Development of a co-designed behaviour change intervention aimed at healthcare professionals recruiting to clinical trials in maternity care

Abstract

Background

The evidence on what strategies can improve recruitment to clinical trials in maternity care is lacking. As trial recruiters, maternity healthcare professionals (MHCPs) perform behaviours (e.g. talking about trials with potential participants, distributing trial information) they may not ordinarily do outside of the trial. Most trial recruitment interventions do not provide any theoretical basis for the potential effect (on behaviour) or describe if stakeholders were involved during development. The study aim was to use behavioural theory in a co-design process to develop an intervention for MHCPs tasked with approaching all eligible potential participants and inviting them to join a maternity trial and to assess the acceptability and feasibility of such an intervention.

Methods

This study applied a step-wise sequential mixed-methods approach. Key stages were informed by the Theoretical Domains Framework and Behaviour Change Techniques (BCT) Taxonomy to map the accounts of MHCPs, with regard to challenges to trial recruitment, to theoretically informed behaviour change strategies. Our recruitment intervention was co-designed during workshops with MHCPs and maternity service users. Acceptability and feasibility of our intervention was assessed using an online questionnaire based on the Theoretical Framework of Acceptability (TFA) and involved a range of trial stakeholders.

Results

Two co-design workshops, with a total of nine participants (n = 7 MHCP, n = 2 maternity service users), discussed thirteen BCTs as potential solutions. Ten BCTs, broadly covering Consequences and Reframing, progressed to intervention development. Forty-five trial stakeholders (clinical midwives, research midwives/nurses, doctors, allied health professionals and trial team members) participated in the online TFA questionnaire. The intervention was perceived effective, coherent, and not burdensome to engage with. Core areas for future refinement included Anticipated opportunity and Self-efficacy.

Conclusion

We developed a behaviour change recruitment intervention which is based on the accounts of MHCP trial recruiters and developed in a co-design process. Overall, the intervention was deemed acceptable. Future evaluation of the intervention will establish its effectiveness in enabling MHCPs to invite all eligible people to participate in a maternity care trial, and determine whether this translates into an increase in maternity trial recruitment rates.

Background

It is widely acknowledged that clinical trials frequently fail to reach recruitment targets, with just 50% of publicly funded trials in the UK achieving optimal participant numbers [1]. The consequences for trials are far reaching, including underpowered results [2], increased costs, ethical implications [3], and ultimately contribution to ‘research waste’ [4]. However, within certain clinical areas, trial recruitment can face even greater challenges, for example, in clinical trials set within maternity care [5]. The participation of pregnant people in trials differs from the general population as it potentially affects two participants. The risk of teratogenicity and adverse pregnancy outcomes increases the perception of vulnerability surrounding the mother and baby dyad [6]. Our previous work, a qualitative evidence synthesis on recruiters’ perspectives of recruiting people to pregnancy and childbirth to clinical trials, found that Maternity Healthcare Professional (MHCP) recruiters often act as protective advocates, creating an additional gatekeeping barrier between trial and participant [7]. Factors such as these have likely contributed to limiting the number of eligible people participating in maternity trials [8, 9] and the evidence available to guide researchers recruiting to clinical trials in maternity care—synergistically creating a crisis for effective recruitment for trials in maternity care.

Empirical evidence on how to recruit participants to clinical trials in general is limited, with even less evidence on which strategies are effective. In the most recent Cochrane review of strategies to improve recruitment to trials [3], the majority of strategies were targeted towards potential participants rather than Healthcare Professional (HCP) recruiters. Only three of the recruitment interventions reported were aimed at HCPs, these centred around communication [10], training and ongoing support for clinicians [11], and evaluating email or postal invitations to general practitioners [12]. In addition to a lack of HCP focus, most of the interventions in the review are atheoretical and lack a predefined mechanism of action reported during the development. The National Institute for Health Research (NIHR) recently updated framework for the development and evaluation of complex interventions (which a recruitment intervention would likely be considered) recognises the need for theory to inform intervention development [13]. As Brehaut et al., note, even large-scale methodological initiatives such as the Studies Within A Trial (SWAT) repository [14] contains studies of interventions focusing on a single aspect of recruitment that lacks a theoretical basis [15].

The trial recruitment process has multiple components [16], and many of the process components within recruitment can be considered behaviours. Behaviours such as talking about a trial with potential participants, distributing trial information, and collecting trial consent, are all behaviours that HCPs may not perform during the normal course of their role [17]. Framing trial recruitment in behavioural terms provides a structure for researchers to systematically identify which behaviours are amenable to change and target them with an intervention [18]. Using a behavioural approach to understand trial recruitment barriers has already shown promise across a variety of clinical settings [19, 20] and in trial process interventions specifically [15, 21]. While the use of behavioural theory in developing trial recruitment interventions is a relatively recent advancement [17], the theoretical grounding of this approach can offer a substantiated explanation of the barriers and solutions for trial recruitment in the maternity setting.

The aim of this study is to use behavioural theory in a co-design process to develop an intervention aimed at changing the behaviour of MHCPs recruiting to clinical trials in maternity care and to assess the acceptability and feasibility of such an intervention.

Methods

This study is part of a larger programme of research; the ENCOUNTER project, a multi-phased, mixed methods project exploring behavioural barriers to recruitment to clinical trials in maternity care from the recruiters’ perspective and developing an intervention to target the barriers. Our previous theory-guided qualitative interview study [22] with twenty-two MHCP recruiters identified the factors enabling or inhibiting MHCP recruiters to invite all eligible women to participate in a maternity care trial. Four global themes, Availability and accessibility of resources, Navigating the recruitment pathway, Prioritising clinical responsibilities over research responsibilities, and The influence of colleagues and peers, mapped to the Theoretical Domains Framework (TDF), identified thirteen salient TDF domains relevant to the behaviour. The target behaviour of interest ‘Maternity Healthcare Professionals inviting all eligible people to participate in a maternity care trial’ was specified in our previous study [22] using the AACTT framework [23].

We conducted this study in three distinct steps (Fig. 1). Step 1 was to carry out a mapping exercise of the salient domains of the TDF to Behaviour Change Techniques (BCT) Taxonomy [24]. Step 2 was to hold a co-design workshop with stakeholders to discuss which BCTs to include in a recruitment intervention. Step 3 was to conduct an online survey to determine whether the resulting intervention was acceptable and feasible to a larger stakeholder group.

Fig. 1

Overview of steps in intervention development

Step 1—Mapping salient TDF domains to behaviour change techniques (BCTs)

The first step in the development of the intervention was to identify BCTs that could potentially target the salient TDF domains identified and reported in our interview study [22]. BCTs are theoretically derived and the smallest ‘active’ components of an intervention [24]. Using the online Theory and Techniques tool (an interactive ‘heat map’ matrix of 74 BCTs and 26 mechanisms of action) [25], we mapped TDF domains to BCTs, noting which BCTs were most likely to be effective in changing behaviour within a particular domain. The mapping process, undertaken by VH and LL, produced a long list of BCTs for potential use in an intervention to enable MHCPs to invite all eligible people to participate in a maternity care trial. The long list was narrowed down by the research team (VH, KG, LB) based on whether the BCT was pragmatic to operationalise within the scope of the current study (for example, non-modifiable organisational factors such as requiring additional staffing or consultation rooms were excluded). Next, suggestions were made (VH) as to how each of the BCTs shortlisted could be operationalised as a potential solution either as a stand-alone, or as part of an intervention package. For example, BCT 7.1 ‘Prompts/cues’ could be operationalised by providing MHCPs with lanyard cards listing trial inclusion criteria. Potential solution suggestions were discussed (VH, LB, KG) and checked by behaviour change experts on the team (ED, LL) to ensure fidelity with the BCT linked to the behaviour. The APEASE (Acceptability, Practicability, Effectiveness/cost-effectiveness, Affordability, Safety/side-effects and Equity) criteria were applied by members of the team (VH, LB, KG) to the remaining BCTs which informed the final selection that were taken forward to the co-design workshop [26].

Step 2—Co-design workshops for developing behaviour change recruitment intervention

In this step of the development, we included MHCP recruiters and maternity service users as stakeholders. This was to ensure the resulting intervention(s) were fit for purpose from the perspective of those who may use/receive it. We chose to hold two separate online co-design workshops, with two separate sets of participants, allowing the opportunity for each participant to contribute to the discussion.

Sampling

Previous participants from our interview study [22], who had consented to future contact, were invited (without obligation) to take part. We also invited maternity service users, who had experienced pregnancy within the past 2 years, via the ENCOUNTER Study Twitter account. Interested people were asked to email the research team and were sent a ‘Study Pack’ (including a brief summary of the ENCOUNTER project, workshop agenda, outline of potential solutions for discussion, and consent form). Participants returned a digital copy of signed consent. We allocated participants to one of two workshops based on their preference, clinical/professional role (midwife, nurse, doctor), the trial they were/had recruited for, and geographical location. This was to ensure, in so far as possible, that groups were balanced. Pilot workshops were held with MHCP recruiters and members of the trials community to check the coherence of the workshop content and format. Pilot participants had no previous involvement with the ENCOUNTER project, did not take part in the actual co-design workshops, nor was the data collected included in the analysis.

Workshops were held via Zoom in October 2021 and scheduled to last 90 min. Facilitation of both workshops was led by VH and co-facilitated by ED and LL. A brief summary of the ENCOUNTER project, an explanation of the ‘behavioural approach’ and workshop objectives was presented to participants using MS PowerPoint. In workshop 1, participants were asked to consider seven BCTs: Pros and cons, Goal setting (behaviour), Goal setting (outcome), Information about antecedents, Self-monitoring of behaviour, Review behaviour goals, and Review outcome goals. Each of these was presented with a description of how it could be operationalised as a potential solution to the target behaviour. Workshop 2 followed the same format, participants were asked to consider five BCTs: Information about health consequences, Information about social and environmental consequences, Reduce negative emotions, Information about emotional consequences, and Salience of consequences. The BCT 13.2 Reframing was implicit in both workshops as the premise of the intervention was to conceptualise ‘recruiting’ as ‘inviting’ all eligible people to participate in a maternity trial. During both workshops, participants gave their opinion on each potential solution presented and discussed how or if they could envision the potential solution being used in practice. Finally, participants were asked if any potential solution stood out as particularly helpful or unhelpful in inviting all eligible people to participate in a maternity care trial. In using co-design principles [27], we anticipated that participants would naturally begin to discuss BCTs beyond those presented. We decided a priori to make a list of any additional BCTs suggested by participants and crosscheck them against the original long-list generated by the mapping exercise. This allowed us to reintroduce any BCTs stakeholders deemed important that had been excluded earlier.

Discussions were audio recorded to facilitate transcription and also summarised in real-time to feed back to participants at the end of the workshop (LL). Directed content analysis was performed (VH) [28], focussing on participant-reported concerns and/or advantages related to each potential solution. The summarised findings were checked by all members of the research team by comparing findings to notes taken during the workshops by LL and ED. BCTs from both workshops were brought together and VH, KG, and LB met to ensure all BCTs remained within the APEASE criteria [26]. We used the GUIDED [29] and TIDieR [30] checklists to report the development and content of the intervention (Supp. File 1).

Step 3—Evaluating acceptability and feasibility of the proposed intervention

In the third step, we conducted an online survey to assess the acceptability and feasibility of the proposed recruitment intervention. A 3-min video introducing the proposed intervention and demonstrating a prototype of the ENCOUNTER app was recorded by the team and embedded within the online survey hosted by QuestionPro. The survey design was informed by the Theoretical Framework of Acceptability (TFA), a theoretical framework developed to assess the acceptability of healthcare interventions [31]. Participants were asked to respond using a five-point Likert scale to two belief statements and seven questions focused on each of the seven TFA constructs: Affective attitude, Burden, Ethicality, Intervention coherence, Opportunity, Perceived effectiveness, and Self-efficacy. An open text box was also provided for additional comments.

Sampling

We identified the two stakeholder groups that could potentially interact directly with the intervention: MHCPs and maternity trial team members. We invited members of these groups (based in Ireland or the UK) to complete the questionnaire. MHCPs who had taken part in or previously indicated interest in any phase of the study were emailed an invitation to participate. Chief investigators of relevant maternity care trials (identified through a clinical trials database search of Ireland and the UK) were also invited to take part. In addition, the link to the survey was promoted and shared on social media via the ENCOUNTER Study Twitter account. Consent was implicit by completion and submission of the questionnaire.

Results

Step 1—Mapping salient TDF domains to Behaviour Change Techniques (BCTs)

Initially, the mapping exercise resulted in 31 BCTs being identified as relevant to the target behaviour. From this long-list, 18 BCTs were excluded based on APEASE criteria. Table 1 presents further detail on the mapping process and justification for inclusion/exclusion. The remaining short-list of BCTs (n = 13) were then divided into two groups, broadly based on intervention function (i.e. how an intervention can change behaviour, e.g. education or modelling), to facilitate presentation to stakeholders at the co-design workshops.

Table 1 Results of mapping exercise (short-list of Behaviour Change Techniques (BCT) for potential inclusion in recruitment intervention)

Step 2—Co-design workshop for developing behaviour change recruitment intervention

Twenty-two MHCP recruiters were invited to participate in one of the co-design workshop; 10 agreed to take part; due to unforeseen clinical commitments, three later withdrew. MHCPs had a range of maternity trial experience in Ireland and the UK and clinical backgrounds that included midwifery, nursing, and obstetrics. Six maternity service users expressed an interest in participating; of those, two women who had birthed within the past 6 months consented to take part. Both women had previous experience of participation in clinical research but not in a clinical trial. Further details on participant characteristics are presented in Table 2.

Table 2 Recruitment intervention co-design workshops: participant characteristics

Workshop 1 (duration: 99 min) included five participants. Workshop 2 (duration: 106 min) included four participants. Participants offered comment on each potential solution, and the group discussed and made suggestions about how they might be adjusted, improved, or delivered in practice. Participants in both workshops suggested the intervention could be delivered as part of mandatory training. However, the majority of participants were strongly opposed to this idea, stating that the agenda for mandatory training days was already full and adding something further would necessitate losing something equally important. Participants were concerned that in the current climate (COVID-19 pandemic) MHCPs did not have the appetite for more training and it would become “just another ‘thing’ they had to do”.

Many participants commented on how mobile phone apps had recently become a useful tool in recruitment. Voice notes were discussed as a particularly helpful method of communication between MHCP recruiters. Group members shared scenarios where they regularly used voice notes to instruct other recruiter colleagues about how to communicate trial specific information to potential participants. One participant had found voice notes especially useful as a means of facilitating remote recruitment, by communicating directly with women about the trial during the COVID-19 pandemic.

Following a discussion summary, participants were asked if they considered any potential solution to be particularly helpful or unhelpful in inviting all eligible people to participate in a maternity trial. In workshop 1, the majority agreed that having ‘experienced recruiters give a talk (or video clip) about resolving situations that commonly occur prior to inviting a woman to a trial’ (BCT 4.2—Information about antecedents) was the most helpful. There was no agreement on the least helpful potential solution as participants all selected different BCTs. In workshop 2, participants agreed that all the potential solutions presented to them would be helpful in addressing the barriers. Furthermore, participants creatively suggested that some solutions could be combined to make a more practicable intervention. Tables 3 and 4 provides a brief summary of participant discussion around each potential solution presented at the workshops.

Table 3 Summary of trial recruitment intervention co-design workshop discussion with maternity healthcare professionals and maternity service users: workshop 1

Table 4 Summary of trial recruitment intervention co-design workshop discussion with maternity healthcare professionals and maternity service users: workshop 2

Six additional BCTs were mentioned by participants during the discussions, all of which featured on the initial long-list generated during the mapping exercise in step 1. Five BCTs had already been excluded based on the APEASE criteria and were not carried forward. The remaining one (10.4 Social rewards) was brought forward to the next stage of development.

The APEASE criteria were applied to the aggregated findings of both co-design workshops: this produced the final list of BCTs for inclusion in the intervention (Tables 5 and 6). In total, 10 BCTs were confirmed for inclusion in the final intervention (1.2 Problem Solving; 2.3 Self-monitoring of behaviour; 4.2 Information about antecedents; 5.1 Information about health consequences; 5.3 Information about social and environmental consequences; 5.6 Information about emotional consequence; 9.2 Pros and Cons; 10.4 Social Rewards; 11.2 Reduce negative emotions; 13.2 Reframing).

Table 5 APEASE criteria applied by ENCOUNTER team to behaviour change techniques following co-design workshops: intervention 1

Table 6 APEASE criteria applied by ENCOUNTER team to behaviour change techniques following co-design workshops: intervention: intervention 2

Our choice of intervention delivery was informed by comments made by participants about the convenience and helpfulness of using mobile phone apps in their working practice and the high number of BCTs we wished to incorporate into one intervention. We developed a prototype app to deliver the intervention in this setting (Fig. 2). It should be noted, however, that in this study, we were interested in the BCT content of the intervention, rather than the technological functions of the app, and therefore, we focused predominantly on content during acceptability and feasibility assessment.

Fig. 2

ENCOUNTER app prototype screenshots

Step 3—Evaluating acceptability and feasibility of the proposed intervention

The final step in the development process involved assessing the acceptability and feasibility of the intervention using an online survey capturing the opinions of key stakeholders. Fifty-four participants, twenty-five from Ireland and twenty-nine from the UK, initiated the online questionnaire, providing demographic information and answering the two statements questions. However, nine participants did not proceed beyond the two statement questions, resulting in a total of 45 completed questionnaires (reported in section below). Responses were provided by MHCPs, including midwives (clinical and research), research nurses, doctors (obstetricians and other specialties), allied health professionals, and maternity trial team members. All respondents described themselves as white and were predominately female 92.6% (n = 50). Research midwives represented the largest group of respondents (n = 25, 46%), followed by clinical midwives (n = 8, 15%) and trial team members (n = 8, 15%). Detailed participant characteristics are presented in Table 7.

Table 7 Recruitment intervention survey: participant characteristics

In response to the first statement question on the survey: “Improving recruitment to trials in maternity care is something I care about”, over 98% (n = 53) of respondents agreed or strongly agreed. In response to the second statement: “I have a role to play in helping to ensure all eligible people are invited to participate in a clinical trial in maternity care”, 100% (n = 54) agreed or strongly agreed with the statement.

Forty-five respondents proceeded to complete the TFA questions on the questionnaire. Below, we give a brief report on the responses to the seven TFA questions and provide detailed survey results in Table 8. Just over half (60%, n = 27) of respondents perceived that the ENCOUNTER app would be effective in helping MHCPs offer the opportunity to take part in a trial (TFA construct: Perceived effectiveness). More than 66% (n = 30) of respondents were enthusiastic about using or seeing the app in use (TFA construct: Affective attitude). With regard to the burden of engaging with the app, over 77% (n = 35) thought it would be easy or required some effort, however, 18% (n = 8) believed it would be difficult or require a lot of effort to engage with it (TFA construct: Burden). Respondents were less positive when considering how using the app might impact other work practices, with 58% (n = 25) believing it would interfere with other tasks MHCPs need to do (TFA construct: Anticipated opportunity). Just over half (53%, n = 24) were confident or very confident about MHCPs being able to engage and use the app, while 47% (n = 21) of respondents were doubtful that MHCPs would be able to do so (TFA construct: Self-efficacy). When asked about the clarity of the proposed intervention, how it would be used, and how the app might enable more eligible people to be invited to trials, 73% (n = 33) felt clear or somewhat clear in their understanding of how the intervention might work (TFA construct: Intervention coherence). When asked if they had any ethical concerns about the app, and how or if it would fit into their healthcare systems, 80% (n = 36) of respondents had no ethical concerns. However, 53% (n = 24) believed that the app would need to undergo some adjustments in order to fit in with the local health system (TFA construct: Ethicality).

Table 8 TFA online survey results: acceptability of the prototype intervention

Discussion

The majority of existing recruitment interventions concentrate on modifying the information given to potential participants about the trials [3], and do not address recruitment staff training, despite being identified as a top priority for recruitment intervention development [32]. Townsend et al.’s systematic review of trial recruiter training programmes found most evidence originated within the oncology setting, where interventions tended to be delivered as multidisciplinary workshops covering generic and trial specific issues [33]. The review authors note that while there is limited evidence to support the efficacy of any particular training programme, some interventions were reported to increase HCPs communication skills. However, as this increase was self-reported, with no indication of the duration over which the change was sustained, caution should be observed when interpreting this finding. While there is an absence of recruiter training interventions focused on maternity trial recruitment, there has been recent research activity in non-trial specific maternity recruitment research [34, 35]. A qualitative exploration of midwives recruiting to maternity research in a hospital setting found the competing worlds of clinical practice and clinical research governed midwives motivation to recruit [34]. The authors recommended involving clinicians in planning and designing strategies to overcome factors that inhibit recruitment [34]. Rose et al.’s study, a theory-informed online survey which included 22 community midwives recruiting to maternity research studies, identified six TDF domains as key influencers in the perceived barriers and enablers to approaching patients about research participation [35]. Four domains (Environmental context and resources, Knowledge, Social Influences, Goals) share commonality with the eight salient domains identified in our previous findings [22]. However, the domain Behavioural Regulation was reported as rarely evident by Rose et al., which is contrary to our findings, where Behavioural Regulation featured as a salient domain more than once. While the authors suggest addressing behavioural barriers and enabling midwives to approach potential participants requires a package of theory-based interventions, they stop short of specifying which BCTs to include in the package [35].

The opportunities for behavioural theory to improve clinical trials is being recognised across several trial processes [17]. Recent examples in the literature show behavioural approaches being used as a ‘diagnostic’ tool to identify recruitment problems [15, 19, 36]. The advantages of adopting this approach are demonstrated by Brehaut et al. in their study with HCPs recruiting to a trial concerning recurrent venous thromboembolism [15], and Castillo et al. in their study concerning haematologists recruiting patients to a CAR(T) cell therapy trial [19], where the use of behavioural theory revealed a variety of additional factors impeding recruiters that had not been uncovered by previous research efforts. Previous studies have largely focused on ‘diagnosing’ recruitment problems; however, behavioural theory can move beyond ‘diagnostics’ towards ‘treating’ recruitment problems by informing the development of behaviour change interventions [21, 37]. Ellis et al. take this approach in their study developing an implementation intervention for cancer clinical trials aimed at HCPs [21]. The authors ‘brainstormed’ BCTs and potential modes of delivery, to iteratively shape a multi-modal trial process intervention to align with opportunities and support structures within the context of community practice. Of the 27 different BCTs identified by Ellis et al., six shared commonality with our intervention: Information about social and environmental consequences, framing/reframing, review outcome goals, goal setting, self-monitoring, and information about consequences. However, there were several task performance orientated BCTs (i.e. instructions, demonstrations, how to perform) that were not identified in our work. Ellis et al. used a mixture of in-person, telephone, and electronic communications and, similar to our approach, chose online video as a supportive mode of delivery for the intervention in a practice setting [21]. Other teams have explored the use of behavioural theory to develop patient-centred interventions to improve trial participant retention [37]. Newlands et al.’s. study explored two interventions, selected by participants, that centred on goal setting and motivational information. Their goal setting intervention was based on a single BCT (1.1—Goal setting behaviour) for communication from the HCP point of contact or peers from trial. Their motivational information intervention included three BCTs (5.1—Information about health consequences; 5.2—Salience of consequences; 5.3—Information about social and environmental consequences) for communication between participant and trial office at various key touch points in the trial. The intended mode of delivery for both interventions was by verbal, paper based, or electronic means. While the above BCTs also featured in our intervention, the mode of delivery differed. Our intervention included a large number of BCTs and made innovative use of the feedback on BCTs to guide the mode of delivery for the intervention.

There are few examples of trial recruitment intervention development in the literature that use stakeholder accounts or make explicit reference to in-depth assessment of the HCP reported recruitment challenges. The QuinteT Recruitment Intervention (QRI) is an example where stakeholder experiences have been incorporated into the interventions and successfully improved recruitment rates [38]. The QRI is a two-phase intervention that gathers evidence at the clinical site(s) about trial specific recruitment processes and then produces a bespoke plan to address the difficulties [38]. We utilised stakeholder involvement in a different way. From inception onwards, we included MHCP recruiters through all stages of development, building on the accounts of MHCP recruiters from a variety of trials in maternity care. In addition, we included maternity service users in the co-design phase and maternity trial teams in the acceptability and feasibility stage. This stakeholder responsive approach allowed us to include additional BCTs and tailor the prototype, to produce a proactive pragmatic intervention primed for success from the outset.

Evaluation is key for all interventions. It is essential that theory-informed trial recruitment interventions are evaluated in a robust manor to determine whether or not the evidence from changing health behaviours extends to changing MHCPs trial behaviours [37]. One method to consider for evaluation is a Study Within A Trial (SWAT). The self-contained research study embedded within a host trial, offers the potential to test the intervention in a real-world context while making a contribution to the pool of methodological evidence [14]. We kept the necessity for evaluation foremost during development. To pave the way for future robust evaluation, we ensured our intervention aligned with the NIHR Framework for the development and evaluation of complex interventions [13].

Strengths/limitations

We developed a theoretically grounded intervention that was acceptable to MHCPs and maternity trial teams, across Ireland and the UK. While our results are limited to these countries, they may be transferable to regions beyond, but it would be important to determine whether the behavioural barriers present in other countries were similar to those identified in our initial analysis. A key strength of our study was stakeholder engagement. We included MHCP recruiters, maternity service users, and maternity trial team members, at key stages of intervention design and development. Perspectives from MHCPs with experience across multiple maternity trials, covering a range of clinical areas and interventions in maternity care, fed directly into the development of the intervention, increasing the transferability of our findings. Holding two separate online workshops strengthened our findings as it allowed time for rich discussion and helped stave off potential power imbalances by achieving an even spread of participants from different clinical roles across workshops. However, we acknowledge the influence of potential bias in the workshop allocation in relation to the different agendas planned for each workshop. There were limitations in the diversity of our stakeholder groups, and the sample is not a direct reflection of the population the intervention was developed for. MHCPs in this study had previously participated in our qualitative interview study [22] and may have been more agreeable to the proposed potential solution. In mitigation of this, we assessed the acceptability of the intervention using an online survey, to potentially reach a larger more diverse group of stakeholders. However, despite our efforts, the sample demographics of the survey revealed a lack of heterogeneity.

In using behavioural theory, despite the guidance of empirically derived tools, we acknowledge that there remains a degree of subjectivity. The application of APEASE criteria is not an exact science and other researchers may have categorised differently. However, applying APEASE criteria is also a strength as it ensures the content of an intervention is tailored to the specific context. Using behavioural theory and co-design in tandem to approach intervention development added synergistic value beyond that to be gained by using either approach independently [39].

Conclusion

This study reports the development of a theory-based, codesigned intervention aimed at changing the behaviour of MHCPs tasked with approaching all eligible potential participants and inviting then to join a trial in maternity care. We used findings from our previous interview study with MHCP recruiters [22] and applied theory and co-design principles to develop each component of the intervention. We welcome rigorous evaluation of our prototype in a SWAT, primarily to establish if the intervention enables MHCPs to invite all eligible people to participate in a maternity care trial and secondly, to determine if this had an impact on the overall rate of recruitment to the trial. While recruitment to trials in maternity care is highly complex and not limited to changing the behaviour of MHCPs alone, we are hopeful that the theoretical grounding of our intervention will support healthcare professionals to invite all eligible people to participate in clinical trial in maternity care.