Introduction

In keeping with global trends, Non-Communicable Disease (NCD) dominates morbidity and mortality in the UK [1]. This position is unsustainable for the health and care system. A fundamental shift is required to move from a reactive treatment service to a proactive prevention-focused National Health Service (NHS) [2,3,4] that supports self-management by people living with long-term conditions [5, 6]. One way of achieving this is through a values-based, person-centred approach that enables people to effectively self-manage their long-term medical conditions with appropriate support from healthcare services [6]. Done well, up to 80% of people could self-manage their long-term medical conditions using this model of care [6].

Physical inactivity is responsible for almost 10% of the major NCDs, including heart disease, type 2 diabetes, and breast and colon cancers [7]. People living with long-term conditions are amongst the least active and stand to gain the most from even minor improvements in physical activity levels [8]. Regular contact with this hard-to-reach group makes healthcare a critical component of population approaches to addressing inactivity [9,10,11]. Successful integration of behaviour change strategies, which promote self-management, into routine care, including changes in healthcare professionals’ consultation behaviour, remains elusive [12]. Consequently, the management and care of people with long-term conditions are still perceived as the healthcare professional’s responsibility, rather than an active collaboration between empowered patients and a healthcare system delivering effective self-management support [5].

In keeping with the UK’s ‘Make Every Contact Count’ initiative [13], the National Institute for Health and Care Excellence (NICE) identify routine clinical conversations between healthcare professionals and patients as a vital interface to unlocking patient-driven behavioural change [11, 14]. Person-centred conversations and behaviour change are intertwined: conversations can effectively develop autonomous motivation to adopt and sustain desirable behaviours. On the other hand, the conversation itself consists of interaction behaviour that can be learned and thus changed. As part of a whole-system approach, targeting this interaction behaviour between healthcare professionals and patients may be fundamental to changing clinical practice in the NHS [5, 15].

Encouragingly, most healthcare professionals perceive conversations about physical activity to be important (ranging from 70% [16, 17] to over 90% [18,19,20]). Despite this, a gap exists between the proportion of times healthcare professionals perceive patients would benefit from brief opportunistic advice and the frequency with which they deliver such interventions. However, there is a disparity between perceived importance and the frequency of conversations on physical activity [12, 21] demonstrating that although healthcare professionals are receptive to the objective of physical activity promotion, a wide range of barriers exist, both individual and organisational, to putting it into practice [21, 22].

Healthcare professionals among primary and secondary care groups lack the knowledge, skills and confidence to have physical activity conversations underpinned by behaviour change theory [18, 22,23,24,25,26,27,28,29,30,31,32,33]. Whilst healthcare professionals are vocal in their support of behaviour change and self-management, they frequently minimise their ability and responsibility to deliver behavioural change work [12, 32]. Although many factors contribute to this avoidance, time concerns and previous negative experiences are powerful deterrents [22, 34]. Consequently, when healthcare professionals attempt to engage patients in conversations about change, it is often a one-sided transaction that focuses on delivering information based on the healthcare professional’s agenda for change, denying the individual the opportunity to take up more time or offer resistance [35]. Emphasis on other components of medical management reinforces this approach, such as medication review and assessment of biomarkers, which are more familiar to healthcare professionals and given greater priority by the systems in which they work. Addressing the broader context of conversations in clinical practice is essential since it is not simply a lack of time that is the issue, but prioritisation amongst the other vital components of medical management.

Collaborative discussion using evidence-based behaviour change techniques to build on a person’s thoughts about and motivation for change is more effective, better received and often more time-efficient than directive interactions [36, 37]. To help promote patient engagement and empowerment, good conversations on physical activity may use a guiding rather than directing style [34, 38]. Focus should be on the likelihood of an individual to change their own behaviour and therefore incorporate skills to emphasise autonomy such as empathic listening, fit into the available timeframe and agree individualised solutions driven by the individual’s agenda [34]. A mindset change is required to move the conversation from “what’s the matter with you” to “what matters to you”. Therefore, changing conversations on physical activity is not as simple as teaching interaction skills or telling healthcare professionals that it is important since these conversations reflect interactions that happen across an entire system.

Change in the medical workforce requires a complex combination of behaviours, decisions and interactions between healthcare professionals, patients, families and other parties. Developing an enabling culture that includes training for core skills and supporting resources to support healthcare professional behaviour change forms a central component of this complex system [39]. The effectiveness of behaviour change interventions to promote change is frequently limited by a lack of integrity with which these complex skills are delivered [40]. Whilst it is perhaps beyond the scope of practitioners to embed a comprehensive treatment fidelity framework within their clinical practice [41], an example of a practical strategy to enhance self-management would be to employ an enactment framework using video or audio recordings for patient consultations. Videos can then be reviewed and coded by the practitioner or independently by a trained third party, potentially including the use of instruments to assess specific behaviour change skills such as the implementation of Motivational Interviewing [40].

Healthcare professionals require the capability, opportunity and motivation to change their own behaviour in order to influence their patients through conversations about physical activity. We set out to understand these behavioural determinants of physical activity conversations to develop educational resources for healthcare professionals that could help them in routine clinical practice. We identified a range of high-quality pre-existing educational resources available on physical activity, such as the Swedish scientific handbook Physical Activity in the Prevention and Treatment of Disease [42]. However, we observed a lack of translation into clinical practice.

The Medical Research Council consider a theoretical basis essential for the successful development of complex interventions in healthcare [43]. The Behaviour Change Wheel (BCW) is an implementation model developed from synthesising 19 different behavioural change frameworks [44]. It provides a comprehensive structure addressing behavioural factors within nine intervention functions and seven policy categories and is advocated for use in this context by NICE [45]. The BCW helps contextualise change on an individual level and the determinants of what needs to happen to achieve organisational or system-level change. It has been used successfully to develop similar interventions improving the Capability Opportunity Motivation-Behaviour (COM-B) of healthcare professionals to deliver physical activity interventions in cancer care [46], therapeutic radiography [47], gestational diabetes [48], depression [49] and prevention of psychosis in ultra-high risk young people [50]. We believe this is the first study to utilise the COM-B framework to inform interventions focused on improving the frequency and quality of conversations on physical activity in managing long-term conditions. We are aware of studies that have used the COM-B framework to assess changes in self-reported knowledge and skills to deliver brief advice on physical activity following training [47] and studies that have explored health professionals’ practice in health care contexts such as mental health using the framework [51].

This study aims to use the BCW to analyse the behaviour of healthcare professionals and outline a coherent approach for developing interventions to improve the frequency and quality of conversations on physical activity by healthcare professionals across clinical practice.

Methods

Overview

The BCW outlines eight steps towards interventional design incorporating behavioural analysis using the COM-B model to understand and explore behaviour [52]. This model allowed us to draw on a parallel workstream using the COM-B model to understand the behavioural factors influencing healthcare professionals’ capability, opportunity and motivation in a national pilot developing a physical activity service in secondary care [53].

We worked through each stage of the BCW following the recommended structure summarised in Table 1 [52]. Each behaviour change component maps onto nine intervention functions (education, persuasion, incentivisation, coercion, training, enablement, modelling, environmental restructuring, and restrictions) and seven policy strategies (Environmental/social planning, communication/marketing, legislation, service provision, regulation, fiscal measures and guidelines) [44, 52]. The COM-B model recognises that behaviour is seated at the heart of this complex interacting system involving the capability (both physical and psychological), opportunity (social and physical) and motivation (reflective and automatic) of an individual or group to perform a particular behaviour [44, 52]. We expanded our behavioural analysis with the Theoretical Domains Framework (TDF), which is a framework compromising 14 domains to help identify and describe the factors influencing a particular behaviour [54, 55]. The TDF helped us achieve a deeper exploration of the barriers and facilitators to behavioural change amongst healthcare professionals and strengthen the links between the theories and techniques of behaviour change in order to anticipate and address implementation challenges [54, 56]. Following behavioural analysis and the identification of intervention options, we identified promising Behavioural Change Techniques (BCT) to inform successful intervention design [52, 57]. BCTs make up the active ingredients of interventions that allow them to be evaluated and replicated when identified in the design and evaluation of projects [45, 57,58,59].

Table 1 Developmental stages of the COM-B model
Full size table

Stage 1 understanding behaviour

Step 1 define the problem in behavioural terms

To address step 1, over 8 weeks, the authors worked with a range of healthcare professionals working across different clinical pathways in the Active Hospital project to understand the problems faced by healthcare professionals. In addition, we, the authorship team, had weekly meetings to discuss and refine answers to the following questions defined in step 1 of the BCW:

  • What is the behaviour?

  • Where does the behaviour occur?

  • Who is involved in performing the behaviour?

In addition, we undertook a broader scoping review [60] interrogating published evidence around physical activity conversations in clinical practice to address the question:

  • What are common barriers to performing the behaviour?

Step 2 select the target behaviour

In step 2, we considered all factors that interventions to increase the quality and frequency of physical activity conversations could target in routine clinical care.

When deciding which behaviours to target, we considered the following four factors recommended by the BCW to inform which options are likely to be the best intervention targets:

  1. (1).

    The potential impact of behaviour change

  2. (2).

    Likelihood of the intervention leading to behavioural change

  3. (3).

    The impact of this behavioural change on other system components, for instance, engaging in an education program to improve skills, may increase the use of a resource to help conversations in clinical practice. The behaviour change wheel categorises this as a ‘spillover score.’

  4. (4).

    How easy and practical it will be to measure the target behaviour

We explored possible solutions through (1) our clinical networks across two regional meetings involving 70 multidisciplinary professionals and patients from community and hospital rheumatology and musculoskeletal services and (2) service managers, multidisciplinary leads and patients across three inpatient and one community pathway in the active hospital project [53]. We met to rate potential target areas in the four domains as unacceptable, unpromising, but worth considering, promising or very promising [52]. We made decisions by majority consensus, following discussion that considered results from the scoping review and reflected on clinical group feedback and personal experience.

Step 3 specify the target behaviour

In step 3, we explored the nature and characteristics of the target behaviours defined in step 2 in more detail and considered the context in which each behaviour occurs. Questions we addressed as a group included who needed to undertake the behaviour, what they needed to do, and where and when they might do it. If we were unclear about the application of the target behaviour, we spent time exploring contrasting clinical pathways to identify common characteristics. This step helped generate discrete target areas for influencing behaviour.

Step 4 identify what needs to change

We identified the capability, motivation and opportunity factors required to change the identified target behaviours based on the scoping review of the literature, feedback from clinical groups, departmental meetings with clinical service leaders, and discussions with inpatients involved in the active hospital project. We subsequently used the theoretical domains framework to add context to each behavioural target by working the targets through diverse clinical pathways and identifying areas of commonality. This process helped us generate discrete targets with a theoretical rationale to change practitioner behaviour successfully.

Stage 2 identify intervention options

Step 5 intervention functions

In a paper exercise, we mapped COM-B components from the behavioural diagnosis onto intervention functions according to the BCW. First, we met as a group to discuss and assess each intervention function using the APEASE criteria – Affordability, Practicability, Effectiveness and cost-effectiveness, Acceptability, Side effects and safety and Equity [52]. Following this, we mapped selected intervention functions onto behavioural targets.

Step 6: policy categories

We identified policy categories reported in the literature and those highlighted by healthcare professionals during active hospital pathway development. As with intervention functions, we met to assess policy categories according to the APEASE criteria and mapped relevant policy categories onto behavioural domains defined in the BCW [52].

Stage 3 identify content and implementation options

Step 7. Behaviour change techniques

BCTs form the active ingredients of interventions and enable coherent approaches to evaluation [61]. The BCW identifies the most frequently used BCTs by intervention function referencing ‘BCTTv1’ - a comprehensive taxonomy of 93 BCTs developed by international expert consensus [57]. We used a snowballing approach to augment BCT data from studies identified during our scoping review [60]. We identified systematic reviews reporting BCTs with promising/proven efficacy in physical activity behaviour change interventions in clinical practice [62,63,64,65,66,67,68,69]. Following this exercise, we met to map the promising BCTs onto intervention categories and identify suitable implementation strategies drawing on results from the scoping review and feedback from healthcare professionals in the active hospital clinical pathways and working groups.

Step 8. Mode of delivery

The final step of the BCW is to develop a delivery framework based on a recognised taxonomy of delivery modes [52]. A review of interventions that change healthcare professional behaviour [70] informed our delivery framework development, and we assessed each category using the APEASE criteria and consensus amongst ourselves.

Results

Stage 1 understanding behaviour

Step 1 define the problem in behavioural terms

We agreed on the following answers to the questions posed in step 1:

  • What is the behaviour? Healthcare professionals initiating person-centred conversations on physical activity at all appropriate opportunities in routine medical care

  • Where does the behaviour occur? Across the spectrum of healthcare provision for managing and treating people living with long-term medical conditions. Delivery will range from community and primary care settings to hospital inpatients in secondary and tertiary care facilities. Settings may include clinic rooms, wards, day rooms and all other environments delivering healthcare services, including remote or telehealth consultations.

  • Who is involved in performing the behaviour? All healthcare professionals.

  • What are common barriers to performing the behaviour? Consistent barriers to physical activity conversations in clinical practice are reported amongst a range of healthcare professionals practising in various clinical domains across numerous countries. Table 2 summarises barriers identified during the scoping review [60]:

Table 2 Summary of barriers to physical activity conversations in clinical practice
Full size table

Step 2 select the target behaviour

Following analysis of the potential behavioural targets, we agreed on the initiation of conversations on physical activity by healthcare professionals as the primary target behaviour. This discrete and tangible target can be easily measured and, if achieved, is likely to prompt healthcare professionals to develop their skills. Furthermore, once initiating conversations becomes a part of healthcare professionals’ routine consulting practice, they are likely to influence others’ practice positively. Table 3 outlines the behavioural analysis.

Table 3 Prioritising behavioural interventions
Full size table

Step 3 specify the target behaviour

After specifying the primary and secondary target populations and behaviours, we specified the behavioural target regarding who, what, where and when the behaviour is performed. This helped us generate the discrete target areas for influencing behaviour outlined in Table 4.

Table 4 Specifying target behaviours
Full size table

Step 4 identify what needs to change

The behavioural diagnosis identified barriers and facilitators to all six core components of the COM-B model. Expanding each domain using the theoretical domains framework helped us define tangible targets for intervention design, as demonstrated in Table 5.

Table 5 Behavioral diagnosis and theoretical domain mapping
Full size table

Stage 2 identify intervention options

Step 5 intervention functions

In total, we selected six out of nine intervention functions. We deemed incentivisation not affordable or practical, coercion not practical or acceptable, and restriction not practical for implementation in general clinical environments. Table 6 maps intervention functions onto behavioural targets.

Table 6 Selecting intervention functions
Full size table

Step 6: policy categories

We selected 5 out of 7 policy categories, although regulation relies on external bodies for practicability. Despite this, we agreed to include regulation because service leaders in clinical pathways we were developing in the active hospital project were keen to change the systems regulating practice by healthcare professionals in their pathways, for instance, through the electronic records system. Fiscal and legislative policy categories were deemed impracticable and unacceptable. Table 7 outlines an assessment of each policy category and examples of potential delivery mechanisms for a resource to support conversations on physical activity.

Table 7 Identifying policy categories to support intervention delivery
Full size table

Stage 3 identify content and implementation options

Step 7 behaviour change techniques

We identified 17 promising BCTs and mapped these onto BCW intervention functions, as demonstrated in Table 8. Selected BCTs included 1. Goals and planning, 2. Feedback and monitoring, 3. Social support, 4. Shaping knowledge, 5. Natural consequences, 9. Comparison of outcomes, 12. Antecedents, and 15. Self-belief.

Table 8 Mapping BCTs to COM-B component, intervention functions and implementation strategy
Full size table

Step 8. Mode of delivery

We focused on population-level delivery approaches to make a resource accessible to as many healthcare professionals as possible. In addition, Digital channels predominate to make the project affordable and broaden its reach. Table 9 outlines our review of the BCW delivery framework.

Table 9 Defining the intervention delivery framework using APEASE criteria
Full size table

Discussion

This study uses the BCW to outline a coherent approach for intervention development to improve the frequency and quality of conversations on physical activity by healthcare professionals managing long-term conditions. Time-sensitive and role-specific resources will help healthcare professionals understand the focus of their intervention. Educational resources aimed at healthcare professionals and patients will have mutual benefit, should fit into existing care pathways and support professional development. A trusted information source with single-point access via the internet will improve accessibility and provide an ideal delivery mechanism for a wide range of resources, including an avenue for distributing free promotional information.

Our concurrent clinical activity in the development of three inpatient and one community clinical physical activity pathway in the Active Hospital pilot provided an ideal environment to explore and test promising ideas from published literature. We balanced our behavioural analysis across community and hospital environments. However, we recognise there is a risk of a bias toward understanding the implementation landscape in a hospital environment, potentially limiting the applicability of our findings to interventions in other settings, such as primary care. Although our clinical backgrounds positively impact the clinical relevance of this study, the quality of our work is potentially limited by the lack of a robust academic background in behaviour change amongst the healthcare professionals in our team. This impacted some of our decision-making; for instance, we included modifying patient behaviour as a secondary behavioural target due to promising literature suggesting the benefits from this approach without behavioural analysis of patients themselves. We identified the risk of our academic limitations at the outset, which informed our decision to use the BCW due to its straightforward and step-wise guidance. As others have reported, we discovered that following the BCW system is an exacting challenge [46]. Systematically following all steps was laborious and time-consuming, but it ensured consideration of all components of effective behaviour change [109, 110]. We followed the model diligently; for instance, we spent time defining the primary behavioural target despite other authors deeming this unnecessary [46]. As we subsequently progressed through the stages, we found that defining the target was a great benefit as it helped us maintain focus on changing the consulting behaviour of healthcare professionals rather than the physical activity behaviour of their patients. In some areas, we found the scope of the challenge exceeded our resources and looked to previously published evidence for guidance. For example, we narrowed down BCT choice by identifying promising BCTs in the published literature. However, failing to fully consider and explore all 93 BCTs on their individual merits may mean we missed effective BCTs whose use may be novel in this area.

A growing body of evidence demonstrates the potential of time-efficient behavioural change approaches in clinical practice [82, 111, 112]. People living with long-term conditions value and welcome behavioural change support on physical activity from healthcare professionals [69, 113]. However, traditional transactional models of clinical consultation offer an over-simplistic and ineffective approach to encouraging behavioural change. This model of medicine, established over generations, is not without limitations when considering straightforward prescription, such as antihypertensives [114] or even major surgery, such as solid organ transplant [115]. Conversations to support behavioural change should start with the individual and consider personal choice, circumstance and behavioural context, suggesting the traditional consultation model of ‘diagnose and treat’ requires a rethink [116]. This study confirms that successful resources should consider individual preference, circumstance, behavioural context, and system constraints such as appointment length to support physical activity conversations effectively.

Whilst education and training alone are insufficient to change healthcare professional behaviour, embedding education and training opportunities into a practical structure to support routine practice improves practitioner engagement and increases the likelihood of behavioural change [117]. Healthcare systems must value and promote such an intervention as the prevailing professional, and organisational culture may be most influential in changing practitioner behaviour [70, 102, 118]. Success may be when healthcare professionals habitually include person-centred physical activity conversations in their practice. Given the range of competing interests on their time, automisation of their behaviour through habit formation is likely to free up cognitive capacity [119]. Although automaticity has been successfully targeted in simple healthcare tasks such as hand washing, complex tasks such as physical activity conversations appear less conducive to habit-forming [120]. However, targeting specific behavioural components may be a way around this challenge [119]. For instance, system support can influence habit formation through intervention such as integrating prompts in computer systems or clinical pathways.

As well as prompting habit formation, building educational resources into routine care by supporting real-time decision-making and providing point-of-care prompts for best practice can enhance professional development [102]. Such education strategies have a more significant impact when derived from influential opinion leaders [70, 98]. Developing strategies informed by likely barriers and facilitators of behavioural change to translate research findings into clinical practice can further enhance effectiveness [121, 122]. Digital approaches successfully support clinical decision-making and the delivery of preventative care [98, 102, 123, 124]. Delivery via the internet supports several behavioural domains identified in this study and is a simple way to deliver a scalable and cost-effective intervention [122, 125].

Future research may include understanding how to leverage the influence of patients on healthcare professional behaviour and improve habituation within complex communication skills. Greater understanding is required of how healthcare systems and the professionals within them can best balance the fundamental medical requirements of long-term condition management with individualised and person-centred behavioural change support. Designing interventions with evaluation in mind is critical to help understand the optimal approach to increasing the frequency and quality of conversations on physical activity across clinical practice. To this end, the findings of this study have informed the development of a hybrid online resource combining educational material with conversational guidance coded with BCTs to support evaluation [44, 59]. We encourage independent evaluation of the resources at www.movingmedicine.ac.uk and call upon researchers to focus on improving our understanding of what works to improve conversations on physical activity across clinical practice.

Conclusion

We have iteratively developed a framework using the BCW to improve healthcare professionals’ capability, opportunity and motivation to have person-centred conversations on physical activity. The framework is grounded in the priorities of busy healthcare professionals addressing a range of barriers, including time, knowledge, skills and system support. At the heart of a successful intervention lies the principles of person-centred care and an approach that may be unfamiliar to healthcare professionals trained in a didactic consultation style. Resources need to be time-sensitive and role-specific, whilst educational resources aimed at healthcare professionals and patients will have mutual benefit, should fit into existing care pathways and support professional development. A trusted information source with single-point access via the internet will improve accessibility and provide an acceptable delivery mechanism for a wide range of resources. All healthcare team members have a role in delivering constructive physical activity support.

Building practical resources based on this framework will improve efficacy, integrate the principles of behaviour change and provide a platform to inform future research and develop clinical physical activity services. Therefore, we encourage open evaluation of resources built using this framework to help improve understanding and implementation of what works.