Feasibility and design of a trial regarding the optimal mode of delivery for preterm birth: the CASSAVA multiple methods study

Maternal: major maternal morbidity (as a composite and as individual components)

• Post-partum haemorrhage ≥ 1 litre.

• Admission to the main intensive care unit for > 24 hours.

• Bowel injury requiring repair.

• Bladder injury requiring catheterisation for > 48 hours.

• Other organ damage.

• Postnatal anaemia < 70 g/dl (or requiring a transfusion).

• Postnatal infection (sepsis) requiring intravenous antibiotics for > 48 hours.

• Anal sphincter injury.

• Postnatal depression requiring inpatient admission.

• Hysterectomy.

• Return to theatre for any other reason.

• Confirmed thromboembolic disease.

• Post-traumatic stress disorder.

• Hospital stay of ≥ 7 days.

Neonatal

• Incidence of Apgar scores of < 7 at 1 and 5 minutes.

• Early mortality (i.e. < 72 hours of life).

• Mortality before 6 months of age/discharge home alive.

• Significant IVH (defined as grade 3 or 4 IVH).

• Cystic periventricular leukomalacia.

Costs

• Costs of major maternal morbidities from recruitment to 6 months postnatally.

• Costs of major neonatal morbidities from recruitment to 6 months postnatally.

A fuller description and a flow chart of CASSAVAplus is shown in Appendix 3, Figure 17.

Health-care professionals would welcome improved evidence, but anticipate difficulties producing this

Health-care professionals largely welcomed the idea of a trial of MoB preterm (Box 4). Health-care professionals described the current evidence base for decision-making about MoB in PTB as suboptimal, with many highlighting gaps that they would like to see filled:

We really like guidelines and proformas and evidence-based work, and it doesn’t really feel like that is at the moment . . . it would be really good to have some really strong evidence to actually base our practice on, rather than just kind of [the] personal preference of the consultant.

M01

Health-care professionals described how the current deficiencies in evidence led to uncertainty (for both health-care professionals and pregnant women), difficulties facilitating informed decision-making and, potentially, suboptimal clinical outcomes for some women and babies. Therefore, interviewees were largely supportive of the idea of a trial:

We need really good evidence to drive change, so that a substantial number of women and babies can get better outcomes.

OB6

However, although health-care professionals were supportive, in principle, of a trial, interviewees envisaged significant challenges to delivering one as proposed in the HTP:

It’s a really good topic . . . it would be great to have the evidence . . . However, I think gathering the evidence is going to be difficult.

RM03

These challenges related mostly to the existence of strong and divergent views in the health-care professional community about the relative merits of vaginal birth and CS in PTB, and associated variations in practice (discussed in detail below). In addition, health-care professionals shared doubts about aspects of trial design, which they suggested might undermine wider support for the proposed trial and raised other, more pragmatic, concerns. These issues are explored in subsequent sections.

Health-care professionals report strong mode of birth preferences (i.e. areas where personal equipoise is lacking)

Despite, or because of, the gaps in the evidence base, interviewees said that they, and/or their colleagues, had developed quite strong and ‘fairly fixed’ (OB1) views about the best or safest MoB, and that views these guided their practice (Box 5). Interviewees also noted how these opinions and associated practices varied between, and sometimes within, centres:

Practice does . . . seem to be different in different centres. And quite strong, different attitudes in different centres, with not a very good evidence base to support it either way.

N02

Women could, theoretically, go down one route or the other if they just rocked up to a different hospital, and had a different person on shift.

OB6

Factors influencing the development of mode of birth preferences

Reflecting on how these different MoB preferences had emerged, interviewees identified several overlapping influences. These included unit/service norms, personal skills and confidence, direct experience, and memorable local and/or high-profile national cases. The first of these influences (i.e. unit/service norms) was portrayed as playing a powerful role in shaping – sometimes by constraining – obstetricians’ (and others’) practices and preferences. Although some interviewees described how practices and preferences had developed through reflection and debate, others highlighted tradition and hierarchy as being pivotal influences:

It will be very much the culture of the hospital, how the deliveries are done . . . the way they’ve always done it.

OB9

Several obstetricians noted how senior consultants’ preferences regarding MoB effectively determined the training experiences of junior staff:

Trainees get trained up with what the consultants are comfortable with.

OB12

For instance, it was noted that where there was a strong tradition of delivering breech presentations by CS, encouragement to specialist trainees to consider vaginal birth, and opportunities to develop the necessary skills, could be very limited.

(Lack of) skills and confidence were further highlighted as important factors, and were implicated, in particular, in some individuals’ reluctance to undertake breech vaginal deliveries:

It’s meant to be evidence-based medicine, but you do . . . what you’re most comfortable with.

OB12

Doctors have become de-skilled in breech vaginal deliveries . . . and they don’t have the confidence or skills.

RM01

Personal experience was also said to play a powerful role in shaping many care decisions. Interviewees noted, for example, how difficult and/or distressing experiences could have a lasting impact on health-care professionals, and leave them keen to avoid repeats (prompting health-care professionals to make choices that supported that goal). Memorable (dramatic and usually traumatic) local cases and high-profile cases from other units/services (e.g. those receiving extensive media coverage because of neonatal deaths and/or disciplinary action against health-care professionals) appeared to exert a similar hold over less-experienced clinicians, as the following quotation from a specialist trainee illustrates:

I think that also, because of the Ninewells case50 – you know, with the consultant and the breech delivery, that was probably about 2 years ago – I think people are also a little bit more mindful, and sort of keen to go for a caesarean section, because . . . people would feel that there’s less likely a chance of sort of neonatal morbidity and mortality associated with a section – in the immediate aftermath at any rate . . . It [the Ninewells case] was a 24-week breech . . . a vaginal breech delivery was attempted . . . and essentially, unfortunately, baby died – and there was a decapitation.

OB4

Patterning of mode of birth preferences by clinical scenario

Health-care professionals explained that their decisions and practices were dynamic, contingent and responsive to a wide range of maternal, fetal and service/clinician factors. However, health-care professionals suggested that presentation (cephalic or breech) and gestational age were particularly prominent (and intersecting) considerations. Although interviewees articulated a variety of views, some broad patterns emerged, which are detailed below.

Health-care professionals suggested that cephalic presentations were generally considered lower risk than non-cephalic presentations. Therefore, a preference for attempting vaginal birth (and avoiding CS), although not unanimous, was relatively strong and widespread. In support of vaginal birth, interviewees cited avoidance of short- and long-term maternal morbidities potentially arising from CS; preservation of women’s reproductive capacity and future birth options; potential benefits to the foetus/baby, including enhanced respiratory and immune functions; and service considerations, such as meeting targets to reduce rates of CS.

For some interviewees, however, support for vaginal birth was contingent on gestational age. These health-care professionals reported that they and/or colleagues would be keenest to undertake vaginal birth – and feel most discomfort about undertaking CS – at later gestations:

30 weeks onwards . . . babies stand a better chance, they’ll have better outcomes generally . . . And if they’re a cephalic presentation, and normally grown, and a normal infant, then a normal vaginal delivery would be preferred.

RM02

The . . . later preterm ones, you know . . . if a baby was head down and it was 34 weeks, I would be quite reluctant to do a caesarean section . . . Personally . . . I’m not in equipoise at that . . . gestation.

OB8

For cephalic presentations at earlier gestations (i.e. less than 28–32 weeks), preferences for vaginal birth appeared less stable/strong, with several health-care professionals (including some obstetricians who reported their usual/default practice as vaginal birth) conceding to some uncertainty regarding what was actually the best/safest option:

The group that I agonise about are the sub-28 weekers, because those babies . . . are so teeny-tiny and they’re so fragile, that sometimes you think . . . ‘Does vaginal birth compromise them too much?’

OB8

Others, although themselves favouring vaginal birth, suggested that colleagues might take a different view, and explicitly advocate birth by CS at earlier gestations:

There’s a hardcore who already think that every preterm baby before ‘X’ weeks, 32 weeks, should be sectioned.

OB5

There was greatest consistency in perspective with regard to the most preterm cephalic foetuses (i.e. 22–25 weeks’ gestation). Here, vaginal birth was almost unanimously favoured, with many interviewees expressing profound unease about conducting CS in this scenario. Obstetricians, midwives and neonatologists all communicated concerns about the challenges of, and significant maternal morbidity associated with, the ‘classical’ section typically required:

Perform[ing] a caesarean at 22 weeks . . . is very difficult . . . [and] for future pregnancies it would be very different than a caesarean section performed later . . . I’m not sure someone would do a caesarean section at 22 weeks . . . [Possibly] from 24 [weeks], maybe 23-plus, but . . . [it’s] very borderline, because, you know, you could have a dead baby, so you just get an injury to the mum.

OB11

22-plus weeks, gosh, that’s really, well, less than the age of viability, isn’t it? . . . Why would you section a woman before 24 weeks? Why would you put a scar on her uterus, for a baby that was less than the age of viability?

RM02

Indeed, many obstetricians said, categorically, that they would not be prepared to conduct a CS at what they referred to as the ‘threshold of viability’.

With breech presentations, a preference for birth by CS was more common. Again, strength of feeling appeared to be affected by gestational age, with preferences for CS being most pronounced at later gestations (i.e. ≥ 28–32 weeks’ gestation). Accounting for this, health-care professionals highlighted a variety of concerns about the potential complications of a breech vaginal birth:

With the breeches, where they’re preterm . . . the lower half of the body can slip through a moderately dilated cervix, but . . . the cervix wouldn’t be large enough for the head to pass through . . . and then it’s an extremely stressful situation – when you’re at the point of no return and, you know, half of the baby is out, but the head gets stuck in the cervix, and then attempts have [to] be made to cut the cervix, with the baby’s head sort of millimetres away.

OB4

Owing to these risks, some interviewees said that they would never, or hardly ever, be happy to deliver a preterm breech baby vaginally:

A breech itself is risk, and a premature breech then is risk [added] to risk. So I would never be happy doing a vaginal breech on less than 37 weeks.

OB12

However, although recognising the challenges, around half the interviewed obstetricians said that they, and/or their colleagues, would at least consider vaginal birth in breech presentations when gestational age was < 28 weeks:

Many of my colleagues . . . are OK if it’s very early, [although] if they get to kind of 28 weeks and upwards, or a bit higher than that upwards, they would be less happy about a woman having a vaginal breech delivery.

OB1

As with cephalic presentations, obstetrician interviewees reported some uncertainty about the best way to manage breech presentations at these early (but not extreme) gestations.

Similarly (to cephalic presentations), at the very earliest gestations included in the trial, some interviewees suggested that they and/or colleagues might be reluctant to perform a CS:

With a breech delivery that was over . . . about 24 weeks, I would say they would do a caesarean section for that . . . [But] the little tiny ones, I don’t think they would . . . They obviously are thinking that they’re probably not going to survive anyway – so if they have a difficult breech delivery, it’s not going to make any difference.

RM03

Interviewees expressed a comparable reluctance to induce labour at these extremely early gestations. More generally, interviewees viewed elective birth scenarios as problematic in the context of the proposed trial. They explained that, in effect, this would involve comparing CS with induction. They considered this fundamentally different from a comparison of CS with spontaneous vaginal birth:

There is an important difference in babies whose birth you’re planning, versus babies whose birth isn’t planned . . . It feels to me that that’s a bit of a different question, because you’re factoring in [other] things . . . There are too many other variables, it’s not just, ‘Should this baby be born vaginally?’

OB8

Obstetricians also stressed that the decision to deliver electively, preterm, was never taken lightly, as the costs of premature birth were profound. Obstetricians explained that typically this would be contemplated only when a serious threat to maternal or fetal health had been identified. Several obstetricians emphasised that such situations tended to require urgent intervention and, as such, would lead them to favour birth by CS. These obstetricians indicated greatest willingness to consider induction at gestations close(st) to term.

Diverging from preferred mode of birth is more problematic in some scenarios than others

Health-care professionals recognised that, for the proposed trial to be successful, they and their colleagues would need to be willing to diverge from preferred MoB practices (Box 6). Although some health-care professionals suggested that, for them personally, commitment to evidence-based practice might act as something of a counterweight to imperfect personal equipoise, others expressed the view that getting people to move away from what they considered best/safest care would prove extremely challenging:

You’re going to be trying to get God knows how many DGHs [District General Hospitals] . . . to change what they see as safe care for the purposes of a research study. That’s not impossible, but . . .

OB6

Trial eligible populations viewed as particularly problematic

Health-care professionals suggested that the trial eligibility criteria included some scenarios/populations in which convincing them and/or their colleagues to change their practice was likely to be more difficult than in others. They identified several populations/scenarios in which, they suggested, the risk of health-care professionals engaging in behaviours with clear potential to undermine the trial (discussed further below) was particularly high. These corresponded with those populations/scenarios in which MoB preferences were most marked (i.e. the extremes of gestation, breech births and elective birth), but also included some additional scenarios/groups (e.g. foetal compromise and women perceived as ‘vulnerable’).

For example, interviewees suggested that they and/or their colleagues would find randomising in situations when women (of any presentation) were very close to, or furthest from, term problematic:

I will randomise to any trial that people can convince me about equipoise. It’s just that . . . some of your inclusion/exclusion criteria are a little wide for me and I might not go as far . . . Would I randomise women at 36 weeks⁺⁶ days with a cephalic to caesarean section or vaginal delivery? In my heart, 100% not.

OB5

My suspicion is that some of the obstetricians, at the very lowest gestations, just won’t be willing to do a caesarean section and so might be unwilling to contemplate randomisation . . . I think some of them might have the view that the outcome is so poor for those infants, and the risks so large, that . . . it’s inappropriate to randomise.

N01

For some obstetricians, breech generally was considered problematic, because of the obstetricians’ strong preference for birth by CS:

The main sticking point for me is the breech thing . . . I wouldn’t really randomise [a] woman if knew she had . . . a premature breech . . . And I think a lot of other clinicians might feel the same.

OB12

Other scenarios in which some obstetricians questioned their own willingness to recruit and randomise included foetal compromise (e.g. growth-restricted babies):

It would depend if you could show me data that there was a potential benefit for having a vaginal delivery in these SGA [small for gestation age] IUGR [intrauterine growth restriction] babies . . . some data to show . . . there’s a definite benefit to having those contractions, to squeeze the fluid out of the lungs, because it means they’re less likely to have transient tachypnoea.

OB9

It was also suggested that clinicians might also be reluctant to recruit and randomise women they viewed as emotionally fragile or vulnerable. This might include:

Women who’ve had bad obstetric history, or are really anxious . . . and don’t want any sort of uncertainty. And prims as well . . . [Women having] first-time babies . . . are likely to be a lot more anxious . . . [Although] to be fair, everybody ought to have the opportunity to take part . . . so maybe we shouldn’t make assumptions . . . It’s difficult.’

RM04

Trial eligible populations viewed as less problematic

Conversely, interviewees suggested that where MoB preferences were less marked, they and/or their colleagues would find recruiting and randomising less problematic. These interviewees suggested that this might be the case between the extremes of prematurity (at roughly 26–32 weeks’ gestation). For example, one neonatologist remarked:

The 25, 26, 27-weekers . . . there is a very real question about how they should be delivered, whether it’s caesarean section or aiming for vaginal delivery. And . . . I think the obstetricians would find that an easier population to randomise.

N01

However, for some, this window of possibility was very tightly circumscribed in terms of gestational range and, for a few, presentation remained a prominent concern (with these interviewees suggesting that they would still only feel comfortable randomising cephalic presentations). As varying positions were expressed within interviews, pinning down the boundaries of personal equipoise often proved difficult:

For me, the biggest question is women who come in in spontaneous labour, and particularly the very preterm but viable infants. So more than 24 weeks – less than 26 . . . 28 weeks . . . 32 weeks would be definite cut-off.

OB08

The presence of such contradictions, corrections and/or qualifications in accounts suggest that, in many instances, interviewees were (re)formulating and (re)considering the certainty of their convictions as they talked. Explaining their hesitancies, interviewees highlighted the range of other factors that might, independently or in conjunction with gestational age and presentation, shift their preferences and practices around MoB. These included, inter alia, multiple pregnancies, degrees of fetal and maternal compromise, and other maternal characteristics, such as age and obstetric history. Interviewees stressed that these considerations were numerous, intersecting and could present in an almost infinite number of combinations. This, they said, made it ‘very difficult to know at exactly what point are you going to say, “I’m in equipoise”’ (OB5).

Marked mode of birth preferences are expected to lead to behaviours detrimental to the trial

Interviewees anticipated that the strong MoB preferences described above would prompt actions and behaviours at both unit/service and individual levels that would present tangible challenges to the proposed trial (Box 7).

Declining to host the trial

Interviewees emphasised that decisions to host a trial are made by teams rather than individuals and that their department would only sign up if there was unanimous support:

As a department [you] will have agreed to take part only if . . . you feel you can [all] support that conversation with women . . . We discuss our individual points of equipoise, and also where we feel there’s maybe not equipoise. And we’d only take it on if, as a group . . . we were all prepared to sign up for it.

N02

Echoing this suggestion, and based on conversations with obstetrician colleagues about the proposed trial, some research midwives (based in less research-active centres) expressed deep reservations that it would be possible to open the proposed trial at their centre:

We wouldn’t be able to agree to open this trial here because there would be no point . . . Equipoise is compromised . . . they’re saying that they’re not happy to do caesarean sections on someone that is not clinically indicated on.

RM03

Selectively recruiting patients

Health-care professionals often observed that, although they and/or their colleagues might make a general commitment to the trial, they/their colleagues might be reluctant (or even refuse) to recruit in those scenarios or populations in which practice preferences were strongest:

Something that you would have to allow for in the . . . protocol, is that, you know, some units may say, ‘Well, absolutely we’re not doing that at that gestation, but we’re prepared to do it at a lower gestation’, and vice versa. You may get some units that absolutely categorically refuse to section a 23-week woman, and others that will.

OB10

Health-care professionals suggested that this reluctance might also lead to more insidious and discreet selecting of patients clearly meeting trial eligibility criteria:

I think there’d be a lot of things like, ‘Oh well, maybe we shouldn’t include this person . . . I don’t think we should do a caesarean on them, so I don’t think we should randomise them’.

RM02

Counselling discouraging enrolment

Several obstetricians suggested that reservations about one MoB (either vaginal birth or CS) might also lead health-care professionals to counsel patients in less than impartial ways. These obstetricians warned that subtle modifications to recruitment conversations could significantly affect their outcome:

I would say . . . ‘We’ve got this trial . . . to see whether it’s safer to do caesarean or vaginal delivery . . . if you’re not on the trial, our general approach is to go for a normal [vaginal] delivery’. Now, If I don’t add that last bit, I . . . end up in an argument, ‘Well, you’re not going into the trial, but actually we’re not offering you [a] caesarean . . .’ So how you counsel people is never unbiased. Some people talk about non-directed counselling – it doesn’t exist!

OB5

Deviating from the protocol post randomisation

Finally, obstetricians warned of how discomfort with the practice(s) required by the protocol might result in deviations, in particular to health-care professionals defaulting to their preferred MoB post randomisation:

I think there will be protocol deviations. So, I think that if someone is randomised for section, they’ll get a section, no probs. I think if people are randomised to vaginal delivery, on that arm you’ll see more protocol deviations . . . documented as ‘Clinical reasons: decided caesarean section more safe’. Just because it’s so much easier.

OB4

Modifying the trial protocol

Interviewees suggested that the chances of successfully delivering the trial would be increased by modifying the protocol, critically, by excluding those scenarios/populations in which MoB preferences were particularly marked (Box 8). Even those who personally viewed the proposed inclusion/exclusion criteria as ‘perfectly appropriate’ recognised that many colleagues would find them problematic. These individuals suggested that, as currently conceived, the trial would ‘bump into what is perceived reality on the shop floor’ (OB6). Different interviewees proposed excluding different scenarios/groups, including the extremes of prematurity, foetal compromise/anomaly, multiple pregnancies and elective deliveries. Excluding the first of these groups [i.e. babies furthest from and/or closest to term (both cephalic and breech)] received the most consistent attention:

I would bring up the lower limit gestation . . . bring the lower limit to 24 weeks . . . Including those 22, 23-week gestation babies . . . involves an area of pretty experimental medicine really. And then . . . presentation in the bigger babies . . . maybe bring your upper limit down to 34, rather than going all the way up to 36⁺⁶ – because I . . . accept that you’re going to put preterm breech babies in there, because that’s what you want to look at really, as much as the cephalic ones . . . but you’ve also got to balance that against the risk of the baby getting stuck.

N03

Ensuring that health-care professionals have confidence in the trial design

Interviewees emphasised that the trial needed to be highly robust in its design and scaled to ensure that it could deliver definitive answers. Reflecting on previous trials’ failures to deliver the clear-cut findings needed to bring about changes in practice, some interviewees suggested that pivotal design features, including the proposed (sub)sample sizes and outcome measures, warranted further thought.

(Sub)sample sizes

Although some interviewees welcomed the breadth/diversity of the study population, many questioned if – and what – subgroup analyses should be planned, and if sufficiently large subsamples would be achieved to support those analyses. These interviewees suggested that (perceived) weaknesses of this sort might undermine both health-care professionals’ support for (delivery of) the trial and, ultimately, their willingness to accept and act on its findings:

They’re going to have a real mixed bag of women, and different indications . . . it’s so diverse that as a practicing obstetrician, [whether] you would have enough confidence in its . . . power for these different groups . . . . I’m not sure.

OB01

I’m worried that you’re going to recruit such a . . . heterogeneous group . . . 20 of these, 40 of those, a 100 of those. And you won’t be able to do any subgroup analysis that’s of any meaning.

OB05

Outcome measures

Interviewees emphasised that the value of the trial hinged on collecting the right data on fetal/neonatal and maternal outcomes. The interviewees stressed that, for clear conclusions about the relative effects of the two MoB to be drawn, outcome measures needed to be capable of identifying relevant differences. Reflecting on the primary neonatal outcome (i.e. composite of alive at 6 months after birth or to home discharge, without significant IVH or cystic periventricular leukomalacia), some interviewees questioned if the frequency of these events was sufficient, especially at later gestations, to differentiate the outcomes of the two MoBs:

[In] that cohort of more-than-32-week infants, the chance of significant IVH is very small. [So] if the aim of caesarean delivery versus vaginal delivery is to avoid that, then it seems that the number needed . . . would be very high . . . that sort of outcome would be very difficult to study in a randomised trial.

N01

In addition, some neonatologists questioned if, and how, these measures were related to MoB:

I guess if . . . it [CS] was going to reduce the incidents of IVH or cystic periventricular leukomalacia, that would be . . . a very, very good outcome . . . [but] I’m struggling to think of a mechanism by which that would happen. Because in my experience, IVH, whilst you can get them fairly early on, often it’s not until after they’re born . . . And cystic PVL [periventricular leukomalacia] . . . [also] tends to be something that comes up later, and usually there’s been something to explain it . . . some kind of post-delivery insult . . . I’ve never really had a feeling that . . . they’re related to the way they were born.

N04

Many interviewees emphasised that what was important to parents, perhaps even more so than clinicians, was how babies/children developed over subsequent months and years. These interviewees noted that measures such as IVH and cystic periventricular leukomalacia were essentially proxies for such outcomes:

As neonatologists we are often thinking that what matters most is the outcome of the child in later childhood, and conscious that when you look at surrogate markers like IVH, although it’s clearly important, that occasionally the later outcome doesn’t agree . . . If this were designed as a neonatal trial, rather than an obstetric trial, I suspect the primary outcome would be neurodevelopment, rather than survival or IVH.

N01

These interviewees said that, ideally, these longer-term developmental outcomes should be measured directly. A few (obstetricians) suggested that capturing longer-term maternal outcomes was important, too:

If your primary outcome measure is fetal handicap [and] survival . . . that’s fine . . . [but] the problem is the longer-term outcome. So like the problems of increasing caesarean section rates . . . 15–20 years [on] people get menstrual disturbance . . . there’s a much higher rate of placenta praevia and accreta, but that’s a consequence in 3, 5, 10 years’ time – studies don’t look at that.

OB05

However, this lack of longer-term outcome measures was also acknowledged as being a weakness common to many (obstetrics) trials, and reflective of the restricted nature of much research funding.

Raising awareness about the (need for) evidence

Although our interviewees portrayed themselves as cognisant of the need for evidence, some questioned whether or not health-care professionals more widely were aware of the limitations of the evidence, with one midwife, for example, describing her colleagues as ‘very shocked’ to hear that there was minimal evidence to support current practice(s). Therefore, some health-care professionals surmised that a starting point in securing health-care professionals’ support for the trial would have to be raising awareness of the current state of the evidence base. Obstetricians emphasised that those clinical scenarios in which MoB preferences were especially marked (e.g. at the extremes of gestation) would need particular, focused attention. Theses obstetricians suggested that, in so doing, it would be critical to have credible advocates/champions making the case for the trial:

. . . a prominent face . . . a big name.

OB4

Others emphasised the importance of identifying and getting the support of local decision-makers and opinion leaders:

People will follow the lead . . . follow the local academic lead . . . So long as it’s not perceived as blatantly unsafe, then they will follow that lead and randomise.

OB10

Concerns about the sufficiency of clinical and research resources

Interviewees drew attention to workload issues that might impair their own, or their colleagues’ at other (smaller) centres, capacity to contribute to the trial (Box 9):

It’s time, isn’t it? It’s like ‘busy-busy’!

RM03

Interviewees emphasised that, as well as having an impact on the workload of obstetricians, midwives and labour ward staff, MoB decisions had implications for anaesthetists’ time, theatre usage and labour ward bed occupancy. Interviewees noted how, with more work to do than time to do it in, health-care professionals might, quite reasonably, prioritise other tasks:

A challenge with any clinical trial, [is] getting people to take part in research when they’re already very busy with their clinical jobs . . . although everybody’s aware of research, you know, it can sometimes just be seen as a bit of a nuisance on top of your . . . busy NHS job.

OB3

Many (obstetricians) expressed the view that recruitment would need to be ‘clinician led’, as discussions about MoB require clinician expertise. However, these interviewees also acknowledged that they would need the help of research support staff, including Good Clinical Practice (GCP)-trained research midwives, in identifying, screening and randomising eligible patients:

Even if the clinician took consent, they’d often need a little bit of support with the randomisation process, and that can be not just inexperience of doing it, but also just workload. And, therefore, when you’ve got . . . research support staff . . . you get better recruitment.

OB7

Interviewees noted how patients could present 24/7, with some commenting that research midwives worked largely office hours. Therefore, even at those sites with more developed research infrastructure and support, interviewees envisaged challenges recruiting and randomising out of hours. It was noted that a large group of (clinical) staff would need to be willing and able to provide support to the study if eligible patients were not to be missed. As interviewees further noted, these people would need GCP certification and also, some suggested, some trial-specific training, as well as a proportionate amount of ring-fenced time:

Centres that become involved in a study like this have to really commit to it, and have a number of individuals who are very dedicated to it . . . And you probably need to have a certain amount of dedicated resource. I don’t think it’s enough . . . to simply expect people to take it on as part of their day-to-day workload.

N05

Concerns about the adequacy of (some) health-care professionals’ clinical skills

Interviewees questioned how many units/services could guarantee the availability of staff with the necessary skills to deliver both arms of the proposed trial effectively. Obstetricians emphasised the criticality of this issue in the more challenging birth scenarios encompassed by the trial (e.g. breech vaginal birth or very early CS):

These extreme preterm gestations, you can’t just leave it to the registrar, the junior registrar, to do these deliveries. These are deliveries that need input from experienced people.

OB9

Several interviewees, although emphasising that their own (large, tertiary) service had 24-hour consultant cover [‘There’s always two obstetricians on-call’ (OB11)], warned that this was not the case at other centres. Interviewees surmised that the availability of staff with appropriate clinical skills and confidence would have implications for both recruitment and study outcomes. Therefore, some interviewees suggested that the trial team should focus on tertiary units where there was greater experience of both difficult vaginal births and CSs and 24-hour consultant cover:

My unit has 24 hours – they can sort of cover – so it makes no difference really what time of the day or night it happens . . . that wouldn’t affect it. But I guess if you’re trying to randomise in other units it would, and that’s potentially a problem . . . I think you probably wouldn’t want to do this . . . in units where there isn’t experienced obstetric cover available for breech vaginal delivery – because that’s necessarily going to bias your results.

OB10

Although some interviewees surmised that in the bigger clinical units appropriate skills and confidence could be assumed, others suggested that for some clinical scenarios, in particular breech vaginal birth, even ‘experienced’ clinicians might vary in their skills and confidence. Health-care professionals drew attention to a general shift away from vaginal breech birth following the completion of the Term Breech trial. 13 Some interviewees remarked that, as a consequence of that trial, there had been limited/diminishing opportunities for junior doctors to acquire and embed the necessary skills and confidence to deliver babies this way. Although many interviewees saw this as a major stumbling block for the trial, some suggested that investment in skill development and support might increase the pool of health-care professionals (i.e. obstetricians) with the skills and confidence necessary to undertake these more complex vaginal births successfully. Proposals included offering ‘a refresher course about vaginal breech’ (OB4) and ensuring that less-experienced obstetricians had ‘some senior support’ (OB12) when undertaking this MoB.

Some interviewees additionally emphasised that birth by CS could also be difficult. These interviewees noted, in particular, that very early CS, especially ‘classical caesareans’ involving a long vertical incision, presented significant technical challenges, again, warranting the involvement of a senior, experienced obstetrician. Some interviewees opined that the profession lacked experience and the necessary technical expertise to successfully undertake CS at the earliest gestations covered by the trial (i.e. 22–24 weeks’ gestation):

I’d expect a caesarean section to be more difficult [at 22–24 weeks], so there is a technical issue here. We’re not really expert in doing a caesarean at these gestational ages.

OB11

Concerns about the adequacy of (some) health-care professionals’ research skills

Alongside identifying gaps in clinical skills, interviewees questioned whether or not they and/or colleagues had the necessary research skills. In particular, concerns related to the challenges of securing informed consent intrapartum. Health-care professionals expected these conversations to be complex and challenging, especially when PTB was entirely unexpected:

Women in high-risk groups . . . they’re an easy-ish group to speak to because they have prior knowledge and prior experience . . . and they’re out with the kind of labour scenario. Now the challenge is . . . a lot of the women that come in in preterm labour are just coming in in preterm labour . . . Speaking to somebody who’s in labour is harder.

OB1

Many interviewees expressed personal discomfort about broaching trial participation with women who were in active preterm labour. Others interviewees said that they expected colleagues to have concerns and some anticipated significant reluctance to be involved:

No-one will want to do the study, with the exception of weirdos like me! Every single R&I [Research & Innovation] midwife and nurse is going to try to not consent women for us . . . because it’s an uncomfortable conversation.

OB6

Intrapartum recruitment (to any trial) was perceived as presenting distinct practical and ethical challenges. Several health-care professionals expressed uncertainty about women’s capacity for decision-making in the light of pain, analgesia and – in the context of PTB – (di)stress. Some health-care professionals questioned if, under these conditions, informed consent was achievable:

A caesarean in one pregnancy . . . restricts your options for the next pregnancy. But that’s a difficult decision for women to make when they’re not in labour and they’re not worried about having a preterm baby . . . Asking them to do that when they’ve got all that going on . . . I would not be entirely happy . . .. I think it’ll be very difficult for a woman and her partner to give truly informed consent under those circumstances.

OB1

The need to ensure that women were aware of the potentially serious and enduring consequences of the two MoBs was viewed as compounding the difficulty of the conversations that health-care professionals needed to have:

That would be a really difficult thing to consent women for . . . it would be a massive conversation.

M01

Health-care professionals’ comfort with the idea of seeking informed consent intrapartum appeared to be related to the extent of their prior experience of intra- and peripartum trials. Those health-care professionals with more experience reported feeling more comfortable with the idea. A few interviewees suggested that training and tools might help less experienced health-care professionals feel more at ease undertaking this work. Some health-care professionals went further, arguing that training around consent would be essential:

[Clinicians] need a lot of prep[aration] about the consent process . . . a lot of support through that discussion, and a lot of case scenarios, and a lot of working out how you can take informed consent in different situations in an intrapartum setting.

OB7

Refining the consent process

Health-care professionals also proposed some refinements to the trial consent process. Many health-care professionals were in favour of introducing the trial to women earlier in their pregnancy (i.e. before they went into labour) and some opined that this could provide a basis for – and, indeed, should be a condition of – the securing of verbal consent intrapartum:

Having an option for verbal consent is a good idea . . . having the written consent afterwards . . . [But] I think you’d have to have quite a good way of informing women about the trial [ahead of time] . . . Because with verbal consent, I don’t think the . . . time they can give verbal consent is the first time they’ve ever heard about the trial, within the consent guidelines that we’ve worked within.

OB07

Several health-care professionals suggested making all women aware of the trial by mentioning it and/or distributing leaflets at key appointments (e.g. at the 20-week scan). Some health-care professionals noted that an alternative, more targeted, approach would be to make women identified as at risk of PTB, who were often seen in specialist clinics, aware of the trial at their appointments. However, interviewees noted that publicising the study – even in the latter, more targeted way – would result in many women who ultimately gave birth at term receiving information. This, they said, could have costs for both health-care professionals and patients, creating work for the former and, potentially, unnecessary anxiety for the latter:

If you introduce the possibility of preterm labour early on in pregnancy, you’re . . . potentially worrying, or raising concerns in, a lot of women who will not go on to deliver preterm.

N05

‘A big decision’ with risk of allocation to unnecessary surgery

Focus group participants highlighted the significant and very different short- and long-term implications of the two MoBs investigated in the trial. In particular, participants raised concerns about the potential for trial participation/randomisation to result in a woman having unnecessary and invasive surgery. Participants also expressed scepticism about women’s willingness to accept randomisation to CS, in the absence of a decisive medical indication for such a procedure (Box 10):

I would be worried that they would assign me to a C-section, although there might not be health indications towards this. I would prefer to give birth naturally.

FG 2/P3

Why would you opt for major surgery, unless you really had to have it?

FG 1/P2

These women highlighted the surgical risks of CS (e.g. bleeding and infection), lengthy recovery times, impact on one’s ability to provide care for the new baby (and any other children) and implications for future pregnancies/births. These women noted how these possibilities might lead them to favour a vaginal birth:

I might have an inclination to want to have a vaginal birth, just because if you have a C-section, you’re more likely to have to have a C-section the second time round.

PG/P2

Having had a C-section, if I could have had a normal delivery, normal vaginal delivery . . . I think I would have taken that over the C-section any day, because I wouldn’t have had a major operation, I wouldn’t have had 6-plus weeks of recovery, you know?

FG 1/P3

However, as the second of the two quotations above suggests, views on the desirability or otherwise of the two MoBs were shaped by women’s prior experiences. Indeed, those who had experienced complications during, or because of, vaginal birth emphasised that this MoB too could have serious and burdensome implications:

I had a third-degree tear and other issues, so literally as soon as I had him, I was off in surgery anyway . . . and that was with a vaginal birth.

FG 1/P1

I’ve had both a natural birth and a C-section and I’m not really sure what is worse – as I had forceps delivery, which was not fun at all, I had a cut, and they stitched it badly.

FG 2/P4

Women without prior experience of childbirth appeared not to have (such marked) MoB preferences. Nevertheless, they still viewed the decision as a significant one.

A bad time: discomfort with intrapartum consent

Women in all groups found the idea of being approached and consented in the intrapartum period deeply problematic:

If she’s in labour, obviously her waters have broken, she’s panicked and she’s being told to make a decision. . . It’s quite a big decision . . . to be introduced to it at that point!

FG 1/P4

It’s . . . about timing, and being able to think straight . . . you have to be in the right frame of mind to make important decisions.

PG/P2

Those who had experienced a spontaneous PTB highlighted how ill-prepared they had been for the premature onset of labour. Moreover, FG participants who had previously given birth emphasised how labour more generally affected women’s ability to concentrate, highlighting the effects of shock, pain, medication and a narrowed focus on ensuring optimal foetal outcomes:

During labour a person stops thinking and focuses only on this eventually ending, on the child being healthy, on themselves being well . . . This is what I think every mother thinks during labour.

FG 2/P3

Having ‘so many more important things going through our minds’ (PG/P2) might, as FG participants suggested, make women unreceptive to approaches about trial participation:

It is just the last thing on your mind, it’s absolutely the last thing.

FG 1/P3

Here I am, about to give birth, I’m in pain all over, I’m worried . . . and someone approaches me and asks about some research study. First thought? ‘Damn, just piss off! I don’t give a damn about your research just now, I just want to make sure that all is well with my child’.

FG 2/P4

Other women noted the potential challenges of securing genuinely informed consent. To illustrate this point, a FG participant who had previously had an emergency CS questioned if the consent she had given could really be considered informed:

I was asked to sign loads of consent forms and I don’t even have a clue what half of them were, because someone was just telling me, ‘You need to sign this in order to get the babies out, to give them a chance to live’, and I’m like, ‘Fine – I’ll sign whatever it takes!’

PG/P2

In the light of these concerns, women in all groups favoured introducing the trial earlier in pregnancy (somewhere between 9 and 20 weeks):

If someone had mentioned something beforehand, so you were aware . . . [of] this possibility.

FG 1/P1

I think planting a seed earlier on, giving some literature or a leaflet to say, ‘Look, this is a study that’s happening. It’s not to say that you’re going to have a preterm birth, but it’s an interesting read. Have a think about it’.

PG/P1

Reflecting further, some women suggested inviting women to indicate interest in trial participation in their birth plans:

She could include this in her birth planner, whether she would like to agree to such a thing . . . It’s obvious that things do not go according to plan every time, but . . . midwives always try to consider the birth plan the woman has with her.

FG 2/P2

Although some participants drew attention to the potential of early information to induce unnecessary anxiety in women who would ultimately deliver at term, others felt that, if information was composed and provided in careful and sensitive ways, this could be avoided. Moreover, some participants were strongly in favour of raising awareness of, and preparing women for, the possibility of PTB, and surmised that trial information might help achieve this:

That would be good, to raise awareness among those who could potentially – just generally, among all mothers, as you don’t know who . . . might have a preterm birth – that there is such an option, and that something like this might happen . . . [Then] they are able to prepare themselves for this emotionally.

FG 2/P2

‘The doctor (still) has their part to play’: relying on health-care professionals to safeguard women’s interests

Focus group participants emphasised that health-care professionals remained an important part of the equation and women’s agreement to enrol in the potential trial did not – or should not – relieve health-care professionals of responsibility to consider and safeguard their interests. To the contrary, women expected health-care professionals to assess individual women’s suitability very carefully and consider whether or not, in the light of factors specific to that individual, they could truly claim to be in equipoise:

I think it’s important that the doctor is considering what the safest option is, and you’re only having the random choice made by the computer as long as there’s nothing else . . . no other factor that would be important to consider.

PG/P2

Participants highlighted how the clinical (and social) circumstances of PTB were highly variable and stressed that women wanted and expected to be treated as individuals, and counselled accordingly. Although participants did not necessarily see this a barrier to trial participation, they stressed that the information given to women about trial participation and the pros and cons of the two MoBs should be tailored to their particular circumstances. For example, FG 2 participants suggested that women’s perspectives and concerns might be quite different across the gestational age range covered by the trial and PG members suggested that health-care professionals should give women detailed information about the respective risks and benefits of the two MoBs at the gestation they were at. The PG members noted, for example, that there were different ways of performing a CS and that the type of incision made would be influenced by individual factors, including the size of the uterus and whether it was a singleton or twin pregnancy:

They could . . . say, ‘Well, these are the risks associated with us doing an incision this way, as opposed to that way, because the baby is extremely premature’.

PG/P1

Health-care professionals also had views on women’s willingness to participate, as described in Box 11.

Family and autonomy

Women from the three groups reflected on ownership of, and influences over, decisions about MoB and participation in the hypothetical trial. Women in the PG and in FG 1 (all of Indian ethnicity) mostly framed the decision as one they would make jointly with their husband/partner, with some questioning whether or not women could make such a decision independently. The potential for women and their husbands/partners to have different priorities, and for different opinions, therefore, to emerge, was noted by members of FG 1 (Box 12). One woman questioned what might happen should there be a marked difference of opinion between a woman and her husband/partner:

There could be a conflict of what happens, because there’s an emotional side and a physical side . . . If it’s a conflict, then what happens in that situation? Because the baby’s joint. Just because the woman’s delivering it, the responsibility is on both.

FG 1/P4

Another FG 1 participant commented that every relationship was different, and warned that not all were supportive/healthy. Therefore, she surmised, discussions about MoB and trial participation should, in the first instance, ‘be had privately with [women]. I mean, ultimately there’s a confidentiality issue’ (FG 1/P3).

Participants in the PG and in FG 1 further noted how the lives of women of Indian heritage might be closely entwined with those of their extended families. Some participants explained that they lived with their husband’s family (e.g. their mother- and father-in-law). Those who lived independently observed that their ‘almost westernised’ living arrangements were perhaps ‘not the usual Indian thing’ (PG/P1).

Women in these two groups noted that extended family could be an important source of support, but also (potentially) a cause of unhelpful interference and stress. In the case of decision-making about trial participation, some women surmised that older family members would be against this and, given the opportunity, might express their views strongly:

I live with my in-laws. They’re quite old-fashioned, so anything like this, they’d be like, ‘No, you’re not going to – don’t take part!’

FG 1/P3

However, others remarked that, in their case, these family members were/had been medical professionals and might, therefore, be quite supportive. Some PG participants suggested that although the perspectives of people other than their husband/partner might have little influence on the decision they made, the airing of opinions might nevertheless make that decision harder to live with. For this (and other reasons), these women said that they might limit the information they shared with their extended family.

Language and consent

Focus group 1 considered the issue of language and the implications of this for consent. Although all group members were professionals who spoke English as their first language, they noted that language might present a barrier to participation for some other women with Indian backgrounds. These group members warned that these women might be unclear as to what they were being asked to do, but feel ‘embarrassed to say that [they] don’t understand’ (FG 1/P2). This lack of understanding, they suggested, might discourage participation and lead to a less diverse sample than the triallists had hoped for. In addition, the group members emphasised that unresolved language barriers could have significant implications for the quality of women’s consent and raised serious doubts as to whether or not consent could, in such situations, be considered ‘informed’. Therefore, the group members suggested that, when English was not a woman’s first language, the involvement of an interpreter was absolutely key:

You need to have a translator explain everything, even if you think they understand, . . . because . . . it’s a big decision and . . . it’s important that they are actually fully informed.

FG 1/P2

Expectations of health care

Women in FG 2, all of whom were first-generation arrivals from Poland, reflected at length on the differences between reproductive and maternity health care in Poland and the UK. Theses women’s accounts suggested that experience of other health-care systems might create different expectations of UK health-care professionals and health care, as well as affecting perceptions of the desirability, or otherwise, of specific medical procedures, with these issues potentially having implications for decisions about participation in an intrapartum trial.

These women emphasised how pregnant women in Poland were monitored very closely, from conception onwards, and contrasted this approach with practices in Scotland/the UK:

[Here] until the 12th week, the pregnancy is not considered a pregnancy . . . they simply don’t support pregnancies prior to the 12th week . . . [whereas] in Poland they do everything to fight for these babies.

FG 2/P2

Another thing I’ve noticed here . . . is the attitude towards pregnant women . . . In Poland you have ultrasound scans done very frequently . . . they test your glucose levels . . . there’s loads of checks.

FG 2/P3

Women suggested that (what they perceived as) the more hands-off approach that is usual in Scotland/the UK could create unease among a community of women accustomed to, and/or expectant of, closer and more attentive care. Some women reported that this had prompted them to consult a Polish health-care professional privately, with one woman warning others that this had, however, led to further worries (because of the conflicting medical opinions that emerged on the appropriate management of their pregnancy).

Women noted how differences in care/management practices continued up to and, indeed, beyond birth, with several group members suggesting that there was a reluctance to admit pregnant women to hospital in Scotland/the UK until labour was quite advanced. Once admitted, some women felt that their choices were limited, with one woman asserting that ‘[i]n this country . . . there is an extreme pressure to give birth naturally’ (FG 2/P2). Women suggested that this included scenarios in which (in Poland) a CS would be considered the more appropriate option. For example, women noted how assisted/instrumental deliveries considered outdated and, indeed, ‘dangerous’ in Poland remained relatively common in the UK:

In Poland . . . forceps are not used any more . . . this practice has stopped many years ago.

FG 2/P4

With regard to how these (unmet) expectations of care and perceptions of good management/birth practices might affect decisions about trial participation, women suggested that they might fuel an innate (national) anxiety:

In my opinion we, Polish women, are anxious.

FG 2/P3

We are anxious, yes.

FG 2/P2

Women further surmised that trial participation might conflict with Polish women’s need to feel ‘in control of everything’ (FG 2/P3). Moreover, these women’s expectations of close and tailored care would seem incongruent with the protocol-driven care inherent in a trial and, as such, might plausibly discourage/disincentivise trial enrolment further.