When should women be screened for postnatal depression?

Maternal psychological morbidity following childbirth is one of the commonest complications of childbearing, with the prevalence of mild-to-moderate depression in the first few months after birth reported in one meta-analysis to be approximately 13% [1]. A more recent meta-analysis of 28 prospective studies, which reported findings from 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% [2]. Therefore, depression after birth is an important public health priority. The UK Confidential Enquiries into Maternal and Child Health, who review and report on maternal deaths during the previous 3-year period, have consistently found psychiatric disorders to be one of the leading causes of maternal deaths, often through suicide [3,4]. Similar findings have been reported in other industrialized countries [5]. These enquiries have highlighted the need to implement strategies to identify women with mental health problems owing to concerns that women who took their own lives were unlikely to have had their mental health needs identified or managed appropriately.

Postnatal depression can have adverse consequences on maternal–infant interactions [6], on a woman’s relationship with her partner and other family members [7], on her child’s longer-term cognitive and emotional development [8,9] and has been associated with the development of later psychiatric and behavioral problems in the child [10,11]. Despite evidence of the extent and impact of postnatal depression, it is estimated that less than 50% of cases are identified in routine clinical practice, highlighting why there is ongoing dialogue relating to screening and case identification strategies [12]. In addition, as the onset of depression may occur before or during pregnancy [13,14], there is increasing recognition that the focus on case identification should not only be in the postpartum period but is also important during pregnancy. Concern has also been expressed that the term ‘postnatal depression’ has been used by professionals for all mental disorders in the postpartum period, with policy and service development focused on interventions that potentially neglect the full range of perinatal mental disorders, including the increased risk of psychosis in the postnatal period [14].

Screening for postnatal depression

There has been much debate about the most effective and timely approach to routinely identify women at risk of depression during the perinatal period. A widely used self-report screening tool, the Edinburgh Postnatal Depression Scale (EPDS), includes ten items in which women are asked to rate how they have felt in the previous 7 days [15]. Each item is scored from 0 to 3, with a range of 0–30. A cut-off score of greater than 12 is recommended to indicate ‘probable’ depression and a score of greater than 9 to indicate ‘possible’ depression, with the need to confirm diagnosis through clinical assessment [15]. The American College of Obstetricians and Gynecologists has endorsed the routine screening of all patients for symptoms of depression and advocated the use of the EPDS for postpartum women [16]. In Australia, the ‘beyondblue’ National Action Plan for Perinatal Mental Health includes a proposal to roll out routine screening for all women during pregnancy and the postnatal period using the EPDS [17].

However, despite being widely used in routine practice during and after pregnancy in the UK, the use of the EPDS is not recommended by the UK National Screening Committee (NSC) unless used alongside clinical judgment. The UK NSC recommendation was based on evidence from a number of validation studies, which suggested that the EPDS has a relatively low positive-predictive value. A recent systematic review of studies validating the EPDS in antenatal and postnatal women compared with a ‘gold standard’ structured or semi-structured clinical interview, which included 37 studies from different countries, found significant heterogeneity in study methodology, cut-off score used and timing of diagnostic interview [18]. The reviewers concluded that use of the EPDS may not be equally valid across different settings and contexts, but also pointed out that there is little evidence to suggest that other instruments perform better than the EPDS.

In view of the recommendation of the UK NSC and increased interest in use of focused questions to identify mood disorders, particularly in primary care, the UK NICE guideline on Antenatal and Postnatal Mental Health [14], which informs NHS care in England and Wales, recommended use of the following ‘Whooley’ questions to identify women with possible depression:

• • During the past month, have you often been bothered by feeling down, depressed or hopeless?

• • During the past month, have you often been bothered by having little interest or pleasure in doing things?

A third question should be considered if the woman answers ‘yes’:

• • Is this something you feel you need or want help with [14].

Evidence considered by NICE found that the positive-predictive value of the EPDS varied across studies from 33 to 93%, the variability probably reflecting differences in the populations sampled and when and how the EPDS was administered. However, the Whooley questions also had a similarly low positive-predictive value (32%), and have not been validated in an antenatal or postnatal population [19]. A recent meta-analysis found that requiring these questions to be endorsed caused substantial numbers of depressed patients to be missed [20], raising concern that their use may not be helpful in the detection of postpartum depression [21]. Paulden and colleagues evaluated the cost–effectiveness of screening interventions for postnatal depression in primary care [22]. Alternative methods of screening at 6 weeks postpartum informed a cost–effectiveness decision model, outcomes of which showed that formal identification methods were not cost effective when compared with routine care because of the high number of false positives likely to be detected, although the authors acknowledged the lack of data on the use of Whooley questions in the perinatal period.

Timing of screening

Current evidence indicates that screening to improve the current low detection of depression after childbirth is feasible in maternity care settings and that estimates of sensitivity and specificity, although limited, are equivalent to those reported in other primary care settings [23]. Less attention has been paid to obtaining evidence of when screening for postnatal depression should be undertaken or whether timing is likely to influence the uptake of screening and identification of women at risk of depression. The majority of postpartum screening studies appear to have been conducted within the first 3 months postpartum, with administration of screening tools, such as the EPDS, commonly undertaken at 6–8 weeks, probably because this coincides with the end of the postnatal period when the woman is discharged from maternity care. In the systematic review by Gibson et al., 22 of the 37 studies included in the review, which were from a range of countries, administered the EPDS at 6–8 weeks [18]. NICE recommend that women are asked the Whooley questions at their first antenatal booking contact, as well as postnatally at 4–6 weeks and 3–4 months [14]. The Scottish Intercollegiate Guidelines Network (SIGN) guideline for the screening and management of postnatal depression recommends routine initial screening for depression at 6–8 weeks postpartum, with a second assessment at 3 months [24]. In New Zealand, women are not formally screened for depression, although the Well Child Schedule currently recommends use of the EPDS at core child health contacts at 6 weeks, 3 months and 5 months [25]. In the USA and Australia, similar recommendations have been made [17,26]. These timings appear to coincide with routine service contacts rather than an evidence base to support optimal screening outcomes, which will encourage pragmatic use of screening in practice and identify some women at risk, but may not be the most appropriate times to ensure as many women as possible are identified.

If two or more screening times are recommended, it is likely that many women may not receive a recommended second or third assessment as they may no longer be in contact with the maternity services and/or healthcare staff may fail to coordinate and implement further screening, especially if responsibility for care has been handed from one health professional group to another. In Scotland, a survey of whether health professionals were adhering to SIGN guidance found that only 55% of primary care practices were offering the recommended second assessment [27]. A study from Australia, which evaluated a screening program that included three applications of the EPDS over a 12-month period in 267 women, found that 34% of women received one of the three screns, 28% received two of the three and 15% received all three screens [28]. In total, 22% of women were not screened at all. Reasons for the lack of compliance included failure of the nurse to administer the tool or the woman not attending for her appointment.

A further consideration when attempting to elicit evidence of an optimal time to screen for depression during pregnancy and following birth is the difficulty of separating out the contribution of the content of the screening intervention, the timing of screening and the impact of any subsequent referral and management. A systematic review to evaluate the clinical effectiveness and cost–effectiveness of antenatal and postnatal identification of depressive symptoms highlighted that the effects of the screening component alone were difficult to disentangle from interventions linked to a positive screen, as some studies included tailored care and/or an intervention to support women who screened positive [12]. Moreover, recent US guidelines recommend against routinely screening adults for depression when interventions are not available [29] because, as a recent systematic review highlights, there is no benefit in screening programs for depression when interventions are not in place [30]. An important question still to be answered is whether screening antenatal and postnatal women and subsequently providing an intervention for those at risk of depression is clinically and cost effective.

Acceptability of timing of screening to women & healthcare professionals

If the timing and content of screening are to support maximum identification of women at risk, the intervention also has to be acceptable to the intended target group and the healthcare professionals responsible for implementing screening. One of the reasons given by the UK NSC for not recommending use of the EPDS was evidence that women found the tool unacceptable [31]. One UK qualitative study interviewed 39 women, who were between 11 and 19 months after childbirth, about completion of the EPDS at 3 months and over half of these women reported that they were anxious about the consequences, did not want to admit to feeling depressed and viewed the process of screening as simplistic. However, by contrast, other studies showed that women were ‘comfortable’ with use of the EPDS and found the ten items acceptable and appropriate [32]. Data from an evidence synthesis of methods to identify postnatal depression in primary care showed that women preferred to talk rather than complete a standardized questionnaire [12]. It was of note that the interpersonal relationship between the woman and healthcare professional was deemed to be important. In the majority of studies identified, the EPDS was acceptable to women and health professionals when women had advanced notice of the intervention, when the scale was administered at home and when health professionals had appropriate training in use of the EPDS and consideration of positive responses to item ten of the EPDS about self-harm [12].

Members of the healthcare team charged with implementing a screening intervention to identify maternal mental health needs require the skills and competency to administer the intervention, must ensure contacts are planned in line with recommended screening times and must be aware of care and referral options [14]. For the majority of women with identified or suspected mental health disorders, the major source of effective assessment and treatment will be in primary care, usually coordinated by the woman’s GP. There is an increased policy focus in the UK on improving access in primary care to psychological therapies for those experiencing mild-to-moderate depression [33]. Evidence of the effectiveness of these interventions in perinatal populations is required. Nevertheless, the relationship between the woman and her health professional is clearly important and could influence earlier intervention if signs and symptoms of depression are detected. As women may be reluctant to discuss how they are feeling, even in response to a screening tool, this aspect of care is even more important.

Conclusion

Depression is one of the more common maternal complications of pregnancy and the postnatal period. Screening interventions for depression can only be justified if the ‘test’ is accurate and leads to more effective treatment than would otherwise be the case. In the UK, the clinical effectiveness of screening for depression is lacking, as is evidence of the most appropriate timing to do this. Screening interventions do not appear to be cost effective at 6 weeks postpartum, although such conclusions are based on limited data and longer-term outcomes have not been considered. If timing is to be addressed, more epidemiological studies are required to confirm prevalence rates at different times and confirm what time point(s) would identify the greatest number of depressed women. Further evidence of how best to identify maternal mental health problems in the perinatal period is also needed. Although most postpartum screening studies to date have been conducted within the first 3 months of the birth, estimates of prevalence suggest that rates of depression may remain high for several more months.

Financial & competing interests disclosure

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.