The MRC described complex interventions as: 1) including several interacting components; 2) sensitive to the context in which they are delivered; 3) having a causal chain linking the intervention to outcomes; 4) having a range of possible outcomes (Craig et al. 2008). It was considered that a new intervention would need to operate within different maternity settings and be delivered to different populations of pregnant women. The choice of intervention components should include consideration of how the mechanisms of change would function within the context of maternity care structures and propose ways the mechanisms would influence women’s symptoms of anxiety. Therefore, the intervention was considered as ‘complex’ and the stages of the intervention development followed the general principles outlined by the MRC theoretical and modelling phases for complex interventions (Craig et al. 2008) (Figure 1).
Identifying the evidence base
The MRC state that the development of a complex intervention should begin by identifying the relevant, existing evidence base (Craig et al. 2008). Existing reviews which have evaluated the effectiveness of interventions on anxiety outcomes in pregnancy have focused on depression, mind-body or pharmacological interventions or included women with severe anxiety. Therefore two systematic reviews were completed to identify the evidence base for non-pharmacological interventions delivered to women with mild to moderate anxiety in pregnancy (Evans et al. 2017, 2019). The systematic reviews asked the following questions:
- How effective are non-pharmacological interventions in reducing the symptoms of mild to moderate anxiety in pregnancy?
- How acceptable and beneficial are non-pharmacological interventions for reducing the symptoms of mild to moderate anxiety?
The two systematic reviews concluded that interventions, specifically designed to support pregnant women with mild to moderate anxiety have mainly been evaluated in small scale studies. Studies evaluated different intervention designs for different populations and overall results were inconclusive regarding intervention effectiveness. Although no particular design which could be directly recommended for clinical practice was identified, the synthesised review findings helped identify components likely to increase the effectiveness and acceptability of the intervention.
There was some evidence of benefit for group interventions; women valued the opportunity to share experiences, reducing feelings of isolation and accessing group support (Bastani et al. 2005, Bogaerts et al. 2013, Breustedt & Puckering 2013, Dunn et al. 2012, Goodman et al. 2014, Guardino et al. 2014, Faramarzi el a. 2015, Satapriya et al. 2013, Vieten 2008, Woolhouse 2014). Some women felt they benefitted from having an individual discussion with their healthcare professionals (HCP) (Burgha et al. 2015, Côté-Arsenault, et al. 2014). Women were motivated to self-select into intervention studies however, some had concerns about disclosing anxiety symptoms and joining groups. There was some evidence of benefit for multi-session interventions and women identified group sessions as helpful once groups became established. Studies which reported an improvement in anxiety scores included group mindfulness (Guardino et al. 2014), mindfulness based cognitive therapy (Faramarzi el a. 2015), motivational interviewing (Bogaerts et al. 2013), relaxation (Bastani et al. 2005, Chang et al. 2008, Teixeira et al. 2005, Ventura et al. 2012) or CBT interventions (Milgrom et al. 2015). Women welcomed interventions which presented options for managing their symptoms and included peer and professional support (Evans et al. 2019).
Identifying appropriate theory
The theory underpinning the potentially beneficial intervention components as identified in the two reviews were explored (Table 1). This process strengthened the rationale for the final intervention design and helped to define the process of change in relation to anxiety symptoms in pregnancy (Moore et al. 2015). The development of complex interventions requires researchers to develop an awareness of the relevant theory underpinning intervention components to increase the likelihood of the effectiveness of the intervention design (Craig et al. 2008, Garber & Weersing 2010). A description of the intervention’s underlying theoretical basis should include specific theories, theoretical positions, and frameworks as well as empirical evidence which may have been conducted in different settings or countries (Mohler et al 2015).
Social support theory
Social support may have a positive effect on wellbeing, such as providing: 1. compassion, reassurance and a sense of self-worth; 2. access to new contacts and information to help develop problem solving skills; 3. reducing feelings of uncertainty and develop a sense of control; 4. providing instrumental support to reduce the frequency and duration of stressors; 5. influencing positive health behaviours (Heaney & Israel 2008). Social support pathways include components of experiential knowledge; social learning theory; social comparison theory and the helper-therapy principle (Salzer & Shear 2002). Individuals resolve their problems through sharing their experiences of mental illness with others who are experiencing similar situations (Borkman 1999) and can benefit by learning from others with who have succeeded in managing their symptoms (Simoni et al. 2011).
Therapeutic relation theory
Collaborative therapeutic relationships enable pregnant women to feel physically and psychologically supported which facilitates confidence building and self-efficacy (Carolan & Hodnett 2007). Continuity of carer from a midwife known to the woman throughout pregnancy and the intrapartum period has been associated with improved health outcomes for women and babies (Sandall et al. 2016). Benefits include an increased sense of trust, choice and control. Social influence theory recognises that the HCP’s may be seen as a source of social power due to their access to information, resources and services. While this may be beneficial, it is also associated with negative outcomes if individuals are influenced or coerced into compliance to gain access to services or information. Excessive information seeking and reassurance seeking are common features of anxiety disorders and can have a negative impact on outcomes and the practitioner–service user relationship (Osborne & Williams 2013). A pregnant woman with health anxiety may continually or excessively seek reassurance about fetal growth, the progress of their pregnancy and about the birth (Bayrampour et al. 2016). HCPs need to be aware of possible service user motivations for seeking reassurance about their health and wellbeing and suggest strategies, such as CBT, to help modify negative behavioural patterns (Williams 2012).
Awareness of mind and body experiences enables an individual to direct their attention to their breathing or another object of focus, to prevent elaborative ruminative thought processing (Gard et al. 2014, NurrieStearns & NurrieStearns 2013). Mind-body interventions like yoga, guided imagery, mindfulness or hypnotherapy may be effective for reducing anxiety as they are thought to induce mental relaxation and alter negative thinking related to anxiety ((Marc et al. 2011). Mind-body approaches are intended to modify an individuals’ perceptions of stressful events which can lead to improvements in adapted behaviour and develop coping strategies (Marc et al. 2011). The relaxation response is thought to counteract the stress response of anxiety. Physiological mechanisms and adjustments are activated when an individual engages in repetitive mental or physical activity and is able to passively ignore anxious thoughts (Manzoni et al. 2008).
In the treatment of anxiety disorders, the aim of CBT is to reduce anxious feelings by undoing prior learning or by providing new, more adaptive learning experiences, changing cognitive and behavioural responses to anxiety (Williams & Garland 2002). Increasing an individual’s awareness of unwanted emotions and behaviours is thought to generate a number of alternative responses. This helps the individual to decide on a course of action and monitor the outcome to re-enforce positive coping strategies (Brewin 1996). CBT for anxiety disorders may include components of:
- Psycho-education on the nature of fear/anxiety.
- Cognitive restructuring to challenge the truth of anxious thoughts and develop alternative thoughts to better reflect their experience (Brewin 1996).
Behavioural exposure components of CBT require further consideration in the context of pregnancy. There are very few studies evaluating exposure-based CBT due to concerns around potential harm to the fetus (Arch et al. 2012, Lemon et al. 2015).
Many of the interventions identified in the systematic reviews had multiple components: psycho-education; relaxation; peer support; and professional support. This multi-component approach was reflected in the interconnected theoretical approaches which underpinned existing intervention components. For example, CBT techniques are often incorporated within therapeutic relationship approaches and can be accessed as a resource within peer support models.
A theoretical model was developed to map the potential mechanisms and their usefulness in meeting the needs of pregnant women with symptoms of mild to moderate anxiety (Figure 2). Exploring the theoretical base highlighted that positive change can occur though: 1. developing collaborative relationships with women which aim to promote women’s choice and control over their care. 2. receiving support from HCPs who both understand women’s individual needs and can also help them access services; 3. accessing support and learning from other women who have experienced / are experiencing similar feelings or situations; 4. developing strategies to help women develop an awareness of their thought processes and learn techniques to improve the way they cope with anxiety. Mind-body and/or CBT approaches were considered as appropriate components of the intervention design.
Additional considerations and motivations informing the intervention design
In response to the increased focus on the role of the midwife to support the psychological and emotional wellbeing of women in pregnancy (MMHA 2013), the development work explored ways in which women could be supported by midwives within midwives current scope of practice (Nursing & Midwifery Council 2013). It was considered that a midwife could facilitate peer groups, acting as a resource to the women. Midwife facilitation may be more appropriate when groups are establishing, however the role of the professional in peer groups should not interfere with the potential benefits derived when group members help others in the group (Brown & Lucksted 2010). In maternity care, the role of the HCP in breastfeeding support groups has been reported to “normalise or counteract extreme views and help women to distinguish between fact, anecdote and myth” (Hoddinott et al. 2006, page 143). In a group based antenatal care study (Andersson et al. 2012), women welcomed midwives who were less structured in their approach to group facilitation. They appreciated midwives contributing their expertise in antenatal care and helping to address topics women found difficult to introduce. To maximise the benefit of social learning mechanisms, women may benefit from hearing the experiences of other women who have been through similar experiences who can share their success stories and inspire hope (Davidson et al. 2012, Miyamoto & Sono 2012, Repper & Carter 2011). Women who feel isolated in pregnancy or have poor social support may benefit from peer group approaches, however some women may not feel confident to share their situations or feelings within a group. Women may have additional pregnancy related or mental health concerns which they would prefer to discuss individually with a midwife who can provide maternity expertise and support referrals or signposting to other specialist services such as Increasing Access to Psychological Therapies (IAPT). The options for the delivery of the intervention components, considering the feasibility of employing midwife facilitators and facilitator training requirements were mapped (figure 3). The advisory group raise concerns that the training to deliver CBT and mindfulness-based interventions was intensive, with training usually taking one year or more to complete. Also, at the time of the study development readily accessible training courses were not focused on delivering interventions to pregnant women. Recent studies have highlighted the effectiveness of interpersonal psychotherapy and CBT interventions to prevent postnatal depression which can be delivered by nurses, midwives and health visitors in antenatal care settings and requiring brief initial training (Jensen et al, 2018; Johnson et al 2018), and a brief midwife-led CBT intervention for maternal anxiety is in progress (Wilkinson et al. 2016). For this intervention, it was considered that the therapeutic intervention components (mind body and cognitive behavioural approaches) could be delivered through supported use of self-help resources. Guided self-help has been reported as an effective intervention for depression and anxiety in general populations (Seekles et al. 2011) and has been used as a stand-alone intervention or alongside group interventions for pregnant women with anxiety, stress and depression (Evans et al. 2017). Potential self-help resources were identified evaluated using IAPT criteria (IAPT 2010).
Modelling process and outcomes
For this study, potential intervention components and processes were tested through consultations with a study advisory group and a maternity research public involvement group. The advisory group consisted of the head of nursing and midwifery research at the local NHS trust, a community psychiatric nurse, a midwife manager, a service user, consultant clinical psychologist and mental health training providers. Service users provided insight into how the intervention would be accessed and used and ensured the intervention was relevant to the needs of pregnant women (INVOLVE 2012, Moss et al 2016). Both groups supported the proposed intervention components and helped to identify methods of delivery for the intervention which considered: the context and methods for introducing the intervention, assessing eligibility, method of delivery and facilitation of peer groups; and delivery of the therapeutic components. Rather than having two midwife facilitators, service managers identified that a midwifery support worker (MSW) could provide support to the midwife during the groups and co-facilitate the intervention. A bespoke training framework was developed for midwives and MSWs which referred to existing perinatal competency frameworks (NHS Education for Scotland 2006, NHS England & The Tavistock and Portman NHS Foundation Trust 2016). Experienced mental health training providers developed a three day training workshop which included a range of educational and learning approaches e.g. role play, lectures and the completion of an information and reflective workbook.
Following the evaluation of the evidence base, exploring the theoretical base and consultations with stakeholder groups, a protocol was developed for the intervention (Mohler et al. 2015, figure 4 & table 2).
The MRC (Craig et al. 2008) state that the future implementation of the intervention needs to be considered at an early stage of development. This should ask questions about whether implementation would be possible, who the key stakeholders are and what information they may need to implement changes in practice. De Silva et al. (2014) proposed that the current MRC guidance could be strengthened by incorporating Theory of Change (TOC) into the design and evaluation of complex interventions To help identify the intervention processes and success indicators a TOC map was developed (figure 5). TOC defines how and why an initiative works, providing a pragmatic framework to describe how the intervention affects change (Weiss 1995, de Silva et al 2014). Each pre-condition for the intervention is evidence based and measured through an indicator. The TOC can help reduce future implementation failures as weak links in the causal pathway can be tested, revised and strengthened. The TOC map set out to answer a series of questions which asked how the intervention could be integrated into routine practice and identifying how the intervention could be empirically tested in future definitive research (Bonell et al. 2012, de Silva et al. 2014).