The intervention was developed in accordance with the MRC guidance for the development of complex interventions (13,17) and took a theory and evidence based approach. The BCW (14) in conjunction with elements of the Person-Based Approach (15) provided a systematic guide to intervention development, incorporating user perspectives, with the aim of changing CBEPs behaviour. The intervention was refined through an iterative and dynamic process based on evidence, theory and feedback from intervention recipients, patient and public involvement (PPI) members, stakeholders and an expert working group with expertise in: behaviour change (SR, LS, RT), complex intervention development (ST), cancer research (LB, DR), urology (DR) and qualitative methodologies (ES) (Figure 1).
Regulatory and ethical approvals, in accordance with the Helsinki Declaration, were sought prior to the commencement of research activities from Sheffield Hallam University (Reference: ER10748795) and the NHS (REC reference: 18/NW/0738 / IRAS project ID: 254343). Written informed consent was collected from all participations prior to research activity.
Step One: Understanding the behaviour
To develop a complex intervention centred on supporting patient behaviour, we firstly sought to understand patient needs i.e. what support do men on ADT for prostate cancer need to exercise. We then identified how to train CBEPs to deliver the required support to men on ADT, as detailed below.
Identification and selection of target behaviours
A list of potential target behaviours for patients and CBEPs were generated by reviewing i) existing literature examining the effectiveness of interventions aimed at exercise initiation and maintenance of cancer survivors and the role of exercise professionals (18) and ii) results of 5 focus groups with men on ADT for prostate cancer (n = 26), exploring their experiences of physical activity and beliefs about participating in a structured exercise programme (methods reported elsewhere (8)).
The comprehensive list of behaviours was reviewed by the expert working group for replication or cross-over between behaviours. Each behaviour was then coded according to defined criteria from the BCW (14), to identify behaviours that should be targeted in the intervention: a) the likely impact of change, b) the likely ease of change, c) the centrality of the behaviour (e.g. the likelihood of changing one behaviour having an impact on another behaviour) and d) the ease of measurement.
Specification of target behaviours
Next, the target behaviours were specified in behavioural terms by the expert working group, and reviewed by the PPI group (7 men with prostate cancer and 4 family members); who needs to deliver the behaviour, what does a person need to do differently, when will it happen and where will it take place. To consider dose we also considered how often the behaviours are required and with whom.
Identifying what needs to change
To further develop our understanding of how CBEPs may support men on ADT to exercise, barriers to, and enablers of the patient target behaviour were identified in the previously reported focus groups (8).
Similarly, barriers and facilitators of the CBEPs target behaviours that could be addressed in a training programme were identified from four focus groups. Participants were identified using purposive sampling to ensure inclusion of the target population (community-based personal trainers, physiologists, and fitness managers) and relevant stakeholders (gym general managers and operational managers). Each focus group was led by an experienced facilitator (ES, LS, SR) and the topic guide, reviewed by the PPI group, was designed to explore the 14 domains of the Theoretical Domains Framework (TDF) version 2.0 (19) (Additional file 1). The TDF is an integrative framework originally based on the synthesis of 128 theoretical constructs from 33 theories of behaviour change (20) to understand behaviour at an individual level and support the development of implementation interventions (21). Furthermore, CBEPs were probed on their preferred mode of delivery of the professional training programme (e.g. online versus face-to-face).
Responses from the focus groups were mapped onto the TDF to identify domains through which change may occur. Focus group transcripts were analysed deductively supported by the use of NVivo (22); the TDF was used to generate the framework for content analysis (SR) and coding specifically identified the type of professional (or patient) who provided the information and the type of professional (or patient) the information was related to (e.g. themselves, a different exercise professional, a HCP or patient). The lead researcher (SR) systematically went through each transcript, coding according to the framework. Text was attributed to more than one domain where applicable. Text relating only to the target behaviours were coded for relevance (21). Subsequently, 25% of coding was cross checked by a second independent reviewer (LS), and discussions informed the development of the coding framework as described above (i.e. who provided the information and the type of professional the information was related to).
Theoretical underpinning and evidence base
Following the behavioural analysis, psychological theories from the field of behaviour change that related to the identified constructs within the TDF were then reviewed to consider how best to apply them to the current context and advance our understanding of the likely mechanisms of change. In line with MRC guidance for the development and evaluation of complex interventions, a logic model was developed to present the proposed mechanisms of change of the intervention (23).
Step Two: Identifying behavioural content and implementation options
Behaviour change techniques
Having identified which theoretical constructs required change to achieve both the patient and CBEPs target behaviours, content for the intervention was developed and guided by the inclusion of behaviour change techniques (BCTs). BCTs are defined as observable, replicable, and irreducible components of behaviour change interventions (14). We used the labels and detailed definitions of BCTs as those included in the BCT Taxonomy version 1 (BCTTv1) (24).
BCTs to be delivered by the CBEPs to support long-term exercise behaviour were identified from previously drawn links between constructs of the TDF (14), mechanisms of action (25) and previous literature exploring BCTs for promoting exercise behaviour in people living with and beyond cancer (18). Context-based decisions for each BCT that was identified as likely to be effective (i.e. most frequently used with the selected TDF domains), and supported by evidence, was made in relation to its affordability, practicability, effectiveness and cost-effectiveness, acceptability, side effects and safety and equity considerations (APEASE criteria) (14).
Similarly, inclusion of BCTs for the delivery of the exercise professional training were guided by the CBEPs target behaviours and theoretical constructs requiring change. BCTs were identified from previously drawn links between constructs of the TDF (14) and mechanisms of action (25). BCTs that were identified as effective (i.e. most frequently used with the selected TDF domains) were reviewed against the APEASE criteria and CBEPs focus group findings to determine their suitability in the intervention (14).
Mode of delivery
Once the BCTs had been identified, the behavioural content was then developed, specifying the mode of delivery. A range of feasible modes of delivering the selected BCTs were considered and reviewed against the APEASE criteria (14) to identify content to be delivered via traditional face-to-face approaches and online learning technologies to promote active and self-directed learning (26).
Furthermore, intervention materials including patient booklets and professional training worksheets were created alongside a training manual for CBEPs and a facilitator manual for the training providers.
Step Three: Delivery and refinement of the intervention
Modelling components of a complex intervention prior to a full-scale study provides important information about the design of the intervention (13). We sought to operationalise the CBEPs training package during an iterative process, delivering the intervention face-to-face to CBEPs (n = 11) and stakeholders (n = 28) between February and June 2019, for feedback and refinement of the intervention.
Rehearsal delivery and subsequent focus groups
The exercise professional training package was delivered to CBEPs who were purposively sampled to ensure a diverse range of target users (e.g. gender, job role and previous experience) (27). Feedback on the content, format, location and delivery of the intervention were collected immediately post intervention, via an audio recorded focus group with an independent researcher (RT). The topic guide was based on Kirkpatrick’s Four Level Training Evaluation Model (28) (Additional file 1) and considers reaction - concerned with understanding how participants feel about the training programme, learning - the extent the attendees have learnt something new such as skills, behaviour - the extent training is put in to practice and results that are obtained as a result of the intervention. Focus group data were analysed in NVivo (22) following Braun and Clarke’s six step process of thematic analysis (29). Suggestions for change were highlighted from the key themes and rated against MoSCoW criteria (Must have, should have, could have, would like) for prioritising change (30). Modifications were made if they were considered likely to impact on behaviour change, uncontroversial and easy or recommended by multiple participants.
Stakeholder workshop and feedback
Following the rehearsal delivery, the exercise professional training package and patient intervention materials were presented at a one-day stakeholder workshop. Participants were identified via existing clinical, professional and patient networks, national charity representatives and our PPI group. The stakeholder workshop ran in the format of presentations from the research team, presenting ‘key uncertainties’ for discussion. Stakeholder discussions were facilitated (maximum of 8 per group) by a research team member using a broad topic guide (Additional file 1) based on the Normalisation Process Theory (NPT); a framework that utilises the core concepts of coherence, cognitive participation, collective action and reflexive monitoring to capture the work that participants do when implementing a new practice (31). Verbal feedback and researcher field notes were collated and circulated to all stakeholders to ask for any further comment or clarification within a two-week deadline. Feedback was then mapped onto the NPT and reviewed against MoSCoW criteria for prioritising change (32), to support intervention refinement, as described above.