Exploring implementation of intrapartum trial evidence: a qualitative study with clinicians and clinical academics

Abstract

Background

Implementing research evidence into clinical practice is challenging. This study aim was to explore implementation of two intrapartum trials with compelling findings: BUMPES (position in second stage of labour in nulliparous women with epidural), and RESPITE (remifentanil intravenous patient-controlled analgesia).

Methods

A qualitative interview study set in UK National Health Service Trusts and Universities. Purposively sampled investigators from RESPITE and BUMPES trials and clinicians providing intrapartum care: midwives, anaesthetists, and obstetricians, were recruited using existing networks and snowball sampling. Semi-structured virtual interviews were conducted. Thematic analysis was underpinned by Capability Opportunity Motivation Behaviour Change Framework.

Results

Twenty-nine interview participants across 19 maternity units: 11 clinical academics, 10 midwives, 4 obstetricians, 4 anaesthetists. Most (25/29) were aware of one or both trials. BUMPES had been implemented in 4/19 units (one original trial site) and RESPITE in 3/19 units (two trial sites). Access to sufficient resources, training, exposure to interventions, support from leaders, and post-trial dissemination and implementation activities all facilitated uptake of interventions. Some clinicians were opposed to the intervention or disagreed with trial conclusions. However competing priorities in terms of staff time and a plethora of initiatives in maternity care, emerged as a key barrier to implementation.

Conclusions

Compelling trial findings were not implemented widely, and numerous barriers and facilitators were identified. Large-scale improvement programmes and evidence-based national guidelines may mean single trials have limited potential to change practice. There is a need to examine how intervention implementation is prioritised to optimise safety outcomes in the context of workforce restrictions, limited resources and large arrays of competing priorities including statutory requirements, that have increased in maternity care.

Contribution to the literature

• Implementation science theories and frameworks are largely absent from maternity services research, particularly intrapartum care.

• This study is the first to explore factors influencing implementation of findings from two compelling randomised-control-trials in intrapartum care, using the COM-B behaviour change framework

• Trial findings were not widely adopted. Multiple barriers were identified relating to clinicians’ capability and opportunity including skills to apply evidence.

• Clinicians are required to implement multiple policy and research recommendations whilst often not having the adequate training, tools or capacity to do so. Incorporating implementation science within maternity services can help guide realistic evidence-based practice whilst optimising implementation processes.

Introduction

Randomised controlled trials are the gold standard of clinical evidence, however implementation of findings can take years, or may never occur [1, 2]. Although there have been studies on uptake of evidence in maternity care going back to the last century, [3] the use of theory, models and frameworks to underpin implementation in maternity is limited compared with other settings [4, 5]. It has therefore been suggested there is a need for implementation science approaches to be tailored and tested in maternity contexts [6]. The lack of maternity implementation research has been highlighted as a priority in the USA to improve evidence-based practice, and address disparities in maternal health [7].

BUMPES [8] compared upright with a left or right lateral position in the second stage of labour for nulliparous women with epidural. It found a 5.9% absolute increase in the chance of spontaneous vaginal birth when lying down (adjusted risk ratio 0.86, 95% confidence interval 0.78 to 0.94) [9]. RESPITE [10] explored remifentanil intravenous patient-controlled analgesia as an alternative to intramuscular pethidine for women in established labour. The absolute risk of the primary outcome, progression to epidural, was lower with remifentanil (19%) than pethidine (41%) (risk ratio 0·48, 95% CI 0·34–0·66; p < 0·0001). Remifentanil resulted in fewer instrumental and more spontaneous vaginal births. Both trials, if implemented, have potential to support reduction of assisted births, which have increased risks of maternal [11] and neonatal [12] morbidities compared to normal birth. The trials’ data collection was completed in 2014 (BUMPES) and 2016 (RESPITE) with articles published in 2017 and 2018 respectively (see Supplementary material 2 for more trial information) while this study was conducted early 2023.

This study aimed to explore the influences on implementation of trial findings in NHS maternity care using behaviour change theory. Two recent intrapartum care intervention trials published in high impact journals were selected as exemplars to explore implementation, BUMPES [8] and RESPITE [10]. In the years following publication of trial findings, researchers and clinical colleagues in the authors’ professional networks had expressed concerns that the interventions had not been widely adopted, but reasons were not understood.

Methods

Study design and theoretical framework

This was a qualitative study involving semi-structured interviews. We used the Capability, Opportunity and Motivation (COM-B) behaviour change framework [13, 14] to explore influences on clinicians’ implementation of trial findings. COM-B has been applied to understand implementation in other maternity contexts [15]. It defines three domains of influence on individual behaviour (B): capability (C), opportunity (O), and motivation (M). These domains interact, and influences on behaviour can be negative or positive. The domains are described in Table 1.

Table 1 Themes and subthemes arranged using the COM-B framework

Setting and eligibility

Clinicians and clinician researchers working in UK National Health Service inpatient maternity units and/or higher education institutions were eligible to take part (Table 2).

Table 2 Participant characteristics

Participants, sampling and recruitment

Participants were purposively sampled from four groups to gather a range of experiences of maternity care and trials: maternity clinical academics affiliated with higher education institutions who held leadership roles in the RESPITE and/or BUMPES trials (chief, principal or co-investigator); senior midwives (e.g. intrapartum matron), obstetricians and anaesthetists working in NHS intrapartum care settings (Table 2). We aimed to recruit 4–12 individuals from each group (total 24–36). Participants were identified via the authors’ existing clinical and research networks, open recruitment at conferences, and snowballing via other participants.

Data collection

Virtual semi-structured video interviews were conducted November 2022-April 2023. Participants provided written or verbal consent following information sharing about the study. A structured topic guide (Supplementary material 3) explored trial implementation facilitators and barriers. This was developed to include topics from each of the Capability, Opportunity and Motivation framework domains. Question clarity was assessed, and wording slightly adjusted after the first three interviews to ensure an effective flow. Participants were offered coffee vouchers as compensation. Interviews were audio recorded digitally, transcribed verbatim and anonymised. Interviews were conducted by [FCS] a midwife and research fellow (described below). A reflexive diary was maintained throughout [16].

Data analysis

Thematic analysis was conducted using a deductive codebook approach [17, 18]. Following familiarisation with the data, two researchers (FCS, BT), open coded an exemplar transcript. They developed an initial coding index, organising codes and categories within COM-B domains, with separate codes for each trial. In Table 1 we describe the themes and present them alongside the relevant domains of the COM-B framework. The index was applied to the remaining data. Codes were reviewed and discussed iteratively by FCS, BT and ND, analytical summaries were written by FCS, integrating and comparing analysis from both trials. Initial analysis was discussed with public contributors and clinical and non-clinical academics before writing up final themes. Analysis continued until data saturation was established. Qualitative data were managed using the QSR NVivo 12 software programme [19]. Consolidated criteria for reporting qualitative research (COREQ) reporting guidelines were followed as recommended to improve the standard of qualitative study reporting [20].

All authors are experienced maternity services researchers, with backgrounds in midwifery (FCS, SK), public health (BT), obstetrics (RL) and implementation science (ND).

Patient and public involvement and engagement

The design, conduct, analysis and interpretation of findings were regularly discussed with public contributors from NIHR ARC West Midlands at throughout the study.

Results

Out of 39 participants approached, 29 (74%) were interviewed (mean duration 40 min, range 29–53) from 19 NHS hospital boards/trusts and five Higher Education Institutions (Table 2).

Themes and subthemes are presented alongside the relevant COM-B domain in Table 1.

Knowledge and skills

Knowledge of the evidence

Most participants (n = 25) were aware of one or both of the trials, whether they were based in a trial site or not. However, many had limited recall and understanding of the trials, with some stating incorrect findings.

I thought that it had shown that the upright posture in labouring women with an epidural was associated with a higher change of having a successful vaginal delivery. RB27 Anaesthetist Non-trial site.

Skills to interpret evidence

It was suggested that clinicians faced challenges in engaging with, interpreting and applying evidence, particularly midwives who had fewer undergraduate and postgraduate research training and practice opportunities.

Our midwives in our unit they’re wonderful, but they’re overworked, they don’t really get to attend a lot of study days… I feel that quite a lot of the midwives in our unit are a bit behind with their knowledge stuff. RB18 Anaesthetist Non-trial site.

The physical environment

Workforce

Workforce issues were a repeated and prominent barrier, as staff did not have capacity to adopt new practices with widespread reports of burnout and low morale.

Midwives at the moment are extremely burnt out, and any suggestion of extra work, extra research, people are just shutting off. So it’s a really hard time to be going up there and be like, “When you’re in there remember to do this,” because people are just fed up… RB06 Midwife BUMPES trial site.

Remifentanil is a narcotic and can suppress respiration, and therefore requires one-to-one midwifery care and respiratory monitoring. Often participants saw implementation of remifentanil as almost impossible to operationalise due to insufficient staffing.

We used to use remifentanil PCA [Patient-Controlled Analgesia] in on the labour suite, and we stopped using it… the concerns really are about the level of monitoring required compared to the midwifery staffing. RB29 Anaesthetist non-trial site.

Guidelines and policies

Clear and accessible guidelines and policies were described as promoting adoption. However, these were not always available and varied between hospitals.

We have a great guideline… Our on-calls are run by trainees, so registrars or SHOs [Senior House Officer]. I think they’re pretty open minded about its use, but as they come and go they want really clear guidelines, which I think we have. So they’re happy if they know they’re safe and okay doing it. RB27 Academic.

Resources, funding and supply

Participants described varied availability of equipment such as pumps to administer remifentanil, and balls, beds and leg supports to support a lateral position in labour. Funding impacted on equipment and training availability, and supply shortages were reported for remifentanil.

We were rationing it to women who needed it because they couldn’t have an epidural rather than to anyone who wanted it. So then we had women coming through wanting to have it but having to be told unfortunately we can’t offer that. RB20 Academic.

The clinical and social context

Clinical cultural norms

Adoption of both interventions was hindered by local embedded practices, a culture of slow adoption of change, negative perceptions of interventions and a perception that evidence was not locally relevant.

You can do the most amazingly robust trial with the most amazing findings, which are clearcut, which BUMPES was, and you will still get people saying, “Oh well, it wouldn’t work in our unit,” or, “Our women are different.” RB12 Academic.

Leaders’ influence

It was suggested that staff deferred to respected medical leaders, and that midwives had relatively low influence. Endorsement by professional bodies such as Royal Colleges was described as helpful.

[Our hospital] does not use remifentanil as a routine…our anaesthetists did not believe that it is a safe drug for labour… there’s enormous respect for our anaesthetists, and if they say they don’t want to use it we go “Okay fine.” RB01 Academic.

Training and exposure to the intervention

Staff exposure to the interventions varied depending on whether organisations had participated in the trials, implemented RESPITE and BUMPES and/or provided training. Staff turnover or rotation necessitated continuous training. Seeing interventions in practice was reported to influence attitudes and behaviour.

Most participants stated that findings had not been implemented locally. Four reported that the BUMPES intervention was in regular use (representing 4/19 units, one BUMPES trial site). Three reported that remifentanil was in regular use (representing 3/19 units, two RESPITE trial sites) (Table 3). The lack of widespread adoption reduced staff exposure to the intervention and training opportunities.

Table 3 Awareness, links and reported adoption of intervention among participants working clinically or linked to BUMPES and RESPITE sites

Midwives need to be trained in monitoring and looking after women with remifentanil, to be familiar with it… if this is a technique that you’re only reserving for women with a contraindication to an epidural, are you then facing a situation where you have staff who are not that familiar with it suddenly doing it in a thrombocytopenic woman. RB07 Anaesthetist non-trial site.

Women’s experiences

Participants widely reported that women did not receive information about interventions, for example posters and antenatal classes promoted upright labour position, and leaflets did not include remifentanil. Participants described how women’s positive or negative experiences of the interventions during the trials influenced subsequent adoption.

Women who had been in the [RESPITE] trial with their first pregnancy… it was getting such good feedback, and the number of women coming back and saying “Can I have it again?” was big. RB09 Academic.

The trial and intervention context

Trial dissemination and implementation efforts

Findings from both trials were published in high-impact peer-reviewed journals, though at the time of data collection they did not feature in national guidelines. It was suggested that academic outputs did not reach many clinicians.

I personally don’t think that midwives read the BMJ [British Medical Journal] any more than an obstetric anaesthetist would. I think the journal was wrong. Even though it’s a really high impact journal and that’s what we aim for. RB29 Anaesthetist non-trial site.

BUMPES was published in BMJ and Midwives journal (which all members of the Royal College of Midwives receive). Other dissemination activities were reported by participants to have included sending summaries and posters to trial sites, presentations at national and international conferences, and a BUMPES celebration event. Participants also suggested that dissemination should focus on collaboration and include support from external peer organisations.

While there were examples of implementation activities (for example a ‘BUMPES champion’ in one site encouraged and reminded midwives to offer women lateral position in second stage), participants articulated a need for more work to translate evidence into practice. This was a general issue rather than specific to RESPITE and/or BUMPES.

You’re not going to reap the benefits of that investment in the trial unless you change practice. So I think it’s quite a big issue about wasted resources, and research waste, which is such a big issue. If we do all this really well conducted research and then don’t implement the findings, then all that investment was pointless. RB03 Academic.

Involvement in the trials

Being a trial site provided clinicians with a head-start with resources, training and exposure to the interventions compared to non-trial sites.

For a unit which didn’t recruit into the trial, who are trying to introduce something entirely new, however well evidence-based, that as well as the fact that resources are so difficult…I could quite understand why they just hold up their hands and saying, “We can’t do this, it’s just not possible.” RB20 Academic.

However, trial participation was not a guarantee of sustained practice, with delays between study end and reporting, and limited influence and involvement of researchers in clinical areas.

We sadly because of the hiatus…it [practice] didn’t change I would say, but partly it’s because we don’t have a remit within the clinical research team about evidence implementation. RB08 Academic.

Individual clinician motivation

Competing priorities

RESPITE and BUMPES interventions had to be balanced with women’s needs and preferences, and clinical and administrative responsibilities. Implementing research findings was challenging in a context of competing priorities, including national safety and quality initiatives.

I think most people just go “Gosh we ought to do everything”…it is also led by things we have to do like CQC [Care Quality Commission reporting] and Ockenden [recommendations to improve maternity safety], and we become so full of change that research sometimes can fall to the bottom of that pile. RB13 Obstetrician non-trial site.

Perceptions and beliefs about the intervention and trial findings

It was suggested that individual clinicians varied in their willingness to change established routines and practices. Some clinicians additionally perceived trial effect sizes as small, that single trials had limited relevance, or were sceptical about the validity of results. BUMPES evidence was considered challenging as it contradicted strongly held beliefs about the role of gravity in labour.

You think that is the right answer, and when something goes against that idea…there’s a very, very big fixed paradigm, and then quite a small effect, and it’s one trial. RB01 Academic.

Polarised views regarding the safety of remifentanil were reported. Some clinicians cited fear and anxiety regarding remifentanil use with concerns about safety, and the confidence, competence and resources required to care for women using it. It was suggested that this could be tackled through training and exposure to the intervention.

The staff I think were nervous…they think people are going to go into respiratory arrest at the drop of a hat. Of course that hasn’t happened because the dose is so low, and the controls are in place… there’s training around its confidence of course. RB14 Midwife RESPITE site.

Discussion

Main Findings

RESPITE and BUMPES findings had not been widely adopted, with limited knowledge of the trials and concerns that many, particularly midwives, lacked skills to apply evidence. Workforce shortages and resource availability were prominent barriers. Clinicians described difficulty in implementing new practices alongside complex clinical tasks and mandatory duties. Maternity unit cultural norms and senior leader perspectives could enable or hinder implementation, as could individual clinician beliefs and attitudes. Women’s perspectives had limited influence as they were rarely informed or offered the interventions. Incorporation of interventions in guidelines or policies, and clinician exposure to interventions through training and practice facilitated adoption in some units. Being a trial site encouraged but did not guarantee sustained adoption. While dissemination and implementation activities were welcomed, many felt more was needed. Most influences were common across contexts and trials, with some intervention and trial-specific barriers. Remifentanil, a potent opiate, involved specific safety concerns and workforce requirements. BUMPES findings contradicted reported beliefs, teaching and practice about the role of gravity in spontaneous vaginal birth (SVB).

Interpretation

The data provides evidence that our initial impression was correct: neither intervention was widely adopted or sustained. Publication does not guarantee adoption [1, 2, 21,22,23], and the many generic barriers to evidence-based practice are also well-known [24]. Publication of RESPITE and BUMPES in leading journals did not effectively reach professionals or women. While the need for diverse approaches to dissemination is recognised, [23] the practicalities and resource implications are substantial. Our findings also suggest gaps in research knowledge and evidence interpretation, particularly among midwives, supportive of calls for increased capacity building for this group [25,26,27].   Including findings into midwifery and obstetric training would be beneficial but the practicality of incorporation into a busy curriculum may be challenging. While research active organisations more readily adopt evidence, [28] our findings indicate that effective interventions may not be sustained even where hospitals host trials. This highlights the need to optimise the translation of evidence into practice [6, 7, 29]. It is sometimes argued that trial clinicians should bear a responsibility for dissemination and engagement regarding findings. However, they may be seen as conflicted and professional and government bodies like NICE may be better positioned to champion findings in the public interest. Possibly the main function of investigators is to push for their findings to be assimilated by policy makers and professional associations at home and abroad. Work before, during and following trials, underpinned by implementation theory, can identify and mitigate challenges, including whether clinicians perceive interventions as necessary, theoretically sound, safe, feasible, affordable and relevant [30, 31]. RESPITE and BUMPES interventions are complex, and increasingly, process evaluation is undertaken alongside complex intervention trials to interrogate implementation [32].

Additional ‘generic’ implementation barriers, included shortages of staff, equipment and drugs, training and support, and gaps in guidelines, policies and patient information [4, 23, 24]. The study was conducted following the COVID-19 pandemic at a time of low morale and staff shortages [33]. This was a very different context to the timing of the original trials, and clinicians also lacked the clinical research support, common purpose and incentives that trial participation brings [34].

Our study highlights how individual practitioner, intervention and organisational context and policy are critical, [23] and generic implementation barriers interact with more specific issues. For example, remifentanil requires staff to provide one-to-one care because of its potential side effects. There was reluctance to accept trial results, or their generalisability. Safety concerns regarding the use of narcotics added another barrier. The RESPITE trial was powered on effectiveness but not safety. The number needed to treat may be much lower than number needed to harm, but the latter could represent a much more serious event, and a reason for non-adoption [2]. Where effectiveness is not enough to change practice, trials might have to be designed to demonstrate safety [35]. Some interventions do not challenge cultural norms and can be easily incorporated into existing routines, such as adoption of planned caesarean section for breech presentation favoured by many obstetricians, [36, 37]. However, BUMPES and RESPITE required people to change practice. Given the difficulty of making change, staff instead carried on as before, with minimal incentive or pressure to change, meaning cultural norms had not shifted.

It is necessary to temper expectations regarding the role of single studies in changing practice. Regardless of quality, they may not provide the evidence required: when the BUMPES trial was incorporated in an updated Cochrane review exploring birth position, it resulted in a minor alteration, but was not sufficient to change the overall conclusions [38]. Some may be cautious to immediately adopt trial findings before clinical guidelines are updated [2]. Incorporating interventions in clinical guidelines may expedite adoption, but frequently occurs some years after trial publication, to coincide with updates. In 2023 the updated English National Institute for Health and Care Excellence (NICE) Intrapartum Care guidelines [39] added recommendations from both BUMPES [8] and RESPITE [10], six to seven years following publication of findings and after this study’s data had been collected. While the guideline states a lateral position may increase SVB, it recommends women can use any comfortable position in the second stage of labour. Remifentanil was recommended for women wanting ‘ongoing pain relief during labour and birth’ [39].

Even where evidence-based guidelines and policy are mandated, substantial challenges must be overcome, and they may not be implemented [33, 40]. In a recent freedom of information request about intrapartum analgesia 55/108 responding UK NHS trusts offered remifentanil (18 offered it only where epidural was contraindicated), due to shortages of staff and/or remifentanil [41].

As in other areas of healthcare [2], UK maternity services are required to adopt many interventions [42] which they must prioritise in a context of significant safety concerns. It may be that interventions that are mandated and include financial incentive have higher priority such as the Clinical Negligence Scheme for Trusts [43]. Cognitive overload is linked to burnout, errors and poor outcomes [44,45,46]. Mitigating the cognitive burden which the ever-growing evidence base and policy context brings could create capacity to adopt effective practices. Strategies include decision support tools, machine learning [45], incorporating time considerations into guidelines [47] and de-implementation of low-value interventions and policies [48]. Consolidating and prioritising interventions may simplify the task for clinicians. NHS England produced a maternity and neonatal delivery plan in response to calls for “One clear plan that looks to encompasses the recommendations from various reports.” [49]. However, this does not consolidate evidence-based guidelines and emerging research, and further work is required to understand how to support clinicians to navigate this.

Strengths and limitations

To our knowledge this is the first study to explore maternity trial implementation across more than one intervention or study. Participants included intrapartum care clinicians and trial investigators from most UK regions, with a range of professional backgrounds and experiences. The study was underpinned by behavioural science which gave a structured, theory-based framework to explore implementation. The authors were not involved in the trials, but had professional connections to some trial investigators. While their relative ‘insider’ status enabled open discussion, it may have influenced participants’ willingness to share negative views, and the researchers’ interpretation of findings. It is likely that clinicians with more positive attitudes to research were willing to participate. Quantitative summaries of trial awareness and adoption must be interpreted with caution due to the small non-random sample.

Conclusion

This study identified maternity trials with compelling findings were not implemented widely, for reasons documented extensively in implementation science literature. It also indicates single trials may have limited capacity to change practice in the context of evidence-based guidelines underpinned by synthesis of multiple studies. Maternity professionals must navigate and prioritise ever-growing policy, guidance and research recommendations amid quality and safety scrutiny, while tailoring increasingly stretched services to local populations and service contexts. Rigorous, theory-based implementation work must be embedded throughout the maternity research cycle and in practice. However, researchers and policymakers have a duty to deliver evidence and guidance that is realistic and balanced, to support clinicians to prioritise care delivering the best outcomes for women and babies, including de-prioritisation and de-implementation. Further work is required to understand how this can be achieved.