Recovery from postpartum psychosis: a systematic review and metasynthesis of women’s and families’ experiences

Postpartum psychosis is a serious disorder that can result in adverse consequences for the mother and baby. It is important that we understand the experiences of women, to develop effective interventions during this critical period. The aim of this systematic review was to conduct a metasynthesis of qualitative research exploring women’s experiences of postpartum psychosis and factors involved in recovery from the perspective of women and family members. A comprehensive literature search of five databases was conducted and the findings were appraised and synthesised, following a thematic synthesis approach. Fifteen studies, capturing the views of 103 women and 42 family members, met the inclusion criteria. Four main themes incorporating 13 subthemes were identified following synthesis: (1) Experiencing the unspeakable, (2) Loss and disruption, (3) Realigning old self and new self and the integrative theme of (4) Social context. The findings offer new insight into the unique experience of postpartum psychosis and demonstrate that recovery does not follow a linear path. To improve clinical outcomes, a more integrative and individualised approach is needed which incorporates long-term psychological and psychosocial support, and considers the needs of the family. Further areas for staff training, service development and future research are highlighted.


Introduction
Postpartum psychosis (PP) is one of the most severe mental health problems following childbirth, affecting 0.89 to 2.6 per 1000 women (Vanderkruik et al. 2017). It is characterised by a dramatic onset, and rapid deterioration, with symptoms including hallucinations, delusions, disorganised behaviour and depression (Di Florio et al. 2013;Heron et al. 2008;Sit et al. 2006). Continued poor maternal mental health is associated with increased risk of adverse outcomes including suicide and decreased mother-infant bonding (Wilkinson et al. 2017). Immediate referral to a secondary mental health service for assessment is therefore recommended (NICE 2018). Inpatient care, ideally within a Mother and Baby Unit (MBU; Gillham and Wittkowski 2015) and pharmacological intervention, is almost always required (NICE 2018;Jones and Smith 2009). Family members are reported to be integral during this process and should be involved in care and treatment planning (Engqvist and Nilsson 2014;Mohamied 2019).
With appropriate treatment, the most severe symptoms usually resolve within 2 to 12 weeks (Bergink et al. 2015). However, women remain at increased risk of subsequent postpartum and non-postpartum episodes, including depression and anxiety (Robertson et al. 2005;Nager et al. 2013). PP is a highly stressful life event which can have a detrimental impact on well-being and long-term functioning and result in feelings of guilt, loss, fear and shame (Burgerhout et al. 2017;Plunkett et al. 2016;Wittkowski et al. 2014). Psychological intervention and psychosocial support has been recommended (Doucet et al. 2011), yet very little is known about the psychological factors underpinning recovery or the types of intervention found to be effective.
Previous metasyntheses have explored individuals' experiences of serious mental health problems and mental health services in the postnatal period (Ruffell et al. 2019;Smith et al. 2018;Plunkett et al. 2016;Megnin-Viggars et al. 2015;Wittkowski et al. 2014;Dolman et al. 2013). However, given the range of maternal mental health disorders included in those reviews, the applicability of these findings to women's experiences of PP remains unknown. Thus, this systematic review and metasynthesis of published qualitative studies was conducted with the aims of (a) synthesizing the reported experiences of PP from the perspectives of women and family members and (b) identifying factors involved in recovery. The outcomes of which should be used to inform policy, practice and intervention development (Moore et al. 2015;Tong et al. 2012).

Method
A metasynthesis was deemed most appropriate to systematically analyse and interpret multiple qualitative studies with the aim to develop new insight (Lachal et al. 2017

Inclusion and exclusion criteria
All published empirical studies exploring women's or family members' experiences of PP and/or recovery using a qualitative methodology were included. Studies published in any language were eligible for inclusion, as were studies using mixed methods, provided qualitative data could be extracted.
Studies were excluded when they (1) explored 'schizophrenia' and other perinatal disorders, such as postnatal depression; (2) did not stipulate the diagnosis; (3) used a mixed sample in which it was not possible to differentiate the findings based on diagnosis; (4) explored only health professionals' views; or (5) were unpublished research or grey literature.

Search strategy
Five databases CINAHL, MEDLINE, EMBASE, PsychINFO and Web of Science were searched from inception to 3 April 2019. An updated search was completed on 20 August 2019 which revealed no additional eligible studies. Only keywords relating to the 'phenomenon of interest ' (Cooke et al. 2012) were applied to ensure all relevant studies were captured ('postpartum psychosis' OR 'puerperal psychosis' OR 'postnatal psychosis' OR 'psychosis after childbirth') and no limits were applied. Additional studies were sought from the reference lists of included studies using forward and backward searching (Horsley et al. 2011).

Study selection
The search process, following PRISMA guidelines (Moher et al. 2009) yielded 1782 unique studies (see Fig. 1). The title, keywords and abstracts of all studies were assessed for eligibility against the inclusion/exclusion criteria; 1728 studies were subsequently excluded. A random 20% of studies (n = 353) were checked by an independent rater and a strong level of inter-rater reliability was obtained (96%, κ 0.88 (95% CI, 0.82-0.95) McHugh 2012). Of the remaining 54 studies, 39 were excluded when reviewed in full, yielding a final sample of 15.

Critical appraisal
The methodological quality of the included studies was evaluated using the Critical Appraisal Skills Programme (CASP 2018) checklist. The CASP comprises ten questions across three broad domains, relating to the validity, design and utility of the results. The CASP was used to determine rigour and trustworthiness in the outcomes (Hannes 2011).

Data analysis
All included studies were extracted electronically onto NVIVO 12 (2018) data management software and analysed using Thomas and Harden's (2008) thematic synthesis (see Table 1). This approach enabled the researchers to integrate multiple qualitative findings and identify recurring themes, following three key overlapping stages (Noyes et al. 2018;Ring et al. 2011). Distinctions were made across the different samples when analysing the data and constant comparisons were made across and within the women's and family members' accounts to ensure similarities and differences were captured. To test the credibility and applicability of the findings, a peer and member-checking exercise was completed during theme development stage (Lincoln and Guba 1985;Creswell and Miller 2000).

Study characteristics
The qualitative synthesis comprised 15 studies reporting the experiences of 103 women and 42 family members, of which 32 were partners (see Table 2). All studies were published between 2003 and 2018. Ten studies were conducted in the UK, two in Sweden, two in the USA and Canada and one used online data (location unknown). A diverse sample was obtained, in which women who participated in the studies were reported to be at different stages of their recovery, ranging from 2 months postpartum to 32 years. No women were reported to be actively psychotic during the interview and this was frequently cited as an exclusion criteria to ensure informed consent could be obtained. Engqvist et al. (2011) used online narratives as a data source and therefore the duration since onset could not be determined. The duration since onset reported by family members was also broad, capturing experiences up to 19 years postpartum. One study (Boddy et al. 2016) specifically sought individuals' experiences during the acute phase of psychosis and this was obtained from the perspective of fathers only.
Using the CASP checklist, no major methodological issues were identified and all 15 studies were rated to have high methodological quality and low methodological bias (see Table 3).

Synthesis
Four main themes and 13 subthemes were conceptualised in the metasynthesis, representing the reported experience of PP and factors involved in recovery from the perspective of women and family. A conceptual model was developed (see Fig. 2) to provide a visual representation and to illustrate the relationships between themes. This captured themes from both women and family members, due to the similarities found in some Table 1 Thematic synthesis  process Stages of thematic synthesis 1 Free line-by-line coding completed inductively across all data contained within the 'findings/results' sections of each study, including participant quotes, author narrative and interpretation. The women's accounts were coded first, followed by those of family members. New concepts were created according to the meaning and content of each sentence. 2 These codes were then organized into related areas and descriptive themes were developed, using an iterative process which involved the refinement of codes, checking for consistency of interpretation. 3 Higher level analytical themes were generated with the aim of producing new interpretations and key messages as guided by the research aims. This was continuously reviewed until a final set of analytical themes that captured the key meanings across the women's and family members' accounts were agreed upon.  of their experiences. The model demonstrates women's nonlinear progression towards recovery, in which they often moved back and forth between phases and this process was integrated within the wider social context. Illustrative quotes are provided below each theme with additional data presented in Table 4.

'Trapped in an insane mind'
Women reported a range of distressing thoughts, which often related to their newborn baby 'Did I kill my baby?' (Woman p.240; Doucet et al. 2012). Women reported experiencing racing thoughts and unusual perceptions; they worried that their baby would be taken away from them and felt 'trapped in an insane mind' (Woman p.380; Engqvist et al. 2011). These fears led women to mistrust people close to them and attempts to conceal their symptoms contributed to delays in families seeking help. Partners reflected that they missed earlier sign and referred to a key, sometimes 'traumatic' moment (Partner p.402; Boddy et al. 2016), where it became clear they needed help.

'Out of the blue'
Women and family members consistently reflected on the sudden and severe escalation of PP. They reported having minimal to no information prenatally, which added to their sense of being unprepared. As women and family members tried to make sense of why the woman had developed PP, many attributed the diagnosis to a biological cause involving a hormonal imbalance and factors related to pregnancy, such as sleep disturbance or a traumatic birth experience. These birthspecific attributions helped women and families to view PP as a disorder specific to childbirth that can happen 'out of the blue' to anyone (Woman p.78; Roberts et al. 2018), thereby externalising the cause and reducing selfblame. 'I feel that people often think it's postnatal depression, and they do not understand it at all. It's taken me 20 years of really banging my head against the wall trying to make people aware of it' (Woman p.77; Roberts et al. 2018). 'There was never much discussion, even from family, but certainly not from any of the medical services, as to how I was, what concerns did I have. I was very much left to feel that you have got to cope' (Partner p.6; Holford et al. 2018).

Difficulty in caring for and bonding with the baby
During the early phase of PP, many women experienced increased anxiety, reduced confidence and reported delays bonding with their babies. Women described meeting their babies' physical needs, but with a sense of detachment: I fed him, bathed him, changed him and was able to make up his feeds, but I had no real affection for him (Woman p.383;Engqvist et al. 2011).
Women reflected on how difficult they found caring for their baby, when they were extremely exhausted themselves and sedated from medication or fearful of accidentally causing harm. As a result, women often relied on family members and nursing staff to provide practical care for their baby. For some partners, providing the level of support needed for both their baby and partner became extremely challenging: I honestly don't think I could have survived much longer (Partner p.4; Holford et al. 2018).
MBU admission was often described as a relief for family members because they believed both the woman and baby were safe and being cared for.

Theme 2: loss and disruption
Women and family reported a sense of loss and disruption that permeated across multiple areas of their lives and remained with women beyond the remission of their acute psychotic symptoms.

Challenge to sense of self
The experiences associated with the acute stage of PP challenged women's sense of self and personal identity. Women and family members described the woman's presentation as being significantly out of character and resulting in feelings of loss, abandonment and fear for the future: I did not recognise her at all… it just wasn't her… I was just terrified….that I had lost her (Partner p.4;Holford et al. 2018).
Women's experiences of loss remained with them beyond recovery and women believed they had 'undergone a change that was likely permanent' and this change required acceptance (Woman p.340;McGrath et al. 2013).
'Guilt because you missed out' As time progressed, women reflected on multiple losses. In particular, they believed they had missed out on the first few months of their child's life and worried about the impact of this on their child's development. Many women and partners described guilt and disappointment resulting from unfulfilled expectations of parenthood. Some reported a sense of shame for the initial thoughts they had about their baby: 'I was appalled that I could have such an awful thought' (Woman p.382; Engqvist et al. 2011). Some parents decided against having further children due to the increased risk of a subsequent episode of PP. A decision that was often associated with additional feelings of sadness and loss: I loved being pregnant and I loved carrying a child… not to do that is heartbreaking (Woman p.420; Robertson and Lyons 2003).

Powerless
A sense of powerlessness was evident across the accounts of women and family members. In the early stages, women were reliant on family, friends and healthcare staff for information and support which often left them feeling 'helpless and disempowered' (Author p.160; Heron et al. 2012). Family members regretted things they 'should have' done and felt as though they had 'let the mother down' (Partner p.6; Holford et al. 2018), particularly in relation to their involvement in the women's hospital admission. Family members described being ignored and expressed frustration at being denied information due to patient confidentiality.

Theme 3: realigning old self and new self
Recovery was described as a lengthy process which involved women making sense of and integrating their experiences, as well as identifying personal strength and adjusting to a new view of self, in the context of being a parent.

A non-linear process
The 'very, very long journey' (Woman p.174; Posmontier and Fisher 2013) to recovery was described as following a 'progressive but non-linear trajectory' (Author p.163; Heron et al. 2012). During this process women reported anxiety, uncertainty about the future and periods of low mood and depression, which occasionally necessitated a hospital readmission. Recovery was conceptualised as an 'active process' (Author p.5;McGrath et al. 2013) in which women became motivated to develop their understanding and utilised strategies to reduce the risk of further recurrence.

Making sense and integrating the experience
The ability for women to openly disclose their experience was regarded as important for recovery. In order to make sense of why they had experienced PP, many attempted to organise events in time and sought information about the disorder. Developing their understanding helped to alleviate self-blame. Communicating with others with similar experiences and sharing their story was important to foster feelings of hope and reduce feelings of isolation and self-blame.
Women wanted counselling or psychotherapy for themselves and also the wider family to help express and come to terms with their experience. However, barriers to accessing support were reported, including lack of provision of talking therapies and different needs within the family sometimes hindered women's ability to talk through and integrate their experiences: He could not understand at times why I just could not pull myself together which annoyed me even more (Woman p.421; Robertson and Lyons 2003).

A 'return to life'
Women spoke of turning points in which they started to feel more hopeful about the future and made a decision to 'return to life' (Woman p.10; Engqvist and Nilsson 2014). Positive experiences, such as connecting with other women and bonding with the baby helped to enhance women's self-efficacy. Experiences such as these enabled women to feel like 'themselves again' (Author p.424; Robertson and Lyons 2003) and provided hope for the family that recovery was 'within reach' (Author p.13; Engqvist and Nilsson 2014).
Whilst describing the experience of PP as traumatic, some positive aspects were identified. Women reported feeling more confident or stronger in themselves: Before I just thought well, I'll coast along until whenever and life's not that bad but now I'm not really scared of anything (Woman p.424; Robertson and Lyons 2003).
Women valued those close to them, reported greater empathy and often felt motivated to help others and 'give something back' (Woman p.424;Robertson and Lyons 2003). Family members similarly reported improved relationships, increased empathy, openness and understanding, both in relation to their family member and more generally towards people experiencing mental health difficulties.

Theme 4: social context
Social context was conceptualised as an integrative theme that both influenced and was influenced by the women's experience of recovery from PP.
Family relationships-a 'double-edged sword' Throughout the process of recovery, family were concurrently viewed as a 'source of immense support and a source of worry' (Woman p.161;Heron et al. 2012). Women valued the support provided by family but simultaneously worried about their families' well-being. In the early stages, women expressed guilt for burdening their family and sometimes expressed strain within their relationships. Women reported prioritising their relationship with their child, but sometimes felt this was to the detriment of their other relationships. Partners reflected on their own stress throughout this experience, but felt unable to seek help: I was an emotional wreck but felt I had to gather myself together for my wife... I could not have any issues, someone had to be strong (Partner p.241; Doucet et al. 2012).
Psychological support for families was needed to help develop family members' understanding of PP, enhance their coping with the additional stress and enable them to know how to respond and cope with fear around relapse: If my husband had a support group for new fathers to deal with a psychotic wife, it would have changed everything. He would have been far more compassionate had he known about my illness (Woman p.241;Doucet et al. 2012).

Interaction with healthcare professionals
Interactions with mental health services were perceived by women and family members as both a facilitator and barrier to recovery. In one study, women reported a sense of pressure to appear to be coping; this was associated with a fear of having their baby removed and acted as a barrier to seeking professional input (Plunkett et al. 2016). Some women and family members reported difficulties accessing support, felt there was a divide between hospital and community services and believed there had been little consideration of their needs as a family. Women and family members occasionally felt judged by health care professionals 'You do the smallest thing and, oh no, you're doing it wrong' (Partner p.405;Boddy et al. 2016).
Some reported more positive experiences and felt 'cared for' (Woman p.384;Engqvist et al. 2011). Positive experiences improved women's confidence, promoted their relationship with their baby and provided reassurance to the family, many of whom valued the expertise and support provided by healthcare professionals.

Societal expectations of motherhood-'snap out of it'
Women and family members both highlighted the public's limited understanding of this 'hidden illness' (Woman p.77; Roberts et al. 2018) which was frequently confused with postnatal depression. The lack of awareness by professionals and peers prevented women from seeking and gaining support and left women feeling as though their symptoms were minimised or dismissed. Some women experienced an expectation from their personal network to 'snap out of it' (Woman p.261; Glover et al. 2014) which exacerbated a sense of internal shame and hindered seeking help: It's a double whammy. Not only the stigma of being mentally ill, you have got the stigma of being a mentally ill mother, a bad mum (Woman p.78; Roberts et al. 2018).

Discussion
This is the first review to synthesise the qualitative research on the experience of PP and the factors involved in recovery. The findings provide further evidence as to the extreme and distinct nature of PP (Di Florio et al. 2013) in which there are unique aspects to recovery that should be recognised and inform care provision. The findings of this review reveal that recovery can be a lengthy and non-linear process that continues beyond acute symptom remission and is influenced by the wider social context. Furthermore, the review reveals similarities in the experiences of women and family members, including a reported sense of shock, fear for the future, hopelessness and difficulties in coping and seeking help.
During the early phase, the distressing nature of PP left women and family members struggling to identify what was wrong and unable to access information. Perceived societal stigma and fear of negative repercussions delayed help-seeking behaviour; this may be associated with fear of losing their baby; however, further investigation is required. In their review looking at mental health problems in the postnatal period, Megnin-Viggars et al. (2015) similarly found that women often reached crisis point before seeking help. This finding is important, given the increasing rates of suicide amongst postpartum women experiencing psychosis (Lysell et al. 2018) and the potential negative consequence for the mother-baby relationship (Alhusen et al. 2013). The reported delays in help-seeking behaviour provide further support for NICE (2018) guidelines which highlight the importance of healthcare professional vigilance to possible symptoms of PP and provision of prompt assessment and intervention.
The findings reveal contact with healthcare services is inconsistent and can even hinder accessing support, suggesting that NICE guidelines are not being consistently met. This may reflect differences in service provision, including insufficient availability of specialist MBUs which can result in delays receiving appropriate help (Hill et al. 2019;Jones and Smith 2009). It was also found that the needs as a family were often not adequately addressed, echoing work in the broader area of maternal mental health (Megnin-Viggars et al. 2015;Dolman et al. 2013) and providing further evidence to provide healthcare staff with specialised training on PP to enhance their skills and confidence (Dolman et al. 2013).
Perceived loss has featured in previous reviews relating to severe maternal mental health disorders (Smith et al. 2018;Wittkowski et al. 2014;Dolman et al. 2013). However, for women and families experiencing PP, the sense of loss was pervasive and reported across multiple domains of women's lives. Women feared losing their child and reported loss of time, control, freedom and perceived loss of subsequent children. These losses all disrupted the women's adjustment to motherhood. Time and support were needed to reflect on these losses and to explore their new identity. Dolman et al. (2013) similarly found that women experiencing severe mental health disorders during motherhood need to integrate their 'dual identity' of being a 'woman with a mental illness' and a 'mother' for successful transition to occur. Furthermore, across the studies, women and partners reported a sense of loss in relation to their decision-making about future pregnancies. Although NICE (2018) guidelines recommend providing preconception advice for those at risk of mental health problems, the findings of this review suggest this support may need to be more proactive and ensure the emotional impact of this decision is incorporated.
The centrality of family support was clearly communicated throughout participants' narratives and confirms previous reports that family plays a 'key role' (p.3; Plunkett et al. 2016). Differences were noted in the family members' ability to cope and provide the level of support needed to manage at home. Improved understanding of postnatal mental health problems has been found to influence the relationship and can result in greater resilience and confidence by the partner (Ruffell et al. 2019). The importance of involving the family by providing information and fostering feelings of security and hope has also been reported by professionals working with women experiencing PP (Engqvist and Nilsson 2013b;Engqvist et al. 2007). However, as family perceived a need to be 'strong', this sometimes acted as a barrier to seeking additional external help for themselves.

Clinical implications
In order to improve outcomes and facilitate women's recovery, recommendations based on women's and family members' reported experiences are provided in Table 5. Additional recommendations are suggested for healthcare professionals. However, given that the included studies did not report on the views of healthcare providers, these recommendations should be treated with some caution.

Strengths and limitations
The use of thematic synthesis for this review enhanced the rigour and transparency of the analysis. It allowed for the incorporation of multiple findings and the subsequent offer of a new interpretation which can be used to inform improvements in clinical practice. The systematic search was comprehensive, minimizing language and publication biases. All studies included were of high methodological quality, thereby enhancing the trustworthiness of the data synthesised. Furthermore, steps were taken to maximize transparency and enhance rigour, including the use of data management software and independent reviewers (Tong et al. 2012).
The findings report on the experiences of 103 women and 42 family members and capture a broad range of experiences up to 32 years postpartum. This is viewed as a strength of the review, because it allowed for the analysis and interpretation of a diverse range of views and experiences. However, it was not possible to investigate individuals' experiences specifically in relation to their stage of recovery. It is therefore unclear how an individual's perspective may change over time.
Despite not limiting studies by language, only research from the UK, USA, Canada and Sweden, was identified. This means that the experiences of women and family living in other countries, including less developed countries, perhaps with different family structures and healthcare systems are unknown. Caution is therefore advised when generalizing findings. Despite applying no limits on date, no research pre-dated 2003. This may reflect the poor representation of qualitative research in medical literature (Shuval et al. 2011) which is now on the increase due to an increased recognition on its value (Popay 2006). Furthermore, the analysis is limited to the information provided in the studies included in this review. Other topics which may be pertinent, for example, regarding the types of services accessed, were often not clearly stated.

Further research
This review highlights a number of gaps in the literature in which further research is needed. Research should focus on the types of long-term psychological and psychosocial interventions that women may find helpful during their recovery from an episode of PP. Exploration may be required to further understand the professional perspective and how this can be integrated into the conceptual model. Family support could be investigated due to their influential role, with emphasis on how to support and enhance their position. It may also be helpful to investigate the experiences of women with different types of family and healthcare support, and to explore the perceived barriers and facilitators to accessing professional input. Finally, it may be useful to explore the impact of preconception counselling and to investigate the factors involved in accessing this type of support. • Accurate information should be provided, alongside access to peer support to help normalise women's experiences.
• Women could be offered psychological and psychosocial input during the latter stages of recovery not just the early stages.
• Longer term psychological needs should be considered, incorporating the reported feelings of guilt, loss and difficulties transitioning to their new perceived role. • Practical considerations are required to enable access to psychological support, this should be flexible and consider childcare provision. • Women/partners should be offered more proactive support when reaching decisions about future pregnancies.

Family
• The needs of the family should be considered and incorporated into any assessment and intervention plan as much as possible.
• Family members' wellbeing should be monitored and they should be signposted to appropriate support and peer networks as required. • Family members should be informed of reputable sources to obtain accurate information.
• Families may benefit from a therapeutic space in which they can openly explore and seek to resolve any difficulties within their relationships.

Healthcare professionals
• Specialist training and support may help to develop healthcare professionals' confidence and competence meeting the complex needs of women experiencing PP and their family. • It is important that professionals maintain a compassionate and non-judgmental stance; in order to develop a therapeutic relationship which promotes optimism and hope for the future. • Healthcare professionals should pay particular attention to women who do not have supportive family structures in place.
• Help-seeking behaviour should be targeted; for example, through the provision of accurate information (e.g. during antenatal classes) and by improving public awareness of PP.