A mixed methods review to develop and confirm a framework for assessing midwifery practice in perinatal mental health

Aim: To ascertain whether a new framework examining midwifery practice in perinatal mental health is supported by the research literature. Background: The identification and care of women with perinatal mental health problems is increasingly considered part of midwifery practice. Research suggests that many midwives lack knowledge, skills and confidence. It would be useful to be able to determine barriers and facilitators to effective clinical practice. The authors propose a framework comprising five potentially measurable domains which impact on midwives’ ability to identify, assess and care for women with perinatal mental health problems. Design: This mixed-methods review uses an innovative qualitative convergent design based on framework synthesis. Data sources: Relevant electronic databases were searched for the period from January 2007 to December 2016; 33 studies from nine countries met the inclusion criteria. Review methods: Study quality was assessed using critical appraisal tools. Study findings were mapped onto the five domains of the framework: knowledge, confidence, attitudes, illness perception and infrastructure. Findings were then synthesized for each domain. Results: All five domains are substantially represented in the literature, thus supporting the proposed framework. A number of sub-domains and relationships between domains were identified. Varying levels of knowledge, confidence, attitudes and illness perceptions were found; evidence suggests that midwives benefit from further training within these domains. Features of organisational infrastructure act as barriers or facilitators to effective care; these need to be addressed at organisational level. Conclusion: The proposed framework was confirmed and can be used to inform practice, policy and research.


INTRODUCTION
There is significant and continually accruing evidence of the deleterious consequences of perinatal mental health (PMH) problems for pregnancy outcomes (Dunkel Schetter & Tanner, 2012) and the well-being of mothers, fathers and babies (Stein et al., 2014). Identification and assessment of mental health issues can be critical in facilitating timely and appropriate liaison with relevant professionals, discussion regarding treatment and support options, and the development of management plans for the perinatal period. In the ten years since the publication of the UK National Institute for Health and Care Excellence (NICE) guidelines (NICE, 2007) for antenatal and postnatal mental health, which made recommendations and provided guidelines for the assessment and care of women with PMH problems, it would seem reasonable to expect significant changes. However, national reports (Knight et al., 2015;RCOG, 2017) continue to emphasize gaps in service provision and failures to appropriately identify and care for women with PMH problems. It seems timely, specifically in the light of significant investment in new service development in the UK and increasing political mobilisation in other countries, to seek greater understanding of the challenges midwives face, and to provide an explanatory framework which can identify why to date numerous policy drivers and guidelines have failed to translate into practice and how that could be better supported.

Background
Whilst midwives are increasingly involved in the assessment and care of women with PMH problems, evidence suggests that many express concern regarding their knowledge and skills and hence lack confidence (Noonan, Doody, Jomeen, & Galvin, 2017;Ross-Davie, Elliott, Sarkar, & Green, 2006). Research exploring women's experiences has highlighted that a lack of knowledge of PMH among health professionals can act as a barrier to women's access to care (Byatt et al., 2013;Higgins, Tuohy, Murphy, & Begley, 2016b). Low levels of confidence can have a negative impact on midwives' behaviour during assessment and care (Davis, Foureur, Clements, Brodie, & Herbison, 2012;McGookin, Furber, & Smith, 2017). Research with health visitors suggests that confidence in identifying and managing women with PMH problems is closely linked to knowledge .
Negative attitudes and stigmatisation associated with mental illness are an important issue and result in labelling people as 'different ', stereotyping, and discrimination (Mårtensson, Jacobsson, & Engström, 2014). When this is linked to the literature which attributes knowledge deficits and low levels of clinical confidence to a greater likelihood of negative attitudes toward individuals with mental health issues (Schafer, Wood, & Williams, 2011), an interesting context begins to emerge, where a complexity of factors interact to influence practitioner behaviours. Whilst this has not to date been explored in the perinatal context, evidence does seem to indicate that lack of knowledge, experience and familiarity are factors related to more negative attitudes in healthcare staff. It is feasible then to suggest that a similar context could exist for student midwives or midwives who are not specifically trained in mental health. The impact of negative attitudes can be substantial and affects understanding of PMH problems (McGookin et al., 2017) and professional behaviour (Noonan et al., 2017), which may reduce the likelihood of women disclosing problems. Furthermore, women themselves often perceive their mental health problems as a stigma, which can be instrumental in preventing them from seeking help (Bilszta, Ericksen, Buist, & Milgrom, 2010;Byatt et al., 2013;Dennis & Chung-Lee, 2006). Further complicating the issue of effective care of PMH problems is a general acknowledgement that unless it forms an integral part of a resourced infrastructure, with clear pathways that can offer diagnostic assessment, effective and available treatment options and support, then practitioners are less likely to actively engage in the identification and assessment process (Jomeen & Martin, 2014;Noonan et al., 2017). Research into women's experiences echoes the importance of clear referral processes and integrated services (Darwin, McGowan, & Edozien, 2015;Higgins et al., 2016b;Phillips & Thomas, 2015;Rollans, Schmied, Kemp, & Meade, 2013b). Access to a specialist PMH service promotes more positive experiences (Higgins et al., 2016b;Myors, Schmied, Johnson, & Cleary, 2014).
From this narrative snapshot of the literature, it appears that five domains -knowledge, confidence, attitudes, illness perception and organisational infrastructure -are likely to impact on midwives' ability to identify, assess and care for women with PMH problems. The last two decades have seen a growing body of research in relation to the identification and care of women with PMH problems by midwives. A recent review (Noonan et al., 2017) exploring midwives' perceptions and experiences of caring for women with PMH problems concludes that midwives' knowledge, skills and attitudes were of great importance in the care for these women. Midwives need to be supported not just by further training opportunities, but also by appropriate referral pathways and a supportive infrastructure. This paper reviews a similar body of literature, but has a different focus with a different approach. The aim of this review is to systematically identify whether the international research literature supports the proposed five-domain framework in terms of midwives' clinical practice.

THE REVIEW Aims
The overall aim of this mixed methods review using a framework approach was to identify whether research into midwives' clinical practice within the identification and care of women with PMH problems confirms the proposed five-domain framework or whether it needs to be modified.

Design
This review synthesizes findings from qualitative, quantitative and mixed methods studies using a framework approach. While the review includes quantitative and qualitative evidence, the methodological and philosophical differences between the qualitative and quantitative studies (including within mixed methods studies) were not very large. The qualitative studies tended to be descriptive rather than interpretative. The quantitative studies used mostly simple surveys, with some employing a pre-/post-test design, and were also more descriptive in nature; they were not suitable for inclusion in a meta-analysis. Therefore findings were integrated by transforming quantitative into qualitative findings (Sandelowski, Voils, & Barroso, 2006). Following Pluye and Hong (2014), the approach taken can be described as a qualitative convergent design.
Framework synthesis employs a framework as a 'scaffold' against which findings from qualitative studies are mapped (Carroll, Booth, & Cooper, 2011;Carroll, Booth, Leaviss, & Rick, 2013). While Carroll et al (2013) include the systematic identification of frameworks in the literature and the subsequent generation of the a priori framework, for this review a suitable framework had already been identified. The a priori framework used in this review is based on existing research evidence and the authors' expert knowledge, as discussed above.

Inclusion criteria
Studies were eligible for inclusion if they were empirical research, written in English, published between January 2007 and December 2016, and related to midwifery practice in PMH from midwives' perspectives, i.e. with midwives as participants.

Search strategy
The search terms are shown in Table 1. Searches were conducted on PubMed, PsycINFO and CINAHL. Reference lists of identified papers were searched for further papers. After eliminating duplicates, search results were further screened by title, abstract and full text. The search was carried out in March 2017 by one of the authors (FW).

Study selection
The flow diagram in Figure 1 shows the number of studies retrieved and retained at each stage of the screening process. A total of 33 studies were included in the review. The selection of studies was carried out by one of the authors (FW) and agreed by the other authors.

Study characteristics
The eligible studies included 12 qualitative, five mixed methods and 16 quantitative studies originating from nine countries, with the majority conducted in Australia and the UK. The studies' aims, sample characteristics, methodological approaches, and findings are shown in Table 2. Aims, study design and methodological approaches varied considerably.

Quality appraisal
The quality of studies was assessed using the Critical Appraisal Skills Programme (CASP) (Critical Appraisal Skills Programme, 2017) tool for qualitative studies; the tool for cohort studies was adapted for the quantitative studies. CASP ratings and comments for individual studies are included in supplementary information table 1. No studies were excluded on grounds of quality, but ratings were taken into account during the synthesis. Quality assessments were done by one author (FW), with a sub-sample checked by another author (PJ); comparison of assessments showed good interrater agreement.

Data abstraction
Data from qualitative and quantitative studies were mapped against the framework. For mixed methods studies, qualitative and quantitative data were mapped separately. Mapping followed these steps: 1. In each paper, evidence of the domains was highlighted, using different colours for the five domains. These concepts represent distinct 'units of meaning '. 2. Quantitative findings were transformed into qualitative findings. 3. Each concept was represented by a brief quote or description. These were entered into a table, with one column for each domain and an additional column for concepts which did not appear to fit into any of the domains. Each row contained the concepts for one paper.

Synthesis of evidence for each domain
After mapping the data against the domains, findings were drawn together and integrated for each Findings were written up for each domain and sub-domain. The process of mapping concepts and writing up findings for each domain was iterative; referring back to the original papers as necessary helped to ensure that the concepts remained rooted in the original studies.

Comparison of mapped evidence and the a prior framework
The distribution and quality of evidence from qualitative and quantitative studies across all domains was then examined. Rich, in-depth evidence from a range of qualitative and quantitative sources for each domain indicates good evidence for confirmation of the framework (Carroll et al., 2013). The identification of concepts outside the five domains would suggest that the framework needs to be modified (Carroll et al., 2013).

RESULTS
Detailed findings are available in supplementary information table 3.

Knowledge
This domain relates to knowledge of PMH issues, knowledge of assessment, treatment and referrals, and interpersonal skills in interactions with women. There is a clear overlap with other domains, particularly confidence and illness perception. There was considerable variation in midwives' levels of PMH knowledge. Overall, knowledge was good for most conditions, particularly PND, but there were notable gaps, particularly for more severe conditions and for risk factors and consequences of PMH problems Higgins et al., 2016a;Jarrett, 2015;Jomeen et al., 2009;Phillips, 2015). Many studies examined self-rated knowledge, which tended to be more positive than assessed knowledge but is likely to be less reliable. Training increased levels of knowledge McLachlan et al., 2011).
Midwives used a range of assessment approaches (Gibb & Hundley, 2007;Yelland et al., 2007), including informal and adapted techniques (Jarrett, 2014;Rollans et al., 2013a;Salomonsson et al., 2011;Williams et al., 2016). Assessment of depression was more accurate than for other conditions . Assessment skills were improved by training Jardri et al., 2010;Yamashita et al., 2007). Knowledge of treatment and referrals varied, but was found to be poor in some midwives McCauley et al., 2011;Wan et al., 2008;Williams et al., 2016). Interpersonal skills, particularly good communication skills and building rapport, were considered very important in facilitating disclosure and care McLachlan et al., 2011;Mivšek et al., 2008;Rollans et al., 2013a;, but there was limited information in the included studies on the adequacy of interpersonal skills.

Confidence
Not many of the papers discussed confidence explicitly; it is, however, implicit in other domains and there is an overlap with knowledge and to a lesser extent with attitude. Although there were some exceptions, there was a general lack of confidence among midwives in the identification, care and referral of women with PMH problems (Edge, 2010;Gibb & Hundley, 2007;Jomeen et al., 2009;Lees, 2009;McCauley et al., 2011;Phillips, 2015;Reed et al., 2014;, particularly with respect to women with serious mental health problems (Jarrett, 2015;McCauley et al., 2011;. Confidence appeared to be higher in the assessment of women than in caring for women (Jarrett, 2015;. Lack of confidence can lead to increases in referrals to other services . Training increased levels of confidence McLachlan et al., 2011;Williams et al., 2016), as did previous experience , although Jomeen and colleagues (2009) suggest that taught, rather than accrued, conceptualisations of PMH problems are more likely to be effective in increasing confidence.

Attitude
This domain has two sub-domains: (1) midwives' attitude to their role within PMH and (2) their attitude towards women with PMH problems. Midwives' attitudes to their role varied, with some studies revealing partially conflicting views, particularly in the case of more severe PMH problems.
Midwives were more likely to consider assessments as part of their role, but felt that care and treatment, particularly of women with severe problems, was the responsibility of other professionals . Attitudes to women with PMH problems were mixed.
Attitudes are hard to measure accurately due to subjective perceptions and the impact of social desirability bias. Negative attitudes are often not expressed overtly but tend to be hidden or subconscious . Many midwives had positive attitudes towards women with PMH problems, but stereotyping and stigma were also evident Higgins et al., 2016a;Jarrett, 2014;McCauley et al., 2011;Phillips, 2015). In two of the qualitative papers, midwives talked about the reluctance of midwifery generally to adopt a more psychological perspective, rather than a focus purely on physical health (Phillips, 2015), and suggest that mental health is still regarded as a taboo subject (Lees, 2009). Several studies showed that training can help to reduce stigma and increase understanding Higgins et al., 2016a;Reed et al., 2014).

Illness perception
In the included studies, illness perceptions related to contributing factors and prevalence, symptoms and consequences, and treatment and recovery. The accuracy of illness perceptions varied between studies. Illness perceptions appeared to be more accurate for more historically acknowledged conditions such as postnatal depression, which have received more emphasis in training and clinical practice. Perceptions of antenatal depression were less accurate, which may reflect less emphasis in training on depression in pregnancy . Perceptions were also less accurate for more severe conditions, especially in terms of contributing factors and consequences (Jarrett, 2015;. There was evidence of lay or 'common sense' perceptions, particularly in terms of factors contribution to PMH problems . This underlines the importance of adequate training in PMH issues, including severe conditions. One of the included studies illustrates the importance of illness perceptions in the context of antenatal depression: midwives demonstrated varied levels of understanding of the condition, which impacted on identification and led to failure to make referral decisions, despite the majority of midwives believing that antenatal depression was treatable .

Infrastructure
Sub-themes in this domain were time pressures, referral pathways and further services, the organisation of maternity care, training and support for midwives. Many midwives said that time pressures made assessment and care more challenging and could be a barrier to effective assessment (Edge, 2010;Lees, 2009;Nithianandan et al., 2016;Phillips, 2015;Rompala et al., 2016).
Some felt that caring for women with PMH problems required additional time Mivšek et al., 2008). Time pressures were particularly acute in the initial appointment (Lees, 2009;Nithianandan et al., 2016). Many of the studies underlined the inadequacy of referral pathways and a lack of appropriate services for referral. Several studies highlighted the importance of good communication between health professionals and appropriate integration of services for effective care Mivšek et al., 2008;Phillips, 2015). Continuity of care was identified as being important for the provision of effective care as it enabled midwives and women to build up a relationship and makes it easier for midwives to identify problems (Gibb & Hundley, 2007;Mivšek et al., 2008;Reed et al., 2014;Wan et al., 2008;Williams et al., 2016;Yelland et al., 2007).
Lack of adequate training and the need for further training were common themes in many studies Jomeen et al., 2009;Lees, 2009;McCauley et al., 2011;Phillips, 2015;Wan et al., 2008). As one of the studies  showed, students who have just started a midwifery course are likely to use lay conceptualisations of mental disorders, highlighting the need for comprehensive PMH training within the curriculum. Several studies included in this review found that training improved confidence, knowledge and skills and reduced stigma among midwives Higgins et al., 2016a;Jardri et al., 2010;McLachlan et al., 2011;Reed et al., 2014;Williams et al., 2016). This needs to be treated with some caution as in many studies knowledge and skills were self-rated and any increase in knowledge and training was assessed soon after training and may not necessarily be sustained over a longer time period. It is promising that overall the later studies in this review suggest higher levels of knowledge compared to the earlier studies, indicating a general improvement in training and knowledge. A few papers emphasized the emotional impact of caring for women with PMH problems on midwives and the need for both formal and informal support in order to enable midwives to provide good quality care Salomonsson et al., 2011).

Support for the framework
There was rich evidence for all five domains of the a priori framework (Figure 2) in qualitative and quantitative studies, ranging from 19 studies (illness perception) to 29 studies (knowledge) ( Table 3).
There was some overlap between several of the domains and consequently the allocation of some concepts to domains has been ambiguous. In particular, there was considerable overlap between the domains of knowledge and illness perception; in many cases it was difficult to decide to which domain concepts should be assigned. In terms of the knowledge domain, this was the case for knowledge of PMH conditions and treatment options, rather than interpersonal skills. However, overall the domains are distinct enough to represent separate components of the framework.

Additional domains
Originally, several concepts were identified as potentially being outside the five domains. One of these was midwives' interpersonal skills in the interaction with women. The decision was made to include this in the knowledge domain, but it is important to note that it constitutes a separate subdomain to 'knowledge about PMH issues and 'knowledge of assessment, treatment and referrals'.
This raises some questions about the content of the knowledge domain. The emotional impact on midwives of caring for women with PMH problems and midwives' need for support were also originally considered outside the five domains. However, as the provision of support largely falls under organisational infrastructure, these concepts were included within the infrastructure domain. and knowledge of assessment, treatment and referral options. Knowing how is more closely related to skills in assessment and treatment, as well as interpersonal skills. The attitude domain contains two distinct subdomains: attitudes towards midwives' role in assessment and treatment of PMH and attitudes towards women with PMH problems. Illness perception can be divided into perception of contributing factors and prevalence, symptoms and consequences, and treatment and recovery.

Sub-domains
Within infrastructure the subdomains are quite distinct: lack of time, referral pathways and further services, the organisation of maternity care, training and support for midwives, though lack of time could be included with the organisational subdomain.

Relationship of domains to each other
Knowledge is a critical component of the framework and to some extent underpins confidence, attitude and illness perception. The link with confidence is relatively direct: higher levels of knowledge and skills generally increase confidence (Jones et al., 2015). Benner's 'from novice to expert' stages of clinical competence (Benner, 1982) can be used to explore the link between knowledge and confidence. Knowledge is a critical component which develops over time, is reinforced by experience and underpins performance through the five levels of proficiency. The 'novice' has little or no knowledge or experience in the situation in which they are expected to perform and as a consequence lacks confidence (i.e. student midwife). As the practitioner moves through the stages of clinical competence, gaining knowledge and experience, confidence is enhanced and facilitates conscious, analytical contemplation of the problem. A proficient practitioner has a perspective on the whole problem and hence is able to take into account the holistic picture which improves clinical decision-making and efficiency. The 'expert' ultimately is able through knowledge, confidence, experience and deep insight to ensure accurate identification of the problem, even in situations where she has no experience of that particular problem (i.e. experienced midwife with specialist knowledge of PMH problems).
Knowledge is also linked to attitude, as improved knowledge of PMH conditions, treatments and referral options may make it more likely that midwives recognise that assessment and care of women with PMH problems are part of their role Noonan et al., 2017). Higher levels of confidence are likely to result in midwives having a more positive attitude towards their own role in the assessment and care of women with PMH problem. However, attitudes to midwives' role in PMH also go beyond individual levels of knowledge and confidence, as evidenced by the two papers in which midwives talked about a reluctance in midwifery generally to take a more psychological perspective (Lees, 2009;Phillips, 2015). While this is likely to be partly due to a lack of confidence and knowledge, it may also be related to the dominance of the biomedical approach in maternity care which prevails despite challenges from the alternative humanistic normality paradigm (Brubaker & Dillaway, 2009). Better knowledge, and more accurate illness perceptions, can improve attitudes towards women with PMH problems by counteracting negative stereotypes and stigma and reducing fear. This is supported by a number of studies included in this review which suggest that training, and therefore increased knowledge, helps to reduce stigma and increases understanding Higgins et al., 2016a;McLachlan et al., 2011;Reed et al., 2014).
There are number of links between knowledge and organisational infrastructure. Clear, readily available referral pathways and information about further services are likely to increase knowledge of referral and treatment options (Jomeen & Martin, 2014;Noonan et al., 2017). There is good evidence that effective training, both as part of midwifery education and through ongoing professional development, increases knowledge of all types Higgins et al., 2016a;Jardri et al., 2010;McLachlan et al., 2011;Reed et al., 2014;Williams et al., 2016). Training also increases confidence through increased knowledge and the presence of clear referral pathways is likely to increase midwives' confidence in referring women for further treatment (Hogg, 2013). All infrastructure subdomains are likely to have an impact on midwives' attitudes towards their own role in PMH. Lack of time and support and the absence of effective referral pathways and further services may lead to midwives feeling that the identification care of women with PMH problems is beyond the scope of their role (Byatt, Simas, Lundquist, Johnson, & Ziedonis, 2012;Hogg, 2013).

Strengths and limitations
This review used an innovative approach to conduct a mixed methods systematic review using framework synthesis. It was strengthened by the inclusion of a wide range of qualitative and quantitative research studies conducted in a number of countries using different perspectives and research conditions. However, the different contexts and methodologies of studies included in the review made comparison of the studies difficult. Some studies were of methodological quality, with relatively low sample sizes and often high drop-out rates. The frequent use of midwives' self-report assessment of knowledge and confidence and of non-validated questionnaires was also problematic and could undermine the accuracy of the findings from these particular studies. This framework can be used to systematically identify gaps in knowledge, confidence, attitudes, illness perceptions, and training. It can also aid in the identification of barriers to effective care in terms of factors related to organisational infrastructure. In research, the framework can facilitate a conceptual exploration of facilitators and barriers to the identification and care of women with PMH problems. Furthermore, the framework has already aided the development of validated measures assessing mental health awareness in midwives and student midwives (Martin, Jomeen, & Jarrett, 2017) and professional issues in maternal mental health (Jomeen, Jarrett, & Martin, 2018  Yamashita, H., Ariyoshi, A., Uchida, H., Tanishima, H., Kitamura, T., & Nakano, H. (2007). Japanese midwives as psychiatric diagnosticians: application of criteria of DSM-IV mood and anxiety disorders to case vignettes. Psychiatry and Clinical Neurosciences, 61(3), 226-233. doi:10.1111/j.1440-1819.01659.x Yelland, J., McLachlan, H., Forster, D., Rayner, J., & Lumley, J. (2007. How is maternal psychosocial health assessed and promoted in the early postnatal period? Findings from a review of hospital postnatal care in Victoria, Australia. Midwifery, 23(3), 287-297. doi:http://dx.doi.org/10.1016/j.midw.2006

Impact statement
This review adds to the body of knowledge around midwives' practice in the identification and care of women with perinatal mental health problems. The authors propose a framework comprising five domains: knowledge, confidence, attitudes, illness perception, and organizational infrastructure. Relevant research is reviewed to further develop the framework and confirm its domains. The proposed framework, which is supported by the evidence synthesized in the review, can be used to support clinical practice by identifying gaps in effective care, training needs and elements of organisational infrastructure required to support midwives in their role. The framework can also be used in research and the development of relevant measures and can be applied to other conditions and groups of health professionals.     Just over a third of student midwives felt that they had very good knowledge of PND.
Response to 'One of the main reasons women experience mental illness is a lack of self-discipline and will power': 93% negative pre-survey (7% neutral), 100% negative post-survey (p. 367) Participants' perceptions of the amount of control women have over PMH problems was fairly accurate before the module, but improved even more afterwards. A large majority of midwives said they needed more training for PMH problems.
from being stigmatised, more than 60% said other midwives responded negatively to women with PMH problems  43 % worried about the safety of other women and babies when caring for a woman with severe PMH problems, 12% anxious about own safety (Jarrett, 2015) Training helped to reduce stigma and increase understanding Higgins et al., 2016a;McLachlan et al., 2011;Reed et al., 2014) Illness perception

Contributing factors and prevalence
Contributing factors generally: mismatch between expectations and reality, difficulties in coming to terms with the reality of life with a new baby, exhaustion, being older, a lack of support and labour experiences (Gibb & Hundley, 2007); cultural pressures (Jarrett, 2014;Phillips, 2015); social and economic factors, relationship problems, age, isolation and certain personality traits ; low socio-economic and refugee status (Phillips, 2015); history of abuse or trauma and a personal or family history of mental illness  Contributing factors for antenatal depression: relationship problems, psychiatric history, low social support, unintended pregnancy, hormonal imbalances, previous pregnancy loss, and stress or worry  Some inaccurate perceptions around contributing factors Jomeen et al., 2009;Mivšek et al., 2008) Almost half underestimated prevalence of antenatal depression (Jones et al., 2011) A fifth of student midwives underestimated risk of developing puerperal psychosis; only a quarter aware that a woman was more likely to develop puerperal psychosis if she had it previously (Jarrett, 2015) More than two-thirds thought that the incidence of severe fear of childbirth had increased over the last 10 years; two-thirds believed fear of childbirth is different from other phobias  Symptoms and consequences Some believed women often displayed extreme or obsessive behaviours with respect to themselves, their baby or the house (Gibb & Hundley, 2007) Student midwives considered women's environment, attitude, appearance and behaviour to be indicators of psychological well-being (Jarrett, 2014) Midwives believed some sadness, caused by hormonal changes, is experienced by almost all women after birth, but 'real' depression is not very common; postnatal depression was believed not to start until after the third month postnatally  Midwives said it was sometimes difficult to distinguish between symptoms of early pregnancy and symptoms of depression  Generally appropriate illness perceptions of antenatal depression; mostly good awareness of symptoms, but perceived large overlap with symptoms experienced frequently by non-depressed pregnant women  A third of student midwives believed puerperal psychosis had gradual onset in first six months after giving birth (Jarrett, 2015) Not coping with requirements of daily life and the new baby seen as central (Gibb & Hundley, 2007) More than half were aware that PMH problems could lead to attachment problems; most believed women were likely to recover eventually  Most aware that antenatal depression could have major consequences for women, but knowledge of specific consequences less accurate and tended to be based on knowledge of postnatal depression  Almost all underestimated proportion of women with antenatal depression who attempted suicide after birth  Treatment and recovery Recommended mostly support, advice and self-help groups for mild problems; considered referral to other services more appropriate for severe problems  Less likely to suggest antidepressants for antenatal than postnatal depression; more likely to suggest self-help techniques and additional support for antenatal depression, recommended seeking help from specialists for postnatal depression  Third of student midwives not aware that effective treatment for psychosis requires medication and hospital admission (Jarrett, 2015) Student midwives aware that outcome for women with schizophrenia favourable if treated, but lacked some understanding of consequences of not obtaining professional help; understanding largely based on lay perceptions of treatment  Third of midwives felt a visit to the labour ward decreases fear of childbirth, over two-thirds believed specialist fear of childbirth team helped to reduce incidence, almost 80% thought that making a birth plan was beneficial (Salomonsson et al.,

Infrastructure
Lack of time Time pressure makes it more difficult to identify PMH problems (Edge, 2010;Lees, 2009;Nithianandan et al., 2016;Phillips, 2015) Looking after women with PMH problems requires additional time Mivšek et al., 2008) If problems are identified, more time needed to discuss options and referrals (Phillips, 2015) Time pressure most acute in initial appointment as already very busy (Lees, 2009;Nithianandan et al., 2016) Time pressures mean that focus is on women with serious mental illness, leaving women with less severe problems without help they need (Edge, 2010) Almost half said lack of time can be barrier to screening for antenatal depression  More than half said that of time sometimes makes it difficult to provide quality care for women with depression; 60% reported having enough time to assess emotional health