnatal depression, but its clinical features are similar to depression occurring at any time within the general population and include lethargy, tearfulness, oversensitivity, hopelessness, anxiety, guilt, irrational fears and disturbed sleep patterns. The typical gradual onset means that it may not be easily distinguishable from the fatigue and emotional lability experienced by most mothers as they recover from childbirth and adjust to the demands of the baby (Holden 1991).

The International classification of diseases (ICD) (WHO 1990) makes provision for a diagnosis of puerperal mental health disorder if occurring postpartum and cannot be otherwise classified, and the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) (American Psychiatric Association Task Force 2000) allows postpartum onset to be specified for mood disorders that have started within 4 weeks postpartum. However, in both clinical practice and research, a much wider time frame is used to define postnatal depression, often including onset in the antenatal period and for up to a year postpartum. The Scottish Intercollegiate Guidelines Network (SIGN) defines postnatal depression as any non-psychotic depressive illness of mild to moderate severity occurring during the first postnatal year (SIGN 2002).

There have been at least two well-conducted systematic reviews and meta-analyses of the prevalence of depression in the postpartum period. O’Hara & Swain (1996), in a meta-analysis of 59 studies that included assessment using either standardised psychiatric interview-based methods or validated self-report measures, found the estimated average prevalence of postpartum depression to be 12.8%. They found that prevalence estimates were affected by the length of the postpartum period under evaluation and by the nature of the assessment method. In general, the self-report measures produced higher estimates; for example, average prevalence in 12 studies using the Edinburgh Postnatal Depression Scale (EPDS) was 12%, 11.6% in eight using the Beck Depression Inventory (BDI) and 18% in five using the Center for Epidemiological Studies-Depression Scale (CES-D). Average prevalence based on the interview-based methods was, for example, 10.5% in 19 studies using the Research Diagnostic Criteria (RDC) and 7.2% in three studies using the DSM.

A more recent meta-analysis of 28 prospective studies, based only on structured clinical interview assessments, estimated point prevalence including both major and minor depression at various times during the first postpartum year to be between 6.5% and 12.9% (Gavin et al 2005). In 18 of the studies prevalence was assessed by conducting a clinical interview on all study women, and in 10 studies women first completed a self-report instrument, with a clinical interview in those scoring above the predetermined cut-off. This review calculated point prevalence and period prevalence. Point prevalence of major and minor depression was found to rise after delivery and to be highest in the third month postpartum, at 12.9%. In the fourth to seventh months, prevalence declined slightly, staying in the range 9.9–10.6%, after which it declined to 6.5%. Each of these estimates, however, is based only on a few studies so all have wide confidence intervals. There were fewer estimates for period prevalence, the best estimate being that as many as 19.2% of women may have major or minor depression in the first 3 months after delivery, but confidence intervals were wide.

Although the term ‘postnatal depression’ is commonly used by professionals and mothers, some researchers have questioned the extent to which it comprises a specific entity (Green 1998). The review by Gavin et al (2005) identified three prevalence studies that had comparison groups of similar aged non-childbearing women (Cooper et al., 1988, Cox et al., 1993 and O’Hara et al., 1990). None of these indicated a statistically significant difference for prevalence of major or minor depression. Only one of these studies (Cox et al 1993) examined incidence of depression and this showed a threefold higher rate of depression within the first 5 weeks of childbirth amongst the postnatal women, compared with the equivalent time period for control women (odds ratio (OR) 3.26, 95% confidence interval (CI) 1.17–9.06). At 6 months postpartum, however, the difference in incidence was less and was not statistically significant (OR 1.48, 95% CI 0.77–2.82). Cox et al (1993) concluded that the threefold excess within the first month is most likely as a result of the ‘life event’ of giving birth and the immediate impact of a new family member, and that the category of PND remains a useful diagnostic term. Whether or not the prevalence, duration and nature of depression after childbirth are similar to those occurring among women generally, its appearance after birth remains relevant to those involved in providing maternity services.

In the search for possible risk factors of PND, studies have examined a large number of maternal, obstetric and sociodemographic characteristics, in particular, obstetric interventions, sociodemographic characteristics, psychiatric history, interpersonal relationships and support, and hormone disorders. There have been three systematic reviews of the risk factors of PND. O’Hara & Swain (1996) found that the strongest predictors of postpartum depression were past history of psychopathology and psychological disturbance during pregnancy, poor marital relationship and low social support and stressful life events. They found that indicators of low social status showed a small but significant predictive relationship with depression and there was a weak association with obstetric complications. Beck (2001) updated a previous meta-analysis of predictors of postpartum depression, first published in 1996 (Beck 1996), as there had been many more studies. She identified 13 significant predictors of depression, with 10 regarded as strongly predictive. These were prenatal depression, low self-esteem, childcare stress, prenatal anxiety, depression history, life stress, social support, poor marital relationship, difficult infant temperament and maternity blues. Marital status, low socio-economic status and unplanned/unwanted pregnancy were moderate predictors. Robertson et al (2004) completed the most recent systematic review, and gave risk factors in order of magnitude of effect size. These were depression or anxiety during pregnancy, life events, poor social support, previous history of depression, neuroticism and poor marital relationship. Low socio-economic status and obstetric factors had small effect sizes.

The NICE (2007) guidelines on antenatal and postnatal mental health, in a review of risk factors, comment on the lack of overlap of the studies that were included in the three reviews. The guidelines also consider a weakness of the reviews to be the inclusion of studies with PND based on only self-report scales in addition to diagnostic interviews. The review by Beck (2001) has more of the former than the review by O’Hara & Swain (1996). Robertson et al (2004) do not define method of assessing depression in their review, other than it had to have ‘proven reliability’. The NICE mental health guidelines found eight additional studies published since the reviews, thus not included in them, and comment that these largely support the findings of the earlier studies.

There is no good evidence that hormonal changes are a risk factor for PND (Romito, 1990 and Scottish Intercollegiate Guidelines Network (SIGN), 2002). A recent review specifically examining the possible link between caesarean section and postpartum depression concluded that such a link has not been established (Carter et al 2006).

Postnatal depression has been shown to have consequences for the development of the child, with some effects demonstrable in the longer term. Prospective studies have found an association between PND and insecure infant attachment (Murray 1992), behavioural and emotional problems in early childhood (Caplan et al 1989), cognitive development (Coghill et al., 1986, Hay et al., 2001 and Sharp et al., 1995) and later psychiatric and behavioural outcomes (Halligan et al., 2007 and Hay et al., 2003). Cooper & Murray (1998), in a review of PND, proposed that the association between PND and adverse child development was as a consequence of an impaired pattern of communication between the woman and her infant. That PND might have a lasting impact on a woman and her infant highlights the importance of the early detection and management of this problem.

An Australian study of 1336 women followed up at 6–7 months postpartum (Brown & Lumley 2000) found a positive association between PND (EPDS 13 or more) and maternal physical health and postpartum recovery. Similar associations were shown in a more in-depth telephone interview of a subsample of women at 7–9 months.

The management of PND will depend on its severity, but can consist of pharmacological or various forms of psychological or supportive treatments or a combination of these. For a very small proportion of women, admission to a specialised unit will be necessary. Since the last edition of this book there have been several systematic reviews of effects of treatment of PND (Boath and Henshaw, 2001 and Dennis, 2004, Hoffbrand et al 2001, Lumley et al 2004), and in the UK there have been guidelines produced by SIGN in 2002 and by NICE in 2007. In view of this, the individual management trials will not be described and the focus will be on the most recent systematic reviews and guidelines.

There has been much discussion since the first edition of this book about whether and how health professionals should systematically identify or screen for PND. The Edinburgh Postnatal Depression Scale (EPDS) (Cox et al 1987) was developed as a screening instrument specifically for postpartum women and was increasingly used in the 1990s, mainly by health visitors. The EPDS is a 10-item scale, each item having four possible responses, and the woman is asked to choose the response that comes closest to how she has felt during the previous 7 days. The responses are then scored. A score of 12 or more is considered to identify those women more likely to have depression (Cox et al 1987). The questionnaire had been found to be acceptable to both women and health professionals and to be quick to complete (Holden 1991).

In 2002, however, the National Screening Committee undertook a review of screening for PND and recommended that this should not be undertaken, and that the EPDS should not be used as a screening tool but only alongside clinical judgement. This was mainly because among the six EPDS validation studies some showed the positive predictive value (those correctly identified as having depression from the total with a positive test score) to be relatively low.

The NICE postnatal care guideline, launched in July 2006, and the NICE antenatal and postnatal mental health guideline, launched in February 2007, both address the issue of the identification of PND but have different recommendations for implementation. The postnatal care guideline states ‘At each postnatal contact, women should be asked about their emotional well-being, what family and social support they have and their usual coping strategies for dealing with day-to-day matters’ (NICE 2006, p. 6). The antenatal and postnatal mental health guideline states that:

These questions were devised by Whooley et al (1997) and tested in a general population sample. The guideline later suggests that ‘health care professionals may consider the use of self-report measures such as the EPDS, HADS or PHQ-9 as part of a subsequent assessment or for the routine monitoring of outcomes’.

The reason why the NICE antenatal and postnatal mental health guidelines do not recommend use of the EPDS as a screening tool is because the positive predictive value of the test varies considerably across studies, from 33% to 93%. The Whooley et al (1997) questions, however, have a similarly low positive predictive value, at 32%, and have never been tested in an antenatal or postnatal population. Indeed, the guideline proposes that a priority area for research is a validation study of these questions against a psychiatric interview in women in the first postpartum year. It would therefore seem premature to recommend them as standard care and this has resulted in concern that their use may discourage the detection of postpartum depression (Coyne & Mitchell 2007).

Given that this is all very recent, it is difficult to know exactly what should be recommended in terms of how PND should be identified, although it is clear that health professionals should do this systematically in one way or another. In summary, therefore, health professionals in their consultations with postpartum women should check whether the woman seems to be depressed in whatever way is considered to be appropriate, probably based on local practice and always including clinical judgement.