Reflexive statement.

Reflexive accounting allows the reader of the final research product to assess the degree to which the prior views and experiences of the researcher may have influenced the design, data collection and data interpretation of the study or in this case, the synthesis of the findings of multiple studies. This review was conceived with an informed knowledge of caesarean section and a degree of professional distance, which arguably limited bias based on the teams own experiences. APB is a medical officer with over 15 years of experience in maternal and perinatal health research and public health in general, and caesarean section in particular. CK, a medical sociologist, came to the project with prior beliefs about the complexity and interdependency of social factors driving caesarean section rates, principally informed by undertaking earlier primary research with women and health professionals in the UK. SD, a Professor of Midwifery, believed that maternity care organisations are complex adaptive systems, and that the organisational ethos can exert either toxic or enhancing effects that have real consequences for staff morale, engagement, attitudes, behaviours and performance.

Supporting and challenging professional power, roles and relationships (SoFs1).

The power of the medical profession was perceived as an important barrier to overcome where doctors believed their professional identity and the safety of women was compromised by relinquishing lead professional responsibility to midwives. [50,52–54] Some midwives expressed similar concerns where midwifery confidence, skills and support were low within specific organisations [46,50,52–54] and systems. [46,47,49,52–54] As explained by this midwife in Chile, “Neither midwives nor women are empowered enough to question a medical prescription.” ([49]: p.1153). Women too reported observing the negative effects of power differentials between doctors and midwives. [61] In 11 studies, reported in 13 papers [42,46,47,49,50,52–54,57,58,60–62] interventions, including initiatives to promote physiological birth in Iran [57], hospital primary vaginal birth in the US [61], normal labour and birth in the UK [50,52–54] and the humanization of birth in Japan [46] and Chile, [49,62] challenged the structural balance of power between stakeholders. In UK organisations where a more equal balance of power did exist between women, midwives, family doctors, and obstetrician, there was some evidence that midwifery-led staffing interventions to keep birth normal and reduce caesarean section empowered midwives to work more autonomously [50,52–54] by “…sort of put[ting] a little tag on that woman as a way of saying ‘leave her alone’, which I think some doctors respect, and some don’t” ([53]:p.232). A further perspective on power, roles and relationships between stakeholders was offered by a nurse in the US who said; “My job is to empower them [women]. I don’t need to feel powerful…” ([61]:p.341).

Perverse incentives, fee exemption, fee reduction and health insurance reform (SoFs 2, 3).

Financial incentives, for hospitals, doctors, or women, either to reduce caesareans, or to increase access to caesarean section when needed, were not always perceived to have had the desired effect. In one study from China insurance reform was not believed to be as influential on caesarean section rates as women’s preferences for caesareans. [48] From low- and middle-income countries there was evidence that financing structures, in the form of fee exemption policies [28,43,44,58–60] and insurance reform, [27,29,47,48] were mediators of access to both necessary caesarean section and unnecessary caesarean section. Whether financial interventions were successful or not was mediated by local philosophies of maternity care; inter-professional and inter-personal relationships; staff motivation to work with women or with the organisation, or simply for an income; and the expectations and demands of local women, families and communities.

For example, in a study from Senegal, the intervention was government payments for each caesarean performed, with the intention of ensuring that necessary caesarean section was accessible to all. All participants in the study (including women, medical and midwifery staff) perceived all caesareanss conducted as necessary. In a highly telling interview, an administrator spoke of the increased revenue generated by this policy as the cash-cow for the hospital; the “vaches laitières des hôpitaux.” ([28]:p216) It was seen as a source of pride for the obstetric department, providing them with power and influence in the hospital as a whole. In Iran, insurance policy change was met with scepticism by health professionals, amid concerns that women who need a caesarean section may no longer get one, or that there might be a paradoxical increase in the misreporting of indications for caesarean section to satisfy amended insurance criteria. [47]

Birth environment, efficiency concerns, and organisational logistics (SoFs 4).

In 16 studies, stakeholders talked about the built environment (i.e. physical space, facilities), efficiency (i.e. time constraints on labour and staff) and/or logistical concerns (i.e. availability of equipment, theatre access) as powerful mediators of barriers or facilitators to reducing unnecessary caesarean section. [28,36,39,40–42,46,47,49,51,55–58,61,62] In high-income countries where “quick win” changes had been made to labour and delivery rooms to encourage normal labour and birth, the priority an organisation gave to maintaining them was fundamental to their effectiveness in reducing caesarean section rates. [56,61] This included changes to in-room facilities for labour and guaranteed access to operating theatres when necessary. [51,56,61] One study reported midwives’ views about how birth in a home setting reduces unnecessary caesarean section, [36], citing the absence of restrictions on women’s movements, environmental comforts, and efficiency concerns evident in the other 15 studies of institutional birth contributing to this SoFs. In middle-income countries inadequate facilities (lighting, bathrooms, air-conditioning and shared delivery areas), or the actual conversion of delivery rooms into operating theatres, were reported as important barriers. [42,47,55,57] The need to consider the birth environment as comprising of material facilities, but also material relations between humans and systems was evident within and between studies, and across resource settings.

Role of hospital: philosophies, purpose and structures (SoFs 5).

Type of hospital, such as whether the hospital was in the public or private sector of care, a university teaching hospital and/or a regional referral centre, was perceived by stakeholders to influence the acceptability and feasibility of specific interventions to reduce caesarean section rates. [27,36,42–44,46,47,55] This could simply be a consequence of different financing structures, clinical policies, and the working environment. However, it could also be due to the power of the predominant philosophy of pregnancy and childbirth, based on perceptions of the purpose of the particular kind of unit. For example, being a University affiliated hospital was viewed by some stakeholders as a potential barrier to caesarean section rate reduction because of the lack of continuity of care and interpersonal relationships due to task and intervention orientated pressures, [46], or to the organisational need for medical residents to take responsibility for births, in preference to midwives [47]. In contrast, where larger or more academic hospitals were associated with better governance structures, this was perceived to be associated with low caesarean section rates, as in the case of Lebanon, where it was reported that caesarean section were low because rigorous audit systems [are] more common in teaching hospitals. ([42]:p.45)

Apathy to change, interdependency and complexity of system (SoFs 6).

Across settings, the complexity of the healthcare system, were clinical and non-clinical factors inevitably converge was perceived as a barrier to simple, standardised interventions to reduce unnecessary caesarean section. [28,43,46,47,52–55,57–60]^. This was partly due to the powerful impact of non-clinical factors, such as management processes, rules, regulations, and conflicting strategies. [46,47] For example, the interdependency of the British National Health Service’s internal structures and workforce (midwives, obstetricians, junior doctors), and “the hugely complex series of events”, contributing to high rates of intervention in pregnancy and childbirth, meant many participants reported that achieving higher rates of normal birth and lower rates of caesarean section was unlikely to be effectively addressed by the apparently simple solution of a clinical pathway.([53]:p231) In Nicaragua, healthcare providers spoke of high CS rates as a way of compensating for the multi-dimensional weaknesses in their health system (including insufficient human resource, material resource, or coverage). [55] This was evident in other middle- and low-income countries where antenatal care was absent, communication between all levels of the system, and between the system, staff, and women, was deficient, and infrastructural and geographic challenges of reaching skilled labour care existed. [43,47,58,59]

Multi-disciplinary collaboration, role demarcation and respect (SoFs 7).

In 15 studies, the strength of multi-disciplinary teamwork in an organisation or system was reported to be an important barrier to or facilitator of caesarean section rate reduction. [40,42,46,47,49–58,60–62]. The kind of teamwork that mattered was less about working directly on the caesarean section rate, and more about the general ethos and atmosphere of mutual respect. Stakeholders from organisations or systems with high caesarean section rates said working relationships between professionals were poor, with collaboration, communication, and respectful role demarcation between professionals lacking. [56,57,61] As expressed by this Iranian midwife “in many cases of care, we need to ask other colleagues to do the examination, or other things to help but unfortunately, some colleagues do not believe in helping their colleagues” ([57]:p.1277) ^. In contrast, stakeholders working within organisations with low caesarean section rates valued “working together as a team, knowing that everyone’s voice will be heard, and action is taken at every level of the organization.” ([40]:p.45) One explanation as to why respectful teamwork may contribute to lower organisational caesarean section rates was offered by a UK midwifery manager: “everybody has greater awareness; consultants, registrars, SHOs, ultrasonographers, student midwives, student nurses, anaesthetists… they all bring a different perspective and they also take credibility back to their own peer group.” ([56]:p.337)

Whose risks, whose benefits’? Attitudes towards risks, benefits and rates of caesarean section (SoFs 8).

Important differences in stakeholder attitudes towards caesarean section were reported. [36,39,42,46,47,50–56,59,61] Within and between studies, some health professionals described a lack of knowledge about caesarean section rates, indications or outcomes [42,51,55] while other health professionals and women perceived caesarean section as “normal”. [61] Some health professionals acknowledged caesarean section rates were (too) high locally, and that this might increase risks, but perceived them to be less, or no more severe, than the risks associated with vaginal delivery for mother or infant. [42,47,51,55]. In one study some specialists claimed the complications secondary to C-section are ignorable ([47]:p.6), while other health professionals reported concerns about anaesthetic risks, surgical complications, increased recovery time, cost, and longer term consequences for women. [46,47] In a US study, an obstetrician summed up how attitudes towards caesarean section are shaped by cultural context, at the same time as suggesting the potential of human agency; “People are starting to think; are we really doing the right thing? And I think the answer is clearly no … I can’t believe that evolution is pushing us into the operating room. I think we’re pushing ourselves into the operating room… it’s almost like the perfect storm. You’re going to pay me more, I get to worry less, you’re not going to sue me, and I’ll be done in an hour.” ([61]:p.342) Women also had varied views about birth method, some of which were resonant with those of health professionals. One important difference in women’s views was the embodiment of living with the health consequences of caesarean section. For example, in the context of Ghana’s subsistence culture, one woman said “the C-section itself becomes a disease.” ([59]:p.e123)

Beliefs about quality of care mediated by beliefs about caesarean section (SoFs 9).

Related to stakeholders’ attitudes concerning caesarean section, were their varying beliefs about whether care quality is compromised or enhanced by reducing caesareans. [27,39–42,46–55,57–62] In the UK, US and Canada in organisations where care was focused on the promotion of normal birth and reducing, or maintaining, low caesarean section rates, some health professionals viewed this as having a positive impact on women’s birth experiences and quality of care. [40,59,61] However, within these studies [59,61] where a specific facility’s organisational culture endorsed maternal request caesarean section, and across other studies from high- and middle-income countries, health professionals’ inertia to change was based on the belief that women increasingly want caesareans and are inadequately prepared for labour and vaginal birth. [27,42,46–48,51,55,57,59,61]. Twelve studies reported women’s views, [27,39,41,42,46,48,49,58–62] including their choice of caesarean section and lack of antenatal education about labour, vaginal birth and caesareans. Two studies noted that maternity service users’ views about the acceptability of caesarean section may change (positively or negatively) as increasing numbers of women undergo the procedure, and that there is a need to understand how this relates to women’s perceptions of the quality of care. [52,42]

The value of interpersonal relationships during childbirth (SoFs 10).

In 13 studies [36,39,40,41,46,47,49,52–57,61,62] stakeholders reported valuing interpersonal relationships during labour and childbirth (including emotional labour, companionship and advocacy). In twelve high- and one middle-income country, women talked about their positive experiences of labour support from doulas and/or midwives. Health professionals also talked about the importance of partner support and one-to-one midwifery/nursing care in high-income settings where these were available. In middle-income settings the value of labour support was acknowledged, but availability was limited by too few midwives and inadequate facilities for partners to accompany women during labour.

Normative culture of intervention in childbirth (SoFs 11).

Stakeholder’s concerns that there was a normative culture of intervention in childbirth, and that this acted as an important barrier to caesarean section reduction, were voiced across high- and middle income settings. [36,39,42,46,47,49,50,52–57,61,62] These stakeholders were predominantly health professionals who valued medical care when used appropriately, but who also talked about how the over-medicalisation of childbirth may limit both their opportunities to fulfil their role optimally, and the opportunities for women to experience normal pregnancy and childbirth. Some health professionals, women, and managers perceived the advantages of vaginal birth to include increased speed of recovery, improved bonding between mother and child, shorter stays at the facility, lower costs for the health system, and, as stated by a decision-maker professional at a local level in Nicaragua, “it is physiological.” [55:p.2387] In contrast, there was recognition across settings, that “some doctors’ routine prescription is intervention.” [57:1377] That quote, from a participant in Iran, is illustrative of a general culture of intervention. Other stakeholders talked about specific practices, such as shift handover, where it was the norm for some staff engage in the process of “cleaning up”, about which, a paediatrician from the USA said: “I’ll come in and the C-section fairy is on.” [61:p.341]

Widely desired in principle but not universally acceptable in practice: standardising care (SoFs 12).

In 8 studies [40,42,50–55,57,58] health professionals and policy makers reported that shifts to standardise care were widely desired, but not universally acceptable in practice. Many stakeholders said they had high expectations of guidelines, care pathways or screening tools to reduce unnecessary caesarean section. They were particularly confident about such instruments of change if they were evidence based, designed to be used by multi-professional teams, and developed by consensus. However, discrepancies between what policy makers said existed and what health professionals said they were aware of were evident. [55] Participants from organisations with low caesarean section rates recognised that “great effort has been made to ensure that staff are aware of national standards and guidelines.” [40:p.45] Where intervention content imitated existing practices some health professionals welcomed them as legitimising and supportive of their clinical judgement, [50,52–54] while other staff in the same studies, particularly more experienced staff, experienced them as constraining of clinical judgement suggesting they encouraged “robotic care” through a “tick-box-approach.” [53:p.232] The burden of tools to audit and record standardised processes, and the time this took away from direct hands on care was also noted in one cross-country study. [58]

Attitudes towards in-practice use of best-evidence (SoFs 13).

One of the issues that underpinned the theoretical acceptance of standardised care, but the resistance to it in practice, was the notion of which standards are ‘good’ and how far population based evidence should always be used for individuals. In organisations with low caesarean section rates the normative culture was described as “embracing of evidence and the drive to continually improve.” [40:p.45] In organisations where new interventions were introduced with the aim to reduce caesarean section rates, without taking account of local health cultural norms, professionals reported how the underpinning evidence may be seen as credible or not depending on the prior beliefs and values of specific stakeholder groups. This is illustrated by a midwife in the UK who said “It’s written down and because it’s coming from research, you’ve got all the references in front of you as to what type of research has been used and it sort of … just backs you up”, while her obstetric colleague said of the same evidence “We’re swapping one lot of vague-ish evidence for another lot of vague-ish evidence–and wait and see if anything goes wrong or not”. [54:p.728] The selective use of evidence was reported by participants within studies, across resource settings. [42,47,50,54,55]

Leading and following: Effective leadership, stakeholder involvement, and ownership to facilitate more positive attitudes towards changing workloads (SoFs 14 and 15).

In 14 studies from 13 countries, participants reported effective leadership, stakeholder involvement and ownership as crucial facilitators of commitment to reducing unnecessary caesarean section. [28,40,42,47,50–58,60–62] There was talk of the high priority caesarean section reduction should be given in the public domain (including media coverage) to engage women and their wider social networks. It was felt that this should be undertaken simultaneously with interventions across organisations, facilities and systems with respected, identifiable professional leaders at every level (both top-down and within and across peer-groups). The co-ordination of multiple mechanisms of commitment was considered essential to facilitating cultural and system change, because, as summed up by this manager, from the UK, “if you want to implement something new, you need to get lots of stakeholders on board.” [54:p.727] This also illustrates the important point that leaders can only lead effectively if they have followers who are convinced by their vision and the direction they are taking their organisation. Within and between studies, many participants expressed unmet needs for involvement in the development and implementation of interventions. For some professionals, opposition to change appeared to emerge from feelings of exclusion, alienation, limited sense of ownership, or lack of understanding of the underlying rationale for the change. [42,50,52–54,57,61] These factors were also observed in childbearing women, some of whom found it unacceptable that health professionals were making efforts to keep their labour physiological without understanding why. [52–52,61] The degree of opposition encountered was related to the extent to which an intervention was going against the local cultural norms. In such contexts, a lack of effective, sustainable leadership, little overt organisational buy-in, no mandatory requirement to change or no long-term accountability for caesarean section rates were associated with a lack of convinced followership, which was a significant barrier to change. As a midwife in Iran said “One of the problems we have is that by presenting a program, we cannot expect the program to be implemented in the best way.” [57: p1376] In another Iranian study, [47] and in Lebanon [42] and Chile [62], the need for a National Task Force with obstetric and midwifery representation was noted. Hospitals that achieved success in reducing rates identified nursing and medical leaders who endorsed and championed the project, and who made change an institution wide policy priority. [40,50,51,56]

Effective leadership, within and between professional groups, was also an important mediator of doctors and midwives’ openness to change in their everyday work. [42,46,47,49,50–58,60,61] This SoFs (15), is related to SoFs 7 (normative cultures of multi-disciplinary working between professionals) and others (including SoFs 3 and 4). It is distinct in its focus on attitudes towards the reassigning of workloads (shifting professional roles), new work (as a consequence of the intervention) and the importance of pre-existing workload pressures in implementation considerations. Across settings the importance of additional resource allocation was voiced. For example, in the UK, Japan, and Iran, midwives perceived midwifery care models as unmanageable unless more midwives were employed. [46,50,57] In Iran, it was also suggested that increasing the workloads of midwives had had the adverse effect of increasing caesarean section rates, as midwives came under pressure to free-up hospital beds. [47] Where interventions redefined the doctor’s role (family doctors and obstetricians) by shifting lead-professional responsibility to midwives, doctors discontent was evident. In the UK (Wales), doctors expressed concerns that they no longer had an overview of the overall maternity unit workload. Their new role, “placed in a much more technical position”, meaning they were confined to “coming in like the fire brigade.” [53:p.233;54:p732] Other doctors opposition to midwife led care was interpreted by study authors as fear of a shift in medical authority, loss of financial benefits, for both individuals and facilities, and the convenience of scheduled caesarean section, which made workloads more manageable (with less time on the wards, or on-call).

Addressing fears about safely reducing caesarean section rates through education and training (SoFs 16 and 17).

In 14 studies, stakeholder fears concerning the safely of reducing caesarean section rates were reported. [27,36,39,40,42,46,47,49,50,52–55,57,61,62] In the UK (in Scotland and Wales), fears about compromised clinical safety for women were described by doctors, and by some midwives, following a shift to midwifery-led models of care. [50,52–54] In contrast, in Canadian, UK and USA settings with the lowest caesarean section rates, midwives and obstetricians were more confident that support for women to give birth normally was where midwifery’s strength lay, with obstetric colleagues being well-trained to deal with any complications. [36,50,52–54,61] Practices and skill levels identified as facilitators of low caesarean section rates included “well-trained, technically facile obstetricians who feel comfortable allowing a long 2^nd stage, who are competent at delivering breeches vaginally… and who encourage VBAC’s.” [40: p,44] Despite this, while some decision-makers cited several advantages to vaginal birth, many health professionals focused on the risks. Defensive practice was talked about as a barrier to reducing unnecessary caesarean section in seven studies. [27,42,46,47,49,55,61] A lack of confidence in the safety of normal birth on the part of some women was also noted [27,52–54,61], with a Midwife in Iran suggesting one reason for this was that “…society has spent more time on teaching the process of suing rather than introducing the labor to the general public.” [47p:5].

The importance of education and training that prioritises normal birth and continuous quality improvement was reported in eight studies from high- and middle income settings. [40,42,46,47,55,56,57,61] The needs discussed included better prenatal education for women, and training of health professionals in clinical skills, clinical audit and the actual programme content of specific interventions or programmes targeted to reduce unnecessary caesarean section. The need for such training to be available and accessible to all stakeholders is encapsulated in this quote from a nurse in the US: “I would provide the residents with more education on normal … I would want every single nurse on this unit to go through a childbirth education series, not the 1-day class, but a series. I would like to make the series available to every single patient here, at an affordable cost. Every single patient!” [61;p342]

Dealing with complex adaptive systems by understanding, and tailoring to local context (SoFs 18 and 19).

The importance of understanding and effectively responding to local context, culture and pre-existing initiatives was evident in 16 studies as important mediators of negotiating support or resistance to change. [27,40,42,47–58,60–62] At country level distinctions were made between Chile and Lebanon for example. In Lebanon the convenience of caesarean section was suggested to be the foremost consideration with the need to address patience and skills in vaginal delivery in the “new generation” of obstetricians. Within countries there was also evidence of how the same interventions had different effects depending both upon the culture into which they were introduced and how they were accomplished therein. [50,52–54,56,62] Existing practice patterns, including maternal request for caesarean section, staff attitudes, relationships between professional groups and synergy with other initiatives (financial strategies and incentives, other guidelines and concurrent policies, evidence-based practice, local audit priorities) were all discussed. One UK study noted concurrent strategies intended to increase the normal birth rate (i.e. targeting home birth) as potential confounders, nevertheless caesarean section and instrumental delivery rates continued to rise, with the culture of individual units a significant factor. [53] There was recognition of the need for local tailoring of interventions, and for acknowledgment of how local culture must be actively and continuously negotiated as part of a wider system.

The subtleties of change-in-the-making were highlighted in 14 studies that reported how adaptive, multi-faceted interventions that accommodated local adaptation could optimally contribute to successful change programmes. [28,39,40,42,46,47,50–54,56,58–61] Examples of local adaptation included moving elective caesarean sections to a newly opened operating suite, which reduced scheduling conflicts that occurred when sharing space [51], obstetricians learning from midwives in ways they did not learn during their training about how to counsel women in early labour [61], and recognising “obstetricians did not attend the initial meetings related to the initiative”; but when “a separate meeting was arranged to fit with their time commitments”, that “was well attended.” [56: p337] Stakeholders described interventions that were continuously and creatively negotiated on-the-ground in ways that were not easily captured or anticipated. The mechanisms included inspiring confidence, and patience with variation in the length of time required to bring about change in different organisational cultural contexts. Some of the factors that contributed to development and effectiveness of interventions were opportunistic. For instance, they may have capitalised on other developments in other areas of the health system, so they were built alongside a general change in culture, rather than adherence to a particular checklist, or rigid protocol. Successful programmes also tended to have built-in mechanisms for multi-disciplinary collaboration and communication, and a commitment to continuous quality improvement so that adaptations could be made as evidence of local tinkering came to light. Without mechanisms to identify and address such issues, there was some evidence of no effect on caesarean section rates, or they continued to rise, as women previously identified as "normal" were re-classified as potentially "at risk" [52–54] or indications were found to fulfil insurance criteria [47].

In the final interpretive synthesis stage of the analysis (Fig 2) findings were combined to represent our interpretation, through a line of argument.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Summary of findings and summary themes.

https://doi.org/10.1371/journal.pone.0203274.g002

Line of argument synthesis.

Maternity care is a complex adaptive system. Interventions to reduce caesarean section are unlikely to be successful unless account is taken of power, at all levels of the local health system and society, and until cultural norms and relationships are factored into the intervention process. Mechanisms of effect to achieve change include attention to effective leadership and followership; management of resistance to shifting power relations and to fear of responsibility for risk; and fostering of belief in the importance of reducing the caesarean section rate, with corresponding education of women and the training of health professionals. There is evidence to suggest this can be achieved by continuous dynamic assessment of, and tailoring to, local cultural norms and beliefs, as an essential and intrinsic part of the evaluation and implementation process of any new intervention or approach. Specific facilitators include multi-factorial programmes that build belief in, and valuing of, the need to reduce unnecessary caesarean section with all maternity stakeholders involved; authentic buy-in from effective leadership at all levels; three-way communication between women, midwives and doctors that includes listening as well as telling; and turning perceived losses (such as financial penalties, loss of professional roles and power, and perceived vulnerability to litigation) into gains (including pride in caesarean section rate, positive working relationships, better birth environments and improved quality of care for women and families).