Context

Introduction

Practitioners use a range of interventions when working with someone to improve their health. Each intervention will usually involve one or more behaviour change techniques. However, there is a lack of practical advice on which techniques should be used to tackle specific behaviours (for example, in relation to diet, smoking and alcohol) and with people from specific populations or with particular risk factors.

For an individual to improve their health in the medium and long term, behaviour change must be sustained. Maintaining changes to behaviour can involve both helping people to deal with relapses, and ensuring that new behaviours become habitual.

Sustained behaviour change is most likely to occur when a combination of individual, community and population-level interventions are used. In addition, there is a reasonable evidence base relating to motivation to change (Lai et al. 2010; Ruger et al. 2008).

In 2011, the House of Lords Science and Technology Select Committee reviewed a range of factors that impact on behaviour change. In its final report, the Committee recommended that NICE should update its guidance on the topic; in particular it wanted, 'more explicit advice on how behaviour change techniques could be applied to reduce obesity, alcohol abuse and smoking' (House of Lords 2011).

Classifying behaviour-change techniques

Considerable research has been undertaken to specify behaviour change interventions in terms of their component parts. This has led to a definition of behaviour change techniques relevant for a range of health behaviours (Michie et al. 2013) and for specific behaviours:

  • to improve their diet or encourage physical activity (Abraham and Michie 2008; Conn et al. 2002; Inoue et al. 2003; Michie et al. 2011a)

  • to prevent weight gain (Hardeman et al. 2000)

  • to stop smoking (Michie et al. 2011b)

  • to reduce alcohol intake (Michie et al. 2012)

  • to prevent HIV (Albarracin et al. 2005).

Work is currently underway to explore the extent to which techniques may be applicable across different behaviours. The classification system has been shown to be reliable. Its validity is now being assessed (Michie et al. 2013).

Theoretical frameworks

The importance of having a theoretical basis for the design and evaluation of interventions is well established (Medical Research Council 2008; Craig et al. 2008). For example, it can help ensure better outcomes (Albarracin et al. 2005) as well as providing a means of understanding why an intervention is effective or not.

Work has been done to establish theoretical frameworks for behaviour change (Abraham and Michie 2008; Michie et al. 2011a; West 2009) and evidence continues to emerge about these theories (Tuah et al. 2011; Williams and French 2011).

Considerations

Background

This section describes the factors and issues the Programme Development Group (PDG) considered when developing the recommendations. Please note: this section does not contain the recommendations.

There is a wealth of information and recommendations in existing NICE guidance on interventions related to the behaviours covered in this guidance: alcohol use, eating patterns, physical activity, sexual behaviour and smoking. The PDG did not aim to update or critique these recommendations. Rather, it considered new evidence to add value to the recommendations already made.

The PDG agreed that the principles in the NICE guideline on behaviour change: general approaches relevant to the remit of this guidance were still applicable. These were: principles 1 (planning), 3 (education and training), 4 (individual-level behaviour change interventions and programmes), 7 (effectiveness) and 8 (cost-effectiveness).

This guidance focuses mainly on individual-level behaviour change interventions. However, the PDG agreed that these need to be viewed in the context of a range of other interventions. This includes those delivered at population and community level and those related to the environments in which choices about behaviours take place (see the NICE guideline on behaviour change: general approaches).

The PDG noted that tackling behaviour change among people younger than 16, in particular in relation to issues such as alcohol use and sexual risk-taking, is important. However, this was not part of the remit for this guidance.

The PDG used various terms to describe the target group of an intervention. It did not feel that any term was preferable and used 'participants' and 'service users' interchangeably.

The PDG discussed the role of commercial companies in contributing to behaviour change and the potential contribution they could make to behaviour change interventions. Suppliers and manufacturers could, for example, provide (free of charge) useful data to aid understanding about behaviours such as alcohol use or eating patterns.

The PDG agreed that some of the recommendations were ambitious and may prove difficult to resource at local level. However, it was keen to set a 'gold standard' for service delivery as an aspirational target.

Evidence

The PDG did not think it was useful to look at specific behaviour change techniques in isolation. The Group agreed that single technique interventions may be effective for some people (or with some behaviours). However, it also noted that behaviour change often comes about because of a range of techniques working together (as well as other factors, such as context). The question is, which behaviour change techniques work most effectively together? The Group noted that theories of behaviour change may help determine which techniques should work synergistically.

The PDG noted that there was evidence of effectiveness for the behaviour change techniques recommended in relation to specific behaviours. However, the Group also noted, that the effectiveness of techniques across behaviours and populations was not always clear or necessarily supported by the evidence.

The behaviour change techniques taxonomy used in the evidence reviews (Michie et al. 2013) helped in discussions and in informing the evidence synthesis. However, the PDG had some concerns about the findings reported in review 2. This was due to the quality of reporting in intervention studies and the associated difficulty of coding behaviour change techniques on the basis of limited information. First, variations in reporting behaviour change techniques in the published data posed challenges when trying to provide consistent coding across interventions. Second, many tests were undertaken in the analysis in which behaviour change technique data were pooled across interventions for different behaviours and populations. This, combined with coding issues, could lead to the wrong conclusion concerning whether or not a technique is associated with behaviour change. Third, the potential moderating effect of other variables (such as mode of delivery and intervention intensity) was not addressed in the review analysis.

The authors of review 2 coded the behaviour change techniques used for 'usual care' and the 'control arm' in each study. This was to ensure all the behaviour change techniques used (in both the intervention and comparator) in any study were included in the meta-regression analysis. The accuracy of this coding was, however, dependent on the level of detail provided in published studies about the control arm. The PDG noted that, as with the reporting of interventions in published research, detail about control arms was often poor or missing. Generally there was not enough specific detail.

The PDG did not think recommendations could be based solely on the findings of the meta-regression analysis in evidence review 2. It noted that this review provided evidence of the effect sizes of behaviour change interventions and details of the behaviour change techniques used.

The PDG agreed that triangulation – looking for consistent effects across the different evidence considered by the group – would be appropriate. Consequently, if specific behaviour change techniques were evident in effective interventions in the evidence reviews and expert testimony, these findings were used as the basis for recommendations.

The PDG only made recommendations about behaviour change techniques that were identified using the triangulation process. Hence, a particular technique may not be recommended because of a lack of supporting evidence from more than one source, rather than due to evidence that it is not effective.

The evidence reviews that informed this guidance were structured around the specific taxonomy developed by Michie et al. (2013). While two-thirds of the possible behaviour change techniques defined in the taxonomy were identified in the included evidence, relatively few were identified often. This does not necessarily mean that techniques not mentioned were not used in the interventions. It may be that they are not reported or described in enough detail to be identified in published articles.

The lack of evidence on sexual behaviour in the commissioned reviews made it difficult for the PDG to make recommendations on these interventions.

The PDG noted that interventions aimed at changing people's alcohol use, eating patterns, physical activity, sexual behaviour and smoking are generally cost effective. The same is true for a number of other health behaviours that have been subjected to research. The Group also noted that there was little or no consistent association between the presence of any one behaviour change technique (or cluster of techniques) and an intervention being cost effective.

Developing policy and strategy

The PDG ensured the first recommendation highlights the need for an integrated programme of population, community, organisational and individual-level behaviour change interventions. It noted that interventions that target many levels simultaneously tend to be the most effective.

The PDG noted that it was important for all policy and strategy to be in line with the principles of proportionate universalism. This involves providing universal services and additional tailored support to meet the particular needs and choices of those who may find it difficult to use the services.

The PDG discussed whether practitioners and services should aim to change one behaviour at a time or multiple behaviours at once. It also discussed the best strategy to deal with multiple behaviours. Given the lack of evidence on the best approach, the Group made a recommendation for further research.

Commissioning quality-assured behaviour change programmes

The PDG was concerned that if private companies were commissioned to provide a behaviour change service they may not share data because of commercial interests. It noted the importance of data-sharing for the purposes of monitoring processes and outcomes.

The PDG noted the importance of ensuring all behaviour change interventions and programmes are conducted in an ethical manner. For example, this might involve ensuring participants in an intervention are fully informed of its content and how their data may be used. It might also involve ensuring national data protection and confidentiality policies are met.

The PDG considered that sustained changes in behaviour (that is, the maintenance of behaviour change) are vital to improve public health outcomes. It noted the need to plan for this at the start. The Group also noted that measurable changes in health at a population level may happen over the medium to long term, whereas changes in behaviour of individuals could be detected over shorter time periods.

The PDG noted the importance of long-term evaluation of behaviour change interventions and programmes. It also noted that, in reality, effectiveness is often not assessed beyond 6–12 weeks following an intervention.

The PDG noted that details of various study designs, their internal validity, and how to assess the quality of a study can be found in appendix D of Methods for the development of NICE public health guidance (third edition).

Designing behaviour change interventions

The PDG noted that the majority of published journal articles on behaviour change interventions do not provide enough detail to determine the techniques used in intervention and control groups. Where detail is provided, it may reflect the topic covered. For example, scientific studies on alcohol are based on a more standardised way of reporting than, say, scientific studies within the sexual health field. The Group discussed the need for manuals providing practical detail of the intervention techniques used, and for these to be made publicly available.

The PDG noted that a lack of detail in published journal articles on studies claiming to use motivational interviewing had affected the Group's ability to determine the behaviour techniques used. The PDG recognised that motivational interviewing is based on a clear set of principles and components. But as the articles did not specify which principles and components were used, the Group could not assume that motivational interviewing was used. This made it impossible to recommend this approach. It also added further support to the Group's recommendation that manuals should provide details of all the intervention components used.

The PDG acknowledged stakeholder concerns about intellectual property, copyright issues and the potentially inappropriate use of the information kept in manuals on behaviour change interventions. (For example, someone may use the information to set up their own commercial behaviour change intervention without having had the appropriate training needed to deliver it.) However, it was generally agreed that it was important to make manuals for all interventions publicly available, for example, as a condition of funding for projects in receipt of public monies.

As most journals now have web supplements, the PDG noted that it is possible to provide detailed reports of intervention designs, whatever the word limit of the main paper. The PDG discussed the fact that some journals only publish evaluations of interventions that come with publicly available manuals detailing the full protocols used. It welcomed this practice.

The recommendations contained in this guidance reflect the PDG's conclusions about intervention planning, based on the evidence considered. However, the PDG recognised that a number of other planning tools and resources – for example, 'intervention mapping' (Bartholomew et al. 2011) – could be systematically employed to enhance intervention design and effectiveness (see 4.31 below).

Intervention mapping aids collaborative planning by people from different professional backgrounds during intervention development. The approach proposes 6 intervention design stages:

  • Stage 1: A needs assessment determines what (if anything) needs to be changed for whom.

  • Stage 2: Primary and secondary intervention objectives are defined. This involves specifying the precise behaviour changes participants will be expected to make.

  • Stage 3: Designers identify the underlying evidence-based techniques that maintain current (unwanted) behaviour patterns and may generate the specified changes.

  • Stage 4: Practical ways of delivering these techniques are developed and integrated into the intervention.

  • Stage 5: How the intervention will be used or delivered in everyday contexts is considered.

  • Stage 6: Evaluation to assess whether the intervention changed specified behaviours in context.

    These stages are iterative in that, for example, anticipation of how the intervention will be used or delivered may lead to a change in design and a return to stage 4. Similarly, when the exact behaviour changes are defined in stage 2, these may need to be evaluated. The result is an intervention 'map' of matrices and plans that guide the design, implementation and evaluation of an intervention.

The PDG noted the importance of not just the content of an intervention, but who is delivering it (and their core competencies), to whom, how and where.

The PDG discussed the importance of making sure all key components of a given intervention are adopted so that they have high intervention fidelity and are sustainable.

The PDG agreed that a behaviour change taxonomy for designing interventions was a useful tool. However, the Group was clear that the inclusion of a behaviour change technique in a taxonomy did not necessarily mean there was a strong evidence base for that technique.

Delivery

Evidence showed that behaviour change interventions by GPs and other medical staff can be effective. However, the PDG felt that a focus solely on such interventions may lead to a widening in health inequalities, because people from the most vulnerable groups often do not use primary care services. The Group did not want to exclude such interventions, rather it raised the need to find alternative ways of reaching vulnerable groups. Members agreed that understanding how people come into contact with, and access, services was key to the design of behaviour change interventions.

The PDG noted that details of validated tools and measures for assessing behaviour can be found in academic publications. The Group also noted that specific assessment tools have been recommended by NICE. For example, tools for assessing alcohol use are recommended in the NICE guideline on alcohol-use disorders: prevention.

The PDG agreed that although information is usually a necessary precursor to behaviour change, information alone is not always sufficient to influence behaviour.

The PDG noted that social, economic and cultural contexts can have an impact on behaviour. Although a sense of connection and belonging at school, within the family or community promotes resilience, unhealthy behaviour can also be embedded in social processes and patterns. The Group also noted that diverse health outcomes can be established early in life.

Behaviour change techniques

Recommendations were made to include goals and planning and feedback and monitoring techniques in behaviour change interventions. This was based on a cross-examination of behaviour change techniques identified in expert paper 14, the evidence reviews and the NICE guideline on behaviour change: general approaches. (Triangulation techniques were used.) The process indicated that the techniques would be effective as part of interventions on alcohol, diet, physical activity and smoking. These behaviour change techniques are described in detail in Michie et al. (2013).

Time constraints meant it was not possible to review additional evidence on sexual health interventions and behaviour change techniques. However, the PDG noted that in existing NICE guidance, social support was frequently used in effective interventions for all behaviours (alcohol, diet, physical activity, sexual behaviour and smoking).

The PDG noted that social support provided by friends, family and associates could help to create an environment in which people felt able to make changes. However, members also noted that, if not managed effectively, social support provided by non-professionals (such as family members) could sometimes lead to an unhealthy co‑dependency, bullying, manipulation or other negative behaviour.

The PDG noted that principle 4 (on individuals) in the NICE guideline on behaviour change: general approaches recommended specific behaviour change techniques. The Group agreed that, for consistency of approach, these would be 'coded' using the taxonomy applied in the commissioned evidence reviews for this guidance (Michie et al. 2013). This taxonomy identified the following groups of behaviour change techniques: goals and planning, feedback and monitoring, social support, natural consequences, comparison of behaviour, repetition and substitution and antecedents.

The PDG noted that interventions are unlikely to be effective if providers are not properly trained or the setting is not appropriate. The Group also noted that some behaviour change techniques, such as self-monitoring, might be difficult for some people.

The PDG was aware of a growing interest in the use of new technology, including phone and tablet apps, to deliver behaviour change interventions. The Group noted that the evidence is mixed and there have not been many formal evaluations of its effectiveness. But it also noted that evidence to support the use of technology is encouraging (see recommendation 7).

Training

The PDG noted the importance of training. For example, the Group discussed the fact that if one person successfully trained 100 practitioners they, in turn, could help 10,000 people, and the knock-on effects would be huge.

The PDG discussed the importance of communication skills when providing behaviour change interventions. In particular, it noted the importance of knowing how to initiate a conversation, develop rapport and communicate information effectively. Communication skills include reflective listening, the use of open ended questions and affirmation skills.

The PDG did not discuss the accreditation of training. This may be an area where future guidance is needed.

The PDG was concerned that training programmes still describe the stages of change model (also known as the transtheoretical model) as a theoretical basis for behaviour change interventions. The PDG wanted to highlight that, although it may help practitioners and service users to understand the experience of behaviour change, it is not a theory that is able to accurately explain and predict such change. It was noted that interventions based on this model alone have not, according to the evidence reviewed here, demonstrated effectiveness.

The PDG was clear that being trained to deliver one behaviour change intervention does not necessarily mean that a practitioner is then competent to deliver other behaviour change interventions. The Group was aware of the danger that practitioners and service users may assume their competency extends further than it actually does.

The PDG noted that behaviour change training is a behaviour change intervention in its own right.

Evaluation

The PDG noted that this guidance is not intended to make recommendations on how to undertake research in this area.

The PDG noted that well-conducted evaluation studies and randomised controlled trials – with minimal bias – give the best quality evidence. Anecdotal evidence and smaller or poorly conducted studies are much less reliable and the Group agreed that it was best not to use these as a basis for investment decisions. The Group also noted that NICE, NHS Evidence and the Cochrane Collaboration provide guidance to help identify behaviour change interventions or programmes to invest in for a particular topic, population or setting.

The PDG noted that qualitative, as well as quantitative, measures are important when trying to understand why something does or does not work – and under what circumstances. They can also help to identify any improvements or changes that need to be made.

The PDG noted that the setting where an intervention is delivered and the person delivering it may be the two main factors that make an intervention effective (or ineffective).

The PDG discussed the meaning of independent evaluation. The Group was clear that this was not synonymous with external evaluation. Rather, it could be carried out inside an organisation as long as it was not conducted by those actually involved in designing or delivering the intervention.

National support

The PDG noted that local organisations may need support to help them decide on the most appropriate behaviour change interventions and training to commission or provide in their area. Members discussed how a unified national approach might achieve this, with organisations and research funders working together to ensure appropriate data collection, evaluation and dissemination of evidence. The PDG also noted that some organisations, for example Public Health England and NICE (through NICE evidence services) are already working towards this goal.

Choice architecture

On the basis of current evidence, the PDG felt that it would be premature to make any recommendations on the use of choice architecture interventions (see expert paper 8). As a result, only research recommendations were made on this.

A scoping review of the evidence base for choice architecture interventions targeting healthy behaviour indicated that the majority of evidence involved diet (see expert paper 8). However, in the absence of a full systematic review, the PDG questioned whether such interventions did lead to a healthy diet.

The PDG noted that, in the context of choice architecture, 'doing nothing' is not a neutral approach, because this simply maintains the status quo. And the status quo may, for example, be an 'obesogenic environment' constructed by commercial interests.

The PDG recognised that choice architecture interventions may appeal to people working in a local authority setting. The reasons are twofold. First, this type of behaviour change intervention may be perceived to be relatively low cost. Second, it has the potential to reach a relatively large number of people. However, there is only limited evidence on how effective these interventions are at changing health-related behaviour. The PDG agreed that anyone wishing to commission or provide such an intervention as part of a behaviour change service should be aware of this lack of evidence. The Group agreed that choice architecture interventions, if used, needed to be subject to independent evaluation.

Although the PDG was not able to make recommendations on choice architecture interventions, members noted that a further evidence synthesis on this approach is due to be published soon. The PDG advised that if this synthesis is published prior to the routine update of this guidance, the update should be bought forward.

  • National Institute for Health and Care Excellence (NICE)