Behaviour change techniques: the development and evaluation of a taxonomic method for reporting and describing behaviour change interventions (a suite of five studies involving consensus methods, randomised controlled trials and analysis of qualitative data)

Article history

This issue of the Health Technology Assessment journal series contains a project commissioned/managed by the Methodology research programme (MRP). The Medical Research Council (MRC) is working with NIHR to deliver the single joint health strategy and the MRP was launched in 2008 as part of the delivery model. MRC is lead funding partner for MRP and part of this programme is the joint MRC–NIHR funding panel ‘The Methodology Research Programme Panel’.

To strengthen the evidence base for health research, the MRP oversees and implements the evolving strategy for high-quality methodological research. In addition to the MRC and NIHR funding partners, the MRP takes into account the needs of other stakeholders including the devolved administrations, industry R&D, and regulatory/advisory agencies and other public bodies. The MRP funds investigator-led and needs-led research proposals from across the UK. In addition to the standard MRC and RCUK terms and conditions, projects commissioned/managed by the MRP are expected to provide a detailed report on the research findings and may publish the findings in the HTA journal, if supported by NIHR funds.

The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The HTA editors and publisher have tried to ensure the accuracy of the authors’ report and would like to thank the reviewers for their constructive comments on the draft document. However, they do not accept liability for damages or losses arising from material published in this report.

Declared competing interests of authors

Charles Abraham acknowledges funding from the UK National Institute for Health Research and the Peninsula Collaboration for Leadership in Applied Health Research and Care.

Copyright statement

© Queen’s Printer and Controller of HMSO 2015. This work was produced by Michie et al. under the terms of a commissioning contract issued by the Secretary of State for Health. This issue may be freely reproduced for the purposes of private research and study and extracts (or indeed, the full report) may be included in professional journals provided that suitable acknowledgement is made and the reproduction is not associated with any form of advertising. Applications for commercial reproduction should be addressed to: NIHR Journals Library, National Institute for Health Research, Evaluation, Trials and Studies Coordinating Centre, Alpha House, University of Southampton Science Park, Southampton SO16 7NS, UK.

Behaviour change interventions: purpose and reporting

Preventable behaviours, such as smoking, physical inactivity, eating unhealthy diets and excessive alcohol consumption, have been identified as leading causes of morbidity and mortality. 1–4 Behaviour change interventions (BCIs) are typically complex, involving many interacting components. 5 This can make them challenging to replicate in research, to implement in practical settings and to synthesise in systematic literature reviews.

Behaviour change interventions are ‘Coordinated sets of activities designed to change specified behaviour patterns’. 6 The development, implementation and evaluation of effective BCIs are fundamental for advancing behavioural science and its application. 7,8 However, both science and practice depend on having a good understanding of the nature and content of interventions. This includes knowing what was delivered in the intervention [i.e. the ‘active ingredients’ or behaviour change techniques (BCTs)] and how it was delivered (i.e. who delivered, to whom, how often, for how long, in what format, and in what context). 9–11 Poorly described interventions in research protocols and published reports mean that the precise content of interventions is difficult to establish, with the possibility of the same labels (e.g. behavioural counselling) meaning different things to different researchers/implementers and different labels being used to describe the same BCTs. As a result, the content delivered in practice often deviates from that specified in the intervention protocol12 presenting a barrier to scientific progress and effective translation.

Guidance documents have been published aimed at improving methods of specifying and reporting interventions. For example, Consolidated Standards of Reporting Trials guidelines for reporting randomised controlled trials (RCTs)13 advise researchers to report the ‘precise details’ of the intervention as ‘actually administered’, and the Transparent Reporting of Evaluations with Nonrandomized Designs statement for reporting non-randomised trials14 emphasises the reporting of content and context and full description of comparison and intervention conditions. The UK Medical Research Council (MRC) guidance for developing and evaluating complex interventions calls for the specification of the active ingredients as a necessary step for investigating how interventions exert their effect and, therefore, for designing more effective interventions and applying them appropriately across target population group and setting. 5 The Workgroup for Intervention Development and Evaluation Research group has had some success in encouraging journal editors to ensure that transparent and accessible intervention descriptions are available before publication of intervention outcomes. 15 The Template for Intervention Description and Replication (TIDieR)16 provides a checklist of the minimum data required to report interventions, including surgical, pharmacological, psychotherapeutic as well as behavioural interventions. Although progress has been made in improving how intervention content is reported, if descriptions of BCIs are to be communicated effectively and successfully replicated, a shared and standardised method of classifying intervention content is needed. 17

Behaviour change techniques

A BCT is defined as an observable and replicable component designed to change behaviour. It is the smallest component compatible with retaining the postulated active ingredients and can be used alone or in combination with other BCTs. To enable interventions to be evaluated and effective interventions (i.e. those which bring about the desired change in the target behaviour or behaviours) to be implemented, a BCT should be well specified. BCTs are descriptors and vary in the extent to which they have been empirically investigated and the extent to which they bring about the desired change to behaviour(s) in different situations. BCT definitions used for coding have to be practical, non-overlapping and useful in the reliable reporting of interventions.

Specification of interventions according to component BCTs is beneficial for conducting primary research, implementing effective interventions and for conducting evidence syntheses. A comprehensive list of BCTs facilitates primary research, as intervention developers can draw on a wider range of BCTs than is likely to be considered without such a list. Specification of intervention and control conditions using BCTs can increase accurate replication of interventions found to be efficacious in RCTs. BCT methodology is also useful in assessing the fidelity of implementation of interventions. For systematic reviewers, BCTs provide a reliable method for extracting and coding information about intervention content. Reviewers can identify and synthesise discrete, replicable, potentially active ingredients associated with efficacy and multivariate statistical analysis can then be used to identify BCTs and BCT combinations associated with efficacy. By linking BCTs with theories of behaviour change, researchers and reviewers can investigate possible effect modifiers and/or mechanisms of action. There are some intervention components that can be thought of as ‘modifier BCTs’ in that they add value to BCTs but do not in themselves change behaviour, for example, tailoring, giving choice and homework tasks. Specifying intervention content with this degree of precision helps to maximise scientific as well as practical benefits of research investment into the development and evaluation of complex interventions.

Behaviour change technique taxonomies

To provide a more rigorous methodology for characterising intervention content, methods have been developed for specifying the potentially active ingredients in terms of BCTs. 18–20 Abraham and Michie’s20 taxonomy of 22 BCTs and four BCT packages observed in BCIs was demonstrated to have good intercoder reliability (i.e. the extent to which coders agreed on the presence/absence of BCTs) among the taxonomy developers and trained coders across 221 intervention descriptions in papers and manuals. This, and subsequent taxonomies developed for specific behavioural domains, have been widely used internationally to report interventions, synthesise evidence21–24 and design interventions. 25,26 BCT taxonomies have been developed in relation to smoking,27 physical activity and healthy eating,28 excessive alcohol use29 and condom use. 30 Taxonomies of BCTs have also been used to assess the extent to which published reports reflect intervention protocols12 and to assess fidelity of delivery. 31 They have enabled the specification of professional competences for delivering BCTs32,33 and as a basis for a national training programme. 34,35 Guidance has also been developed for incorporating BCTs in text-based interventions. 36

Previous classification systems have either been in the form of an unstructured list or have been linked to, or structured, according to theories and/or theoretical mechanisms19,20,32,33 judged to be the most appropriate by the authors.

A hierarchically structured list provides the advantage of making it more coherent to, and useable by, those applying it. 37 There are at least five potential benefits of developing a cross-domain, hierarchically structured and internationally supported taxonomy for specifying intervention content:

1. To promote the accurate replication of interventions and control conditions in comparative efficacy research, a key activity in accumulating scientific knowledge and investigating generalisability across behaviours, populations and settings.

2. To specify intervention content to facilitate faithful implementation of intervention protocols in research and, in practice, of interventions found to be effective.

3. To extract and synthesise information about intervention content in systematic reviews. BCT taxonomies, combined with the statistical technique of meta-regression or classification and regression tree (CART), have allowed reviewers to synthesise evidence from complex, heterogeneous interventions to identify effective component BCTs and BCT combinations. 21,23,24,38,39

4. To draw on a comprehensive list of BCTs in developing interventions (rather than relying on the limited set that can be brought to mind).

5. To investigate possible mechanisms of action by linking BCTs with theories of behaviour change and component theoretical constructs. 8,19,21,23

In this monograph, we present the development and evaluation of Behaviour Change Technique Taxonomy version 1 (BCTTv1): a cross-domain, hierarchically structured taxonomy based on interdisciplinary consensus as a method for the accurate and reliable reporting of interventions to change behaviour. 40

Abstract

Objectives: To develop an extensive, consensually agreed taxonomy of BCTs used in BCIs.

Methods: In a Delphi-type exercise, 14 experts rated labels and definitions of 124 BCTs from six published classification systems. 20,23,27,29,33,36 The resulting list was refined based on the feedback from group discussions of 16 members of the International Advisory Board (IAB).

Results: This resulting BCTTv1 comprised 93 distinct, non-redundant and non-overlapping BCTs.

Conclusions: BCTTv1 offers a step change in methods for specifying interventions using shared concepts and language. When sufficient data have been collected about its implementation and an international, interdisciplinary consortium established, v1 will be reviewed and v2 released as and when the need is judged by consensus.

Introduction

Literature reviews find that even essential elements of interventions are frequently omitted from intervention descriptions; an analysis of trials and reviews found that 67% of pharmacological intervention descriptions were adequate, compared with only 29% of non-pharmacological intervention descriptions. 41 While this occurs in descriptions of all types of intervention, it is an even more common problem in BCIs. Titles and abstracts of published interventions (i.e. the materials screened for inclusion in systematic reviews) have been found to mention the active components of the intervention in only 56% of published descriptions compared with over 90% in pharmacological interventions. 42 The TIDieR checklist and guide for reporting all types of interventions was developed using consensus methods with international participants from several disciplines and proposes a minimum set of information: brief name, why (rationale), what materials, what procedure, who provided, how, where, when and how much, tailoring, changes, how well monitored and how well delivered. 16 However, for BCIs, further information is required to specify the active ingredients, that is ‘components within an intervention that can be specifically linked to its effect on outcomes such that, if they were omitted, the intervention would be ineffective’. 42

The content, or active components, of BCIs are often described in intervention protocols and published reports with different labels (e.g. ‘self-monitoring’ may be labelled ‘daily diaries’) and the same labels may be applied to different techniques (e.g. ‘behavioural counselling’ may involve ‘educating patients’ or ‘feedback, self-monitoring, and reinforcement’). 43 This may lead to uncertainty, confusion and difficulties in determining the efficacy of specific change approaches; for example, Morton et al. 44 had considerable difficulty in identifying the necessary components of motivational interviewing and as a result found it difficult to synthesise evidence of efficacy. Further, behavioural medicine researchers and practitioners have reported low confidence in their ability to replicate highly effective behavioural interventions for type 2 diabetes mellitus prevention. 45 The absence of standardised definitions and labels for intervention components means that systematic reviewers develop their own systems for classifying behavioural interventions and synthesising study findings [e.g. Hardeman et al. ,18 Albarracín et al. ,19 Mischel (Presidential address given at the Association for Psychological Science Annual Convention, Washington, DC; 2012) and West et al. 46]. This proliferation of systems leads to duplication of effort and undermines the potential to accumulate evidence across reviews. It also points to the urgent need for consensus. Consequently, the UK MRC guidance5 for developing and evaluating complex interventions called for improved methods of specifying and reporting intervention content in order to address the problems of lack of consistency and consensus.

A method developed for this purpose is the reliable specification of interventions in terms of BCTs. 20 Previous classification systems have mainly been developed for particular behavioural domains (e.g. physical activity, smoking, alcohol consumption or safer sex). Abraham and Michie20 developed the first cross-behaviour BCT taxonomy, building on previous intervention content analyses. 18,19 The taxonomy demonstrated reliability in identifying 22 BCTs (e.g. ‘self-monitoring’) and four BCT packages (e.g. ‘relapse prevention’). Identifying the presence of BCTs in intervention descriptions included in systematic reviews and national datasets of outcomes has allowed the identification of BCTs associated with effective interventions. 46,47 Effective BCTs have been identified for interventions to increase physical activity and healthy eating,23,48 and to support smoking cessation,27,46 reduce excessive alcohol consumption,29 prevent sexually transmitted infections19,30 and change professional behaviour. 49

Although the subsequent development of classification systems of defined and reliably identifiable BCTs has been accompanied by a progressive increase in their comprehensiveness and clarity, this work has been conducted by only a few research groups, but with each developing their own methodology. For this method to maximise scientific advance, collaborative work was needed to develop agreed labels and definitions and reliable procedures for their identification and application across behaviours, disciplines and countries. Therefore, the aim of study 1 was to develop a taxonomy that comprises an extensive list of clearly labelled, well defined BCTs that (1) are proposed as the active components of BCIs, (2) are distinct (non-overlapping, non-redundant) and precise, (3) can be used with confidence to describe interventions, and (4) have a breadth of international and disciplinary agreement.

Method

This study is also published as Michie et al. 40

Results

All data tables for this chapter are reported elsewhere, please refer to Michie et al. 40 The evolution of the taxonomy at the different steps of the procedure is summarised in Michie et al. 40

Step 5: feedback from study team members

Further refinement of labels and definitions by study team members resulted in a list of 93 clearly defined, non-redundant BCTs (see Appendix 2, Table 19). The full evolution of the taxonomy across the five steps is shown in Figure 1.

FIGURE 1.

Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow diagram to show evolution of the taxonomy: labels and definitions of BCTs modified.

Discussion

An extensive list of 93 distinct, non-redundant BCTs was developed with labels and definitions refined to capture the smallest components compatible with retaining the proposed active ingredients with the minimum of overlap: BCTTv1 (see Appendix 2, Table 19). Development comprised a series of consensus exercises involving 35 experts in delivering and/or designing BCIs. These experts were drawn from a variety of disciplines including psychology, behavioural medicine and health promotion, and from seven countries (the USA, Canada, Australia, UK, the Netherlands, Finland and Germany). Therefore, the resulting BCTs have relevance among experts from varied behavioural domains, disciplines and countries, and potential relevance to the populations from which they were drawn. Evidence is already emerging to suggest that some BCTs from BCTTv1 occur more frequently than others in descriptions of BCIs. 40 These BCTs are marked with an asterisk in Table 19 (see Appendix 2).

The extent to which BCTTv1 is applicable without adaptation across behaviours, disciplines and countries is an important question for future monitoring and research. The process of building a shareable consensus language and methodology is necessarily collaborative and will be an ongoing cumulative and iterative process, involving an international network of advisors and collaborators. 50 The balance of stability to allow accumulation of knowledge and development to incorporate significant bodies of new knowledge and experience means that classificatory systems are updated at strategic intervals. 37 Examples where this has happened are Linnaeus’s classification of plants and systems based on consensus such as the Diagnostic and Statistical Manual of Mental Disorders-Fourth Edition54 or the International Classification of Diseases and Related Health Problems, Tenth Edition. 55

There was no prior agreed methodology for this work and there are limitations to the methods we have used. The purpose of the Delphi exercise was to develop a prototype taxonomy on which to build. It was the first in a series of exercises adapted to develop the taxonomy. Our Delphi-type methods involved 14 individuals, an appropriate number for these methods,51 but a number that makes the choice of participants important. We attempted to ensure adequate coverage of behaviour change experts. Although we had some diversity of expertise, we acknowledge the predominance of European experts from a psychological background within our sample. At various stages we made arbitrary decisions such as the cut-offs for amending BCT labels and descriptions. In the absence of agreed standards for such decisions, we were guided by the urgent need to develop an initial taxonomy that was fit for purpose and would therefore form a basis for future development. Our amendments of the BCT labels and definitions also depend on the expertise available and, therefore, we based our amendments on a wide range of inputs – the data we collected from Delphi participants and coders, expert modification, international advice and lay user improvements.

The BCTTv1 encompasses a greater number of BCTs than previous taxonomies. Therefore, it requires structure to facilitate recall and access to the BCTs, and thus increase speed and accuracy of use. A true, that is hierarchically structured, taxonomy provides the advantage of making it more coherent to, and useable by, those applying it. 37 As the number of identified BCTs has increased, so also has the need for such a structure, to improve the usability of the taxonomy.

Simple, reliable grouping structures have previously been used by three groups of authors. Dixon and Johnston33 grouped BCTs according to ‘routes to behaviour change’, ‘motivation’, ‘action’ and ‘prompts/cues’; Michie et al. 27 grouped according to ‘function’ in changing behaviour, ‘motivation’, ‘self-regulation capacity/skills’, ‘adjuvant’ and ‘interaction’; and Abraham et al. 30 grouped according to ‘change target’, that is ‘knowledge’, ‘awareness of own behaviour’, ‘attitudes’, ‘social norms’, ‘self-efficacy’, ‘intention formation’, ‘action control’, ‘behavioural maintenance’ and ‘change facilitators’. In order to achieve our aim of a structured taxonomy that is acceptable and useable over diverse disciplines and theoretical orientations, we used a basic method of grouping that does not depend on a theoretical structure. In the next study, we therefore adopted an empirical, ‘bottom-up’ method to developing a consensus of BCT groupings.

Abstract

Objectives: To enhance understanding and use of BCTs by investigating the hierarchical structure of BCTTv1.

Methods: Participants grouped BCTs according to similarity of active ingredients in an open-sort task. This structure was examined for higher-order groupings using a dendrogram derived from hierarchical cluster analysis. This ‘bottom-up’ sort method was compared with a theory-based ‘top-down’ method in which 18 experts sorted BCTs into 14 theoretical domains in a closed-sort task. Discriminant content validity (DCV) was used to identify groupings and chi-squared tests, and Pearson’s residual values were used to examine the overlap between groupings.

Results: Participants created an average of 15.11 groupings (SD = 6.11 groupings, range 5–24 groupings). BCTs relating to ‘Reward and Punishment’ and ‘Cues and Cue Responses’ were perceived as markedly different from other BCTs. Fifty nine of the BCTs were reliably allocated to 12 of the 14 theoretical domains; 47 were significant and 12 were of borderline significance. Two domains had no BCTs significantly assigned to them. An additional grouping of ‘No Domain’ was included to represent these cases. There was a significant association between the 16 ‘bottom-up’ groupings and the 13 ‘top-down’ groupings (χ² = 437.80; p < 0.001). Thirty-six of the 208 ‘bottom-up’ × ‘top–down’ pairings (i.e. 16 × 13) showed greater overlap than expected by chance. However, only six combinations achieved satisfactory evidence of similarity.

Conclusions: The moderate overlap between the groupings indicates some tendency to implicitly conceptualise BCTs in terms of the same theoretical domains. Understanding the nature of the overlap will aid the conceptualisation of BCTs in terms of theory and application. Further research into different methods of developing a hierarchical taxonomic structure of BCTs for international, interdisciplinary work is now required.

Introduction

Study 1 presented the synthesis of existing BCT taxonomies into a single comprehensive, cross-context, overarching BCT taxonomy: BCTTv1. 8,17,40 Previous BCTs groupings have been based on judgements made by the study author. 20,27,32,33 Given the 93 items of BCTTv1, it is necessary to group the BCTs to make the taxonomy more memorable and useable. To achieve this, we need an agreed method for identifying links between particular BCTs and theoretical constructs.

Two methods were investigated: (1) a ‘bottom-up’ linkage and (2) a ‘top-down’, theoretically guided linkage. The ‘bottom-up’ approach allows each respondent to propose linkages inductively and then identifies which linkages are common across respondents. It makes no assumptions about underlying theory and, therefore, the results should be accessible to users from diverse theoretical and disciplinary backgrounds. Nevertheless, it may reflect commonalities in theoretical approaches. The ‘top-down’ approach prompts each respondent to deduce theoretical linkages based on underlying theory. Common linkages across respondents are then identified.

Theories of behaviour change summarise what is known about the mechanisms of behaviour change and the conditions in which behaviour change is most likely to occur. 8 The importance of understanding the theoretical underpinnings of BCTs has been highlighted in previous research. 10,21,56,57 However, a recent meta-analysis found that BCIs are often not designed on a clear theoretical foundation; for example, only 22.5% of 235 implementation studies explicitly used theories of behaviour change58,59 and the majority of those doing so gave no clear explanation for why the selected theories had been used. Therefore, there is a clear need for improving methods for applying theory to intervention design to increase our understanding of how BCTs exert their influences. Grouping BCTs by theory would help guide understanding of the functional relationships between BCTs, the underlying mechanisms through which they exert their effects and the most effective ways in which BCTs can be applied.

In light of the 93 BCTs of BCTTv1 and the large number of behaviour change theories and component constructs,60 grouping by individual theoretical constructs is impractical. An alternative is to group by broader domains of theoretical constructs (e.g. knowledge, skills, etc.) as has been done in the theoretical domains framework (TDF). 45,61 The TDF is an integrative framework of theoretical constructs of behaviour change that was originally developed by 18 psychological theorists in collaboration with 16 health service researchers and 30 health psychologists. 45 It was developed to make theory more accessible to, and usable by, a range of disciplines and theoretical orientations. The first version of the TDF contains 12 theoretical domains synthesised from 128 theoretical constructs related to behaviour change; the validated version61 suggested minor modifications, with 14 domains.

The TDF has been used by research teams across many countries and health-care systems to investigate implementation problems and inform interventions to change professional practice. 62–72 The TDF was validated using two sort tasks,61 producing a refined TDF containing 87 theoretical constructs relevant to behaviour change categorised across 14 domains: knowledge, skills, social/professional role and identity, beliefs about capabilities, optimism, beliefs about consequences, reinforcement, intentions, goals, memory, attention and decision processes, environmental context and resources, social influences, and emotion and behavioural regulation. A previous study linking 35 BCTs to 11 theoretical domains from the original TDF showed good reliability across four researchers, with 71% agreement over the 385 possible links. 43 Building on this work, we aim to link BCTs from the more comprehensive BCTTv1 to the refined TDF, using a larger number of experts in behaviour change.

Study 2 aimed to:

• investigate the hierarchical structure of the groupings of the taxonomy, which were obtained from an inductive ‘bottom-up’ method, using quantitative clustering methods

• identify to what extent the taxonomy can be reliably grouped using a deductive, ‘top-down’ theory-based method into the 14 theory-based domains of the revised TDF

• examine similarities and differences in the groupings that emerged using these two methods of developing a hierarchical structure.

Method

This study is also published as Cane et al. 73 Some text has been reproduced from Cane et al. 73 © 2014 The Authors. British Journal of Health Psychology published by John Wiley & Sons Ltd on behalf of the British Psychological Society. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

Analysis

To analyse open-sort data, a binary dissimilarity matrix containing all possible BCT × BCT combinations was produced for each participant, for whom a score of one indicated BCTs that were not sorted into the same grouping, and a score of 0 indicated items that were sorted into the same grouping. Individual matrices were aggregated to produce a single dissimilarity matrix, which could be used to identify the optimal grouping of BCTs using cluster analysis. Using hierarchical cluster analysis, the optimal number of groupings (2–20) were examined for suitability using measures of internal validity (Dunn’s Index) and stability (figure of merit). 79 Bootstrap methods were used in conjunction with the hierarchical cluster analysis, whereby data were resampled 10,000 times, to identify which groupings were strongly supported by the data. The approximately unbiased p-values yielded by this method indicated the extent to which groupings were strongly supported by the data with higher approximately unbiased values (e.g. 95%) indicating stronger support for the grouping. 80 The words and phrases used in the labels given by participants were analysed to identify any common themes and to help identify appropriate labels for the groupings. For each grouping, labels were created based on their content and, when applicable, based on the frequency of word labels given by participants. After the labels were assigned to relevant groupings, the fully labelled groupings with the word frequency analysis were sent out to a subset of five of the original participants for refinement.

To analyse closed-sort data, mean confidence ratings for each BCT × domain pairing were calculated and analysed using DCV methods. 74,77 BCT × domain pairings that had no confidence rating from individual participants (i.e. BCT was not allocated to that domain by that participant) were scored zero and entered into the mean score for that pairing. A series of one-sample t-tests compared the mean confidence ratings for the assignment of BCTs to a value of zero. This established the extent to which BCTs were related to each domain. In cases for which no experts allocated a BCT to a specific domain (i.e. all scores for a BCT × domain pairing were zero) the BCT × domain pairings was excluded from t-test analyses.

The BCTs were considered to be reliably allocated to a domain if their mean confidence ratings were significantly greater than zero (p < 0.05) after Hochberg’s correction81 [applied using the p.adjust function in R version 3.0.1 (The R Foundation for Statistical Computing, Vienna, Austria)]. This was used to control for the family-wise error rate, given the large number of tests used, and provided a suitable criterion for inclusion and exclusion of BCTs to a particular domain, over and above the use of a subjective cut-off value. Hochberg’s correction also provides a conservative p-value that makes it less likely that a BCT × domain pairing achieving low confidence ratings across the majority of participants will achieve significance. The agreement of BCT allocation across participants was analysed using a two-way intraclass correlation coefficient (ICC) within each domain.

To identify any overlap between groupings, two types of comparisons were made between the ‘bottom-up’ groupings and the ‘top-down’ TDF derived groupings – comparison between the theoretically derived ‘top-down’ groupings and, (1) the higher order strategy groupings found in the ‘bottom-up’ sort task, and (2) the final groupings of the ‘bottom-up’ sort task. To test the possibility of overlap between groupings derived from using ‘bottom-up’ and ‘top-down’ methods, Pearson’s chi-squared test was adopted. To adjust for potential inaccuracy of the p-value estimation (resulting from the number of cells that had expected frequencies < 1), Monte Carlo simulation (using 2000 replications) was used. Pearson’s residual values [(observed – expected)/sqrt (expected)] were used to quantify the extent of overlap between individual BCTTv1 grouping × TDF domain pairings resulting from the ‘bottom-up’ and ‘top-down’ methods. Positive values indicate that the observed overlap in BCT assignment between the BCT taxonomy and TDF domain pairings is greater than expected by chance whereas negative values indicate that it is less than expected.

Results

Participants for the closed-sort task reported moderately high levels of expertise in behaviour change theory (mean = 3.17, SD = 0.71) and in delivering BCIs (mean = 2.17, SD = 1.38) as measured on 5-point scales (scores are reversed so a higher score indicates more experience). This was not significantly different from the level of expertise reported by participants in the open-sort task: behaviour change theory, mean = 3.00, SD = 0.88, t(34) = 0.64; p > 0.10; BCIs, mean = 2.42, SD = 0.96, t(34) = 0.63; p > 0.10. Although the age of participants did not differ significantly between the two sort tasks [t(34) = 0.77; p > 0.10] the number of female and male participants did [χ²(1) = 4.33; p < 0.05] as did the country of residence (χ² = 20.76; p < 0.001) (Monte-Carlo simulation using 2000 replicates was used to compute the p-value given that a number of the expected cell values were < 1). This was an artefact of the selection process as there was no duplication of participants across the two sort tasks.

Identifying overlap between the ‘bottom-up’ and ‘top-down’ groupings

The chi-squared analyses used for the grouping comparisons did not allow us to include domains in which no BCTs were assigned (i.e. not linked to domains through the DCV process); therefore, the domains of memory, attention and decision processes, and social/professional identity were excluded from these analyses. An additional grouping of ‘No domain’ was included in the ‘top-down’ groupings and represented cases for which BCTs included in BCTTv1 were not assigned to any TDF domain. Therefore, the chi-squared analysis was conducted first on 91 (7 × 13) possible pairings for the seven higher-order ‘bottom-up’ sorting strategy groupings and the 13 ‘top-down’ groupings, and second on 208 (16 × 13) possible pairings derived from the original 16 BCTTv1 ‘bottom-up’ groupings and the 13 ‘top-down’ groupings.

Comparison of the BCT groupings derived from the higher-order ‘bottom-up’ sorting strategies, shown in the dendrogram and the ‘top-down’ TDF derived groupings (see Table 3) revealed a significant association (χ² = 236.13; p < 0.001). Figure 2 shows the level of overlap between each of the grouping × TDF pairings within each cell (Pearson’s residual value range –2.10 to 6.61).

FIGURE 2.

Pearson’s residual values for the association between BCT allocation to ‘top-down’ theoretical domain groupings and the ‘bottom-up’ higher-order hierarchical split groupings. Note: ‘No domain’ indicates BCTs contained within a ‘bottom-up’ higher-order grouping not assigned to any TDF domain. Reproduced from Cane et al. 73 © 2014 The Authors. British Journal of Health Psychology published by John Wiley & Sons Ltd on behalf of the British Psychological Society. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

Twenty one of the 91 ‘bottom-up’ and ‘top-down’ TDF domain combinations showed a greater than expected overlap with positive Pearson’s residual values (Table 4). Only two combinations achieved Pearson’s residual values > 5, which were: ‘grouping 1′ with ‘reinforcement’ (Pearson’s residual value = 6.61) and ‘grouping 5′ with ‘social influences’ (Pearson’s residual value = 5.04).

TABLE 4 - Taxonomy grouping (‘bottom-up’) and TDF domain (‘top-down’) combinations achieving positive Pearson’s residual values for similarities in the assignment of BCTs

Reproduced from Cane et al. ⁷³ © 2014 The Authors. British Journal of Health Psychology published by John Wiley & Sons Ltd on behalf of the British Psychological Society. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

Taxonomy grouping	TDF domain	Pearson’s residual value
Repetition and substitution	Skills	6.66
Goals and planning	Goals	6.41
Covert learning	Beliefs about consequences	5.76
Self-belief	Beliefs about capabilities	5.70
Scheduled consequences	Reinforcement	5.22
Antecedents	Environmental context and resources	5.20
Comparison of behaviour	Social influences	4.96
Social support	Social influences	4.14
Reward and threat	Reinforcement	4.10
Goals and planning	Intentions	4.05
Feedback and monitoring	Behavioural regulation	4.03
Self-belief	Optimism	4.03
Feedback and monitoring	Knowledge	3.80
Comparison of outcomes	Beliefs about consequences	3.68
Natural consequences	Emotion	3.05
Associations	Environmental context and resources	2.35
Regulation	Emotion	1.97
Shaping knowledge	Knowledge	1.97
Social support	Emotion	1.97
Identity	No domain	1.83
Regulation	No domain	1.46
Shaping knowledge	No domain	1.46
Associations	No domain	1.45
Natural consequences	Knowledge	1.25
Antecedents	Social influences	0.72
Identity	Social influences	0.72
Natural consequences	Beliefs about consequences	0.63
Natural consequences	No domain	0.46
Repetition and substitution	No domain	0.46
Reward and threat	Beliefs about consequences	0.37
Feedback and monitoring	No domain	0.27
Self-belief	No domain	0.27
Reward and threat	Social influences	0.12
Comparison of outcomes	No domain	0.01

There was also an association between the 16 ‘bottom-up’ groupings and the 13 ‘top-down’ groupings (χ² = 437.80; p < 0.001). Figure 3 shows the level of overlap between structures; Pearson’s residual values range from –1.72 to 6.66. Thirty six of the 208 combinations showed greater than expected overlap, achieving positive Pearson’s residual values (see Table 4). Six combinations achieved Pearson’s residual values > 5 indicating a comparatively high level of overlap and these combinations were ‘repetition and substitution’ and ‘skills’ (Pearson’s residual value = 6.66), ‘goals and planning’ and ‘goals’ (Pearson’s residual value = 6.41), ‘covert learning’ and ‘beliefs about consequences’ (Pearson’s residual value = 5.76), ‘self-belief’ and ‘beliefs about capabilities’ (Pearson’s residual value = 5.70), ‘scheduled consequences’ and ‘reinforcement’ (Pearson’s residual value = 5.22), and ‘antecedents’ and ‘environmental context and resources’ (Pearson’s residual value = 5.20).

FIGURE 3.

Pearson’s residual values for the association between BCT allocation in ‘top-down’ theoretical domain groupings and ‘bottom-up’ groupings. Note: ‘No domain’ indicates BCTs within a ‘bottom-up’ BCTTv1 grouping not assigned to any domain. Reproduced from Cane et al. 73 © 2014 The Authors. British Journal of Health Psychology published by John Wiley & Sons Ltd on behalf of the British Psychological Society. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

Discussion

Examination of the hierarchical structure of BCTTv1 uncovered a ‘higher-order’ grouping strategy taken by the behaviour change experts in the ‘bottom-up’ task and the dendrogram indicates that some groupings of BCTs within the 16-grouping solution can be considered as more clearly distinct from others. The grouping of BCTs in the ‘top-down’ sort task has helped illuminate relationships between particular BCTs and theoretical domains and could aid the selection of BCTs in the construction of theory-based interventions. There was a moderate overlap between the 16 BCT groupings derived from the ‘bottom-up’ inductive approach and the 12 groupings from the ‘top-down’ theoretically driven approach, indicating some common conceptualisation of BCTs across these two approaches. These findings may help to further our understanding of the relationships between BCTs and enable researchers to use common BCT grouping labels to discuss individual, or groupings of, BCTs in behaviour change research.

The grouping methods employed in the ‘bottom-up’ and ‘top-down’ sort tasks improve on previous attempts to group BCTs using consensus approaches. First, use of an open-sort grouping task allowed for the individual groupings of BCTs defined by participants to hold equal weight within the final solution and be aggregated using empirical techniques (hierarchical cluster analysis in the ‘bottom-up’ sort task and DCV methods in the ‘top-down’ sort task). As a result, the groupings reported here are potentially more robust than those derived using consensus methods among a few people. 43

A second advance was that a comprehensive, cross-behavioural domain taxonomy of BCTs was used rather than BCTs relevant for a single behavioural domain (e.g. road safety,82 smoking cessation,32 weight management83). Third, the BCTs were grouped according to the perceived active ingredients underlying BCTs, rather than by categorisations that may not have reflected how people think about BCTs.

In addition to providing 16 groupings, the ‘bottom-up’ open-sort task yielded systematic empirical estimates of how distinct the groupings are. Examination of this hierarchical structure revealed that BCTs related to reward and threat, and those related to cues and cue responses, were conceptualised quite distinctly from the other BCTs. The least distinct groupings (i.e. ‘social support’, ‘comparison of behaviour’, ‘self-belief’, ‘comparison of outcome’ and ‘identity’) comprised BCTs relating to social support, social comparisons, and self and identity, suggesting that there is less clarity about the BCTs within these theoretical domains. Four further groupings of BCTs (‘covert learning’ and ‘natural consequences’, and ‘feedback and monitoring’ and ‘goals and planning’) lay between these most distinct and least distinct groupings. BCTs in distinct groupings are clearly perceived to share a common mode of action in changing behaviour whereas BCTs in less distinct groupings may be viewed as having less distinct or more than one mode of action.

This difference in distinctiveness not only has implications for understanding how BCTs are conceptualised by behaviour change experts but also has implications for the practical use of BCTTv1 in behaviour change research. For example, the groupings increase the practical use of BCTs by aiding recall. Distinct sets of individual items with semantic similarity can be more easily recalled than a single list of individual items both in the short term and long term, particularly when the semantic category is cued. 84–86 This is especially useful when quick reference to BCTs is necessary, for instance when coding descriptions of interventions or in choosing BCTs to develop or report a BCI. Therefore, in those cases where the groupings are less distinct, adopting additional strategies to aid recall the groupings may be of particular advantage.

The ‘top-down’ mapping of BCTs to theoretical domains advances the limited consensus methods used by Michie et al. 43 by using an improved BCT taxonomy, an empirically validated TDF and a larger number of respondents. In this ‘top-down’ task, 59 out of 87 BCTs were reliably allocated to one or more of the TDF domains with a further 12 BCTs having borderline statistical significance. Thirty-seven BCTs were allocated to three domains that also had high confidence ratings and ICCs: ‘beliefs about consequences’, ‘reinforcement’ and ‘social influences’. This suggests that these are the theoretical domains for which there is the greatest number of agreed methods for bringing about change. Other domains also showed high agreement but had fewer associated BCTs – ‘behavioural regulation’ had only one assigned BCT but achieved good agreement, while ‘goals’ had five BCTs assigned with good agreement. In designing interventions, it may be more important to have a few agreed BCTs than to have a large choice of BCTs available to target change in a given theoretical determinant of behaviour. Further evidence is required to ensure that these ‘agreed’ BCTs do in fact achieve behaviour change by changing the proposed theoretical domain. For the two theoretical domains for which no BCTs were reliably assigned, there would appear to be no shared, or recognised, way of changing them.

Most of the commonly used BCTs were associated with a theoretical domain. Of the 24 most frequently identified BCTs in Michie et al. ,40 18 were clearly grouped into one of the 14 domains and the remaining seven BCTs were not reliably allocated to any domain even though they could be identified reliably in the intervention descriptions. This finding suggests that these BCTs may have evolved from several different behavioural domains, theoretical approaches or disciplines and, therefore, may be less associated with a particular theoretical domain.

Abstract

Objectives: To evaluate two programmes of user training in improving reliable, valid and confident application of BCTTv1 to code BCTs in descriptions of BCIs.

Methods: 161 trainees (109 in face-to-face workshops and 52 in distance group tutorials) were trained to code frequently occurring BCTs. Training was evaluated by comparing three measures before and after training: (1) intercoder agreement, (2) trainee agreement with expert consensus and (3) confidence ratings. Coding was assessed for 12 BCTs in workshops and for 17 BCTs in tutorials. Trainees also completed a course evaluation.

Results: Workshop and tutorial training improved trainee agreement with expert consensus [workshops: mean prevalence- and bias-adjusted kappa (PABAK) before training = 0.39, after training = 0.50; tutorials: mean PABAK score = 0.57, after training = 0.72; both p < 0.05] but not intercoder agreement (p = 0.08 and p = 0.57, respectively) and increased confidence ratings for BCTs assessed in workshops (mean number of assessed BCTs, identified with high confidence before training = 8.38, after training = 9.56; p < 0.001). Training was evaluated positively by trainees; all components of both types of training were highly rated in terms of usefulness.

Conclusions: Both methods of training improved trainee agreement with expert consensus and confidence for BCTs assessed, but neither method improved intercoder agreement. This varied according to BCT.

Introduction

In order to enhance understanding and use of BCTs in BCIs, we have improved the quality of their labels and definitions, and developed a hierarchical structure of BCTTv1. 73 However, it is not only the content and form of the taxonomy that drives reliable and valid application of BCT labels/definitions but also the extent to which the user has been trained to use the taxonomy. Identifying (coding) BCTs involves a deductive process of categorising qualitative information (e.g. descriptions of interventions) using an established coding framework and instructions. The process of coding BCTs is a highly skilled task requiring familiarity with the BCT labels and definitions and one which requires coders to make a series of complex interpretative judgements. 87,88 Achievement of good intercoder reliability (i.e. the extent to which coders agree on the presence/absence of BCTs identified in intervention descriptions using the taxonomy as a coding framework) is therefore not only a function of the clarity of the taxonomy and its coding guidelines, but also of the competence of its coders.

Intercoder reliability has been demonstrated in using BCT taxonomies among coders with varying amounts of training and experience. 20,23,28–30,48 Training in taxonomies, when reported, has generally involved manual-based coding instructions, prompting of practice and one-to-one feedback from those familiar with using the taxonomy. Delivery and intensity of training has varied, with some coders receiving one-to-one feedback from experienced BCT coders whereas others have trained themselves using coding manuals. Systematic documentation of the training process and evaluation, involving the comparison of coding competence to apply BCT taxonomies with reliability and validity before and after training, will establish whether or not training can enhance coding competence.

One of the goals of training in the use of the taxonomy is to teach coders to recognise BCTs as defined by the taxonomy rather than relying on their own subjective judgements that are triggered by the BCT label. 89,90 Another goal of training is to enhance the ‘validity’ of coder judgements, that is the extent to which coders agree with expert BCT coders as to the presence or absence of BCTs. Expert coding, assessed here as a consensus between expert coders, is the nearest we have to an objective standard of ‘validity’. An effective training programme is therefore one that not only enhances intercoder agreement between trainees but also enhances agreement with expert consensus about BCTs identified.

The research literature suggests that collaborative or co-operative training strategies (i.e. working together in small groups towards a common goal) and active learning techniques such as discussion are more effective than traditional, lecture-style training for acquiring new knowledge, building skills and increasing motivation to continue to improve skills. 91 An effective training programme is built on four basic principles: (1) setting training goals involving provision of information or concepts to be learned, (2) demonstration of knowledge and skills to be learned, (3) practice or rehearsal of skills learned and (4) provision of feedback to trainees during and after practice. 92,93 Skills are more likely to be retained and improve future practice if trainees feel challenged, receive positive feedback and find the learning process interesting and enjoyable. 93,94 This evidence and these principles clearly reflect the relevance and use of BCTs such as feedback on behaviour, goal-setting, social support and behavioural practice/rehearsal directed at changing skilled behaviours. Coder training incorporating these principles and BCTs has previously been evaluated in applying the taxonomy to specify BCT content in written intervention descriptions of behavioural support in smoking cessation, and training in this context was a 3-hour workshop delivered to a mixture of research psychologists and non-psychologist practitioners. 95 This study found that training improved trainees’ coding competence (i.e. their agreement with expert consensus about which BCTs are present).

Developing user training for applying Behaviour Change Technique Taxonomy version 1

The increasing use of the BCT approach, in particular for specifying interventions in evidence synthesis, has prompted high demand for training in the reliable and valid application of BCTTv1. In response to this, training programmes have been developed building on previous BCT coder training conducted by the study team and integrating the principles drawn from the literature. 92–95 These training programmes are evaluated in this chapter. They were designed to train coders to identify the most frequently occurring BCTs from BCTTv1, that is those that were found to occur most frequently in intervention descriptions. The decision was made to train coders in the most frequently occurring BCTs as training in 93 BCTs was not feasible within the proposed training intensity and duration, and because frequently occurring BCTs would be more accessible to trainees as well as more useful for them to learn. Two methods of training were developed: workshops, which involved face-to-face group training for 1 day, and distance tutorials, which were usually delivered by telephone to groups of two to four people in four, 1-hour sessions held over 1–2 months. The latter allowed training of coders across the world. Two substudies evaluated these training programmes and addressed the following research questions (RQs):

Does face-to-face training (1-day workshops) and distance training (group tutorials) improve the specification of BCIs by BCT as assessed by increased:

1. Intercoder agreement about BCTs identified?

2. Agreement with BCTs identified by expert consensus?

3. Confidence ratings for BCTs identified as present?

4. Do trainees evaluate BCTTv1 training as a useful experience?

5. What proportion of trainees reach an acceptable standard of competence following training?

One-day workshops

Distance group tutorials

Analysis

For both workshops and tutorials, the following analyses were conducted.

Intercoder agreement about BCTs identified was assessed by using the PABAK; see Byrt et al. 99 PABAK was used rather than Cohen’s kappa statistic101 or percentage agreement because it adjusts for (1) coders sharing bias in the use of categories and (2) high prevalence of negative agreement (i.e. when both coders agree the BCT is absent). Trainees were randomly allocated into coding pairs across tutorial groups using a random number generator. When both trainees identified the BCT as present or absent, agreement was recorded and when one trainee identified the BCT but the other did not identify the BCT, disagreement was recorded. The PABAK score was calculated for each trainee pairing and means reported across pairs and for each of the BCTs assessed.

Trainee agreement with expert consensus was assessed by calculating the number of BCTs identified by each trainee that were also identified by expert consensus and by using PABAK. Each trainee was paired with the expert consensus. The number of agreements and disagreements between trainees and expert consensus were recorded and were used to calculate the PABAK score. Means were reported across trainee–expert consensus pairings and for each of the BCTs assessed. When a trainee identified the BCT identified as present by expert consensus, agreement was recorded and when the trainee did not identify the BCT, or identified a BCT not included in the consensus, disagreement was recorded.

Confidence for BCTs identified was assessed by calculating the frequency and percentage of high confidence ratings (i.e. ‘++’: BCT present beyond all reasonable doubt and clear evidence available) for BCTs identified as present. High confidence ratings were included in the analysis so that we could easily distinguish BCTs identified with certainty.

Change following training: paired sample t-tests were used to assess change in (1) agreement between trainees about BCTs identified (intercoder PABAK), calculated across trainee pairings and across BCTs; (2) the number of BCTs identified by trainees also agreed on as present by expert consensus; (3) trainee agreement with expert consensus (trainee consensus PABAK) across trainee consensus pairings (i.e. each trainee paired with the expert consensus) and across the BCTs assessed; (4) high confidence ratings (i.e. ‘++’) for BCTs identified as present; (5) perceived confidence; and (6) reported intention to use BCT taxonomies in the future.

Previous experience: frequencies and percentages were calculated to describe the number and proportion of trainees with previous experience of BCT taxonomy use. Means and SDs were calculated to describe trainees’ expertise associated with BCIs (separate means were calculated for each of designing, delivering, reporting, reviewing experience and for use of behaviour change theories; an overall mean was calculated across these categories).

Evaluation of training: trainee ratings of the content of the training and ratings of the materials used were summarised using means and SDs. A content analysis of the written feedback was conducted to identify training components that participants reported to work well or not well. Given the novelty of this type and format of coder training, we decided that content analysis provided an appropriate first step for investigating patterns within coders’ initial feedback. CW conducted the first round of analysis, allocating coder feedback into categories based on similarity of content. KS (see Acknowledgments) then checked allocation of coder feedback to content categories. Any discrepancies were resolved through discussion between the two researchers.

Proportion of trainees reaching an acceptable standard of competence following training: an acceptable standard of competence was defined in terms of trainee agreement with expert consensus. Landis and Koch102 suggested that kappa values of 0.60–0.79 indicate ‘substantial’ reliability and those > 0.80 are ‘outstanding’. Thus, we used this as a guide and considered trainees achieving a PABAK score of at least 0.60 to have reached an acceptable standard of competence. To assess change in competence from before to after training, a one-way analysis of variance (ANOVA) was carried out. To evaluate the two training methods in increasing competence, a 2 (before vs. after training) × 2 (workshop vs. distance tutorial) ANOVA was carried out.

Results

Discussion

Training in using BCTTv1 in the form of 1-day workshops or group tutorials held over 6 weeks increased average trainee agreement with expert consensus and doubled the proportion achieving the competence threshold to 46% and 78% respectively. Training also increased workshop trainees’ confidence in identifying BCTs assessed in descriptions of BCIs, but did not improve intercoder agreement for either method. The opportunity to apply new knowledge and skills in a number of coding tasks followed by group discussion was evaluated as being a useful approach by trainees.

The difference between training methods in the proportion of those achieving an acceptable standard of coding competence may be partly due to differences in the percentage showing competence before training. Workshops required less commitment and may have attracted participants with less involvement and less competence in BCT coding. The fact that tutorial training allows consolidation of learning and practice between sessions may account for the higher proportion of competent coders following training. This is the model that has been taken forward in developing an e-learning programme of training (see Chapter 7 and BCTTv1. Online Training103).

We identified BCTs for which training was effective, that is when agreement and intercoder agreement were poor before training but good after training; for example, behavioural practice/rehearsal. There was only one frequently used BCT that achieved good reliability without training for workshops and tutorials: material reward (behaviour). Two other BCTs achieved good reliability without training for tutorial trainees: self-monitoring of behaviour and self-monitoring of outcome(s) of behaviour. Overall, training achieved good validity and reliability for between one-quarter and one-third of the BCTs.

Some BCTs consistently achieved poor reliability before and/or after training, for example goal-setting (outcome), social support (unspecified), and information about social and environmental consequences. These BCTs may require further definition and/or refinement of labels and examples for trained coders to be able to identify them with high reliability and validity. For example, drawing on feedback from trainees, refinements could include further clarification of the distinction between BCTs with similar labels that also appear in the same taxonomy grouping [e.g. social support (unspecified) and social support (practical)] or addition of more examples to help coders distinguish between different ‘types’ of BCTs [i.e. ‘behaviour’ vs. ‘outcome’ BCTs such as goal-setting (behaviour) and goal-setting (outcome)]. It is therefore likely that the number of BCTs that can be effectively trained will increase as BCTTv1 develops.

Training increased confidence in identifying BCTs assessed for both training methods and significantly so among workshop trainees. Tutorial trainees also reported increased overall confidence in using the taxonomy. This may be due to the learning environment provided by tutorials which includes a support network built over the period of multiple and regular sessions, and more time for reflection and practice between sessions. In the current study, we did not compare differences in the efficacy of training across the tutorial groups. While guided by a training manual, it is feasible that the engagement style and delivery of support, guidance and feedback differed across expert tutors despite having a similar level of experience to one another. Confidence in applying BCTTv1 requires learning the complexities and challenges of applying it, which may decrease or increase confidence in the short term. Therefore, future research should examine how this changes with training, further experience and across different modes of training delivery.

Although a direct comparison of the two BCTTv1 training methods was not the focus of study 3, some reflections are possible about the comparative merits, acceptability and feasibility of each approach. Workshops may be considered the more cost-effective and time-effective option given that a relatively large group of coders can be trained over the course of a day. They may also be more accessible to participants who simply wish to learn about the approach but with less commitment to becoming expert BCT coders.

Compared with workshop trainees, tutorial trainees started with a greater level of experience and expertise in using BCT taxonomies. It may be that training involves a process of trainees changing their existing interpretations of a BCT as understood from their discipline’s perspective. Greater pre-existing experience may have meant that ‘unlearning’ needed to take place before competence could be reached. This could help to explain why some BCTs achieved poor reliability and validity before and/or after training. Small group tutorial training delivered over a longer period of time provides a learning environment which includes a support network built over the period of multiple and regular sessions and more time for reflection and practice between sessions. Having access to an interactive training resource that provides continued support and feedback beyond that of time spent with the tutorial group would provide the additional space and time that this level of learning requires. Such a resource has been developed and is described in more detail in Chapter 7.

It is important to remember that effect of training was evaluated on trainee coding of just one intervention description before and after training for each of the methods. The training materials used in 1-day workshops and tutorials were written to exemplify particular BCTs, to highlight the learning principles and to ensure inclusion of the frequent BCTs targeted for training. However, the quality of published intervention descriptions is generally poor,95 with techniques described using different terminology and referred to using different labelling. 104,105 As the use of BCT methodology increases, BCT content as reported in intervention descriptions is likely to become much clearer, leading to increased ease of identifying BCTs. 95 It is also important to acknowledge that, despite attempts to ensure diversity of trainees, our sample were predominantly female and from the UK. As the use of BCT methodology becomes more widely used, it will be possible to recruit a broader range of trainees for training evaluation.

For users to maintain good levels of accuracy and reliability after training, efforts should be made to maintain knowledge and skills over time. 106 Therefore, we recommend that coders already trained in use of BCTTv1 regularly review training materials and check their reliability and accuracy.

Abstract

Objectives: To assess frequency and reliability (intercoder and test–retest) of identifying BCTs in written intervention descriptions.

Methods: 40 coders were trained to identify BCTs defined in BCTTv1. Coders identified BCTs in 40 intervention descriptions published in protocols and repeated this task 1 month later. A consensus of judgements reached by coders who were experienced in coding BCTs (and also developed BCTTv1) were compared with those of trained coders and used as the index of current validity.

Results: 80 out of 93 (86%) defined BCTs were identified by at least one trained coder and 22 (28%) of these were identified in 16 or more of 40 descriptions. Good intercoder reliability was observed across 80 BCTs identified in the protocols; 64 (80%) achieved mean PABAK scores of ≥ 0.70 and 59 (74%) achieved mean scores of ≥ 0.80. There was good test–retest reliability; good within-coder agreement was observed between baseline and 1 month. For the 32 coders providing data at both time points, mean PABAK scores for the two occasions ranged from 0.84 to 0.99. Good concurrent validity for 15 BCTs was agreed to be present by the experienced coders and mean PABAK score of ≥ 0.70 was achieved for 14 out of the 15 BCTs.

Conclusions: BCTTv1 can be used by trained coders to identify BCTs in intervention descriptions reliably (both in terms of agreement with each other and over time) and validly (assessed by agreement with experienced coder consensus). Some BCT definitions require further clarification.

Introduction

Results from two initial phases of reliability testing indicated BCTTv1 to be reliable in specifying 26 frequently occurring BCTs when used by the taxonomy developers. 40 These findings helped to identify BCT labels and definitions requiring refinement. It also suggested that user training may be required as high reliability is not only dependent on the content of the taxonomy but also on the extent that the user is skilled in using it. Since this initial assessment of reliability, the taxonomy has been developed further and we have developed a user-training programme for BCTTv1 delivered through face-to-face workshops, distance group tutorials and more recently via an online training programme (see Chapter 4). Study 4 reports a detailed assessment of reliability of use of the taxonomy also reported in Abraham et al. 107 It investigates concurrent validity of BCTTv1 and evaluates the use of BCTTv1 by newly trained users (primary researchers, systematic reviewers and practitioners) in specifying BCTs across a wide range of complex BCIs.

Since Abraham and Michie’s20 initial paper, intercoder reliability has been demonstrated for various subsequent BCT taxonomies. 20,23,27,28,30,48 One indication of intercoder reliability is the percentage of BCTs for which there is agreement between coders that a BCT is present or absent. Abraham and Michie20 reported > 93% agreement between three pairs of coders (including the authors). As agreement scores tend to be inflated by the number of occasions on which neither coder reports the BCT to be present, Cohen’s kappa scores were calculated excluding BCTs agreed to be absent. Across 78 reliability tests (of identification of 26 defined BCTs and BCT clusters across 195 published intervention descriptions) the average kappa per technique was 0.79 with only three BCTs yielding kappa scores of < 0.70. Landis and Koch102 have suggested that kappa scores of 0.60–0.79 can be described as ‘substantial’ and that scores of ≥ 0.80 can be described as ‘outstanding’ but, conventionally, 0.70 is taken to indicative of acceptable intercoder reliability. Thus, these findings demonstrated that BCTs can be reliably identified in published descriptions. Similar methods were applied to BCTTv1.

Initial assessment of BCTTv1 used six experienced coders on the BCTTv1 study team and calculated kappa (PABAK, for full description see Analysis) when there were at least five identifications of a BCT in the 65 coded intervention descriptions. 40 We set the criteria of ‘at least five identifications’ because the more frequent the BCTs, the greater the confidence that PABAK is a useful indicator of reliability of judging the BCT to be present. Of the 26 BCTTv1 BCTs meeting this criterion, nine were found to have ‘outstanding’ and 14 ‘substantial’ reliabilities. Although these are good kappa results, they raise several questions. First, would equally good reliability be obtained by newly trained coders who have not been involved in the development of the taxonomy? The extensive discussions involved in taxonomy development described by Michie et al. 40 might result in enhanced agreement owing to tacit knowledge beyond the explicit BCT labels and definitions. therefore, it is important to examine the extent to which the BCT taxonomy enables newly trained coders (i.e. coders who have not been involved in the development of BCTTv1) to reach agreement about the content of interventions.

Second, are these levels of agreement good enough to enable replicable implementation of intervention content? Most researchers would probably accept that BCTs definitions generating kappa scores of 0.70 can be reliably identified but may have reservations about those with lower scores. 108 Although these reliability statistics indicate levels of agreement, they do not quantify the degree of error that might be incurred when scientists try to replicate interventions from their descriptions. When two practitioners read a report of an effective intervention with kappa agreement of 0.80, which is ‘outstanding’, how different might their implementation of the intervention be? For example, in coding 40 BCTs, it would be possible for each of two coders to identify two BCTs but for each to have identified different BCTs, or for one coder to identify and implement five BCTs while the other identified and implemented nine (i.e. a further four BCTs), would result in quite different interventions.

Third, it is possible to assess reliability of coding of specific BCTs, or of an intervention description as a whole. So, for example, if an intervention contained some BCTs that were reliably coded and some that were not, replication of the intervention might fail owing to the omission of poorly recognised BCTs. There might also be failure to replicate if the intervention description was badly written, either by omitting aspects of the intervention or by lack of clarity in the description. Improved coding systems cannot facilitate detection of a BCT if the description omits to mention that it was included in the intervention; Lorencatto et al. 12 found that published descriptions reported, on average, less than half the BCTs that were included in the longer protocol descriptions. BCTTv1 might enable authors to report fuller, clearer descriptions of their evaluated interventions. 17 On the other hand, when the description is simply unclear, BCTTv1 should enable agreement between coders as it clearly specifies that coders should not infer the presence of a BCT from vague descriptions. It is therefore important to assess not only the reliability of coding of each BCT, but also to investigate whether or not intervention descriptions are satisfactorily coded.

Fourth, it is quite possible for two coders to reach good agreement but be wrong, that is it is reliable but not valid. While reliability of use of BCT taxonomies has typically been evaluated, it is also important to assess the ‘validity’ of these judgements. Poor reliability or a good level of reliability between newly trained coders but without agreement with expert consensus would both be unsatisfactory outcomes. High reliability with poor validity might occur for a variety of reasons such as misleading pre-coding discussion or a misleading cue in the intervention description. Estimates of validity require a criterion against which the codes are judged, but in the case of intervention descriptions such a criterion does not exist.

Abraham and Michie20 assessed reliability between the two coders who developed the 26 BCT definitions (the authors) and also between the first author and two other coders trained to use the taxonomy. Findings showed no significant difference between reliability scores indicating that trained coders could identify BCTs as reliably as coders involved in defining BCTs. In order to assess concurrent validity of BCTTv1, we compared trained coders’ identification of BCTs with BCTs judged to be present in descriptions by a consensus of experienced coders. This consensus took as its starting point the independent coding by pairs from a pool of six taxonomy developers who were experienced in identifying BCTs in intervention descriptions. This resulted in the reliability data for 15 BCT definitions reported in both Michie et al. 40 (see Table 3) and Abraham et al. 107 It was then further developed by discussion of any discrepancies by each of the pairs. Further discrepancies were identified during BCT training of new coders. If a resolution was not obvious, SM and the study researcher (CW) reviewed the remaining discrepancies and proposed a coding. The list of BCT codes resulting from this process was circulated to the group of six researchers and each agreed the final codes. In the absence of any better criterion for valid codes, consensus regarding the presence for these 15 BCTs was used as a criterion by which to judge the validity of trained coders’ judgements across the same 40 intervention descriptions.

Fifth, BCT identifications rated with low confidence by coders could indicate problems with specific BCTs or ambiguity of specific intervention descriptions, which might prevent achievement of good intercoder reliability and accurate replication. However, our analyses have shown that reliability of BCT identification was not positively correlated with coder confidence and indeed tended to be negatively correlated. 109 Therefore, it is important to ascertain, if possible, whether or not codes made with high confidence are also reliable and valid ratings.

Finally, various authors have proposed different methods of assessing kappa that adjust for its limitations. Byrt et al. 99 proposed the ‘PABAK’, which results in considerably more meaningful indices for data such as BCT coding when the prevalence of each BCT is low and bias may occur; therefore, it was the statistic chosen by Michie et al. 40 as the most appropriate to measure intercoder reliability in applying BCTTv1. Gwet et al. 110 suggest that further adjustment should be made for chance agreement that might occur when two coders make random ratings and make allowance for this in the alternative chance – corrected statistic to kappa (AC1) statistic110 (see equations 7 and 8 in Gwet et al. 110 and equation 4.1 in Gwet et al. 111). We compare reliability results generated by applying the PABAK and AC1 formulae.

The aims of study 4 were:

1. To assess the reliability (intercoder and test–retest) and validity of labels and definitions in BCTTv1 when used by newly trained coders to specify BCTs across a range of intervention descriptions.

2. To assess whether or not confidence in codes relates to validity of codes.

3. To identify labels and definitions in BCTTv1 requiring further refinement.

First, we investigated how many of the 93 BCTs were used by participants to describe the interventions coded in order to ascertain whether or not BCTTv1 contained too many BCTs.

1. How often are particular behaviour change techniques identified in intervention descriptions?

We then addressed the following RQs.

Method

This study is also published as Abraham et al. 107

Analysis

The extent to which coders agreed on the absence/presence of BCTs in the descriptions (‘intercoder reliability’) was assessed using PABAK. 99 Coders at time 1 were randomly allocated to pairs working on the same set of intervention descriptions. Mean PABAK scores for each trained coder pairing, for each BCT and for each intervention description were calculated using the number of agreements and disagreements between each coding pair. Trained coder agreement with expert consensus was calculated by pairing each individual trained coder with the expert consensus BCT identifications (as described above). Trained coder agreement with expert consensus was represented by mean PABAK scores for each trained coder–consensus pairing and for each BCT.

Test–retest reliability of trained coder judgements was also assessed using PABAK. Coders at time 1 were paired with themselves at time 2. Mean PABAK scores for each intervention description were calculated using the number of agreements and disagreements. A PABAK score for each coder was then calculated using the mean of these scores.

Prevalence- and bias-adjusted kappa was used rather than Cohen’s kappa statistic. 101 Many authors have discussed inter-rater agreement measures. 111 Kappa tends to underestimate identification reliability when the number of instances is small or has asymmetric distribution between agreements and disagreements, thus PABAK overcomes these problems. Guidance on interpretation of Cohen’s kappa and other reliability statistics has been published and, conventionally, 0.70 is regarded as indicative of acceptable of good inter-rater reliability as indicated by Cohen’s kappa. In the absence of evidence-based guidance on interpretation of PABAK scores, we report means above and below 0.70 and 0.80.

Test–retest of intercoder reliability between time 1 and time 2 and of trained coder agreement with experienced coder consensus was assessed by bivariate correlations, and stability of reliability was assessed by paired t-tests. For each of the BCTs identified, the number of high confidence ratings by each trained coder at time 1 and at time 2 was calculated. The percentage of high confidence ratings was then calculated for each BCT taking into account the frequency of BCT identification across the 40 intervention descriptions. Paired t-tests (i.e. by pairing percentage of high confidence ratings made at time 1 with percentage of high confidence ratings made at time 2) were used to assess stability of high confidence ratings. A multivariate ANOVA was carried out to assess whether or not percentage of high confidence ratings at time 1 and at time 2 differed according to frequency of BCT identification (i.e. between categories of frequently, occasionally and rarely).

A number of alternatives to Cohen’s kappa have been developed. Gwet110 tested a number of such reliability indices and concluded that the AC1 statistic (see equations 7 and 8 in Gwet110) had optimal output characteristics. This is particularly true when the frequencies of occurrence are small. In the present study, a two-tailed chi-squared test applying the Yates’ correction for continuity was used to explore which statistic, PABAK or AC1, gave the higher number of BCTs achieving good reliability (i.e. ≥ 0.70).

The free response feedback provided by coders was subjected to analyses. The lead researcher on these analyses (CW) read coders’ feedback and sorted responses into themes. Themes were developed during inspection of the data. Additional themes were added if data did not fit the existing themes. An independent coder (KS) then checked allocation of feedback to themes. Any discrepancies were resolved through discussion between the two researchers.

Results

Discussion

Eighty of the 93 BCTs defined in BCTTv1 were identified in 40 intervention descriptions by at least one trained coder, 22 BCTs were identified in at least 16 descriptions and 47 BCTs were identified in at least six of 40 descriptions. Thus, coders made extensive use of BCTTv1, justifying the large number of BCTs included. Clearly specification of BCIs requires at least this range of BCTs to describe the active content.

Good intercoder reliability was observed across the 80 identified BCTs, with 80% achieving mean PABAK scores of at least 0.70 and 74% with scores of ≥ 0.80. Poorer reliabilities were more common for frequently occurring BCTs; just 12 of the 22 (55%) achieved mean PABAK scores of ≥ 0.68 and six (27%) achieved ≥ 0.70.

Intercoder reliability was equally good across intervention descriptions and remained similar across two tests separated by 1 month, indicating temporal stability. There was also good trained coder agreement with developer consensus about the 15 BCTs agreed on as present, thus indicating concurrent validity.

Trained coders’ confidence in their BCT identifications increased from time 1 to time 2, and varied across BCTs. There tended to be a negative correlation between trained coders’ confidence of BCT identification and intercoder agreement. Coders’ confidence in identifying BCTs was not related to either intercoder agreement or to agreement with experienced coders. This suggests that perceived confidence does not appear to be a useful indicator of accuracy of identification of BCTs in intervention descriptions. Reliability was consistently higher when represented by the AC1 statistic than the PABAK statistic.

The BCTTv1 taxonomy of BCT labels and definitions is an important development of previous work and provides the most comprehensive listing of BCTs to date. Of the 22 BCTs observed in 16 or more of 40 intervention descriptions, 18 (82%) were also identified in the Abraham and Michie20 BCT taxonomy, confirming the frequent occurrence of these 18 BCTs. However, while Abraham and Michie defined only 22 BCTs and four packages of BCTs, BCTTv1 defines 93 separate BCTs, 80 of which were identified by coders in 40 published intervention descriptions. The data reported here provide good overall intercoder and test–retest reliability for 80 of those definitions among trained coders. Furthermore, the data suggest that newly trained coders can code accurately as defined by agreement with experienced coder consensus. Given the expertise derived from the experience of developing BCT definitions and the previous experience of the developers in identifying BCTs in intervention descriptions, correspondence between trained coder judgements and experienced developers’ judgements was regarded as a measure of the validity of the trained coders’ judgements. Thus, validity of BCTTv1 is demonstrated among new users and does not depend on additional discussion and cues that may have influenced the experienced coders. Therefore, the BCTTv1 taxonomy can be used successfully, with appropriate training, by a wide range of users to identify a wide range of BCTs included in intervention descriptions.

Beyond testing the reliable and valid use of current BCT labels and definitions, a primary purpose of this research was to identify definitions that might be applied more reliably with further clarification or guidance. Sixteen frequently identified BCTs achieved PABAK scores of < 0.70 with nine of these falling below 0.60. These included four pairs of BCTs defining variants of the same underlying change mechanism, namely (1) Information about (a) health consequences and (b) social and environmental consequences; (2) goal-setting in relation to (a) outcomes and (b) behaviours; (3) feedback on (a) behaviour and (b) outcome(s) of behaviour; and (4) social support both (a) practical and (b) unspecified.

It may be appropriate to remain somewhat circumspect about trained coders’ feedback on clarity of BCT definitions as confidence in BCT identification was not related to observed intercoder reliability (and indeed tended to be negatively correlated) or to trained coder agreement with experienced coder consensus. Nonetheless, coders provided clear feedback suggesting that further distinction between such BCTs groupings was needed.

Intercoder reliability was poor for ‘instruction on how to perform behaviour’ and individual coder feedback suggested confusion with ‘demonstration of the behaviour’. This was one of four pairs of BCTs (including two goal-setting definitions) with which coders reported difficulty. Both prompts/cues and adding objects to the environment also fell below the 0.70 threshold. Further testing of the definitions of these BCTs is therefore warranted.

Trained coders commented on the definition of credible source. This BCT showed poor intercoder reliability among coders and poor validity. It was also the lowest intercoder reliability score among experienced BCT coders. This is clearly a BCT that requires further clarification and specification but should be retained because of its relevance to behaviour change. The importance of message–source credibility has been recognised for over half a century151 and coding for source characteristics in intervention descriptions has been shown to predict intervention effectiveness in meta-analytic studies. Perceived professional competence of intervention facilitators has also been found to predict effectiveness of human immunodeficiency virus-preventative interventions. 152 Similarities between intervention facilitators and recipients (e.g. age, ethnicity and behaviour risk group membership) explained differential effectiveness of interventions. This highlights that coding intervention descriptions for message source or facilitator characteristics is important. Plus it may be important to distinguish between perceived professional competence of those delivering the interventions and their ‘credibility’, which could also be based on common group membership. Finally, it may be important to distinguish between characteristics of those delivering interventions and the content of what they deliver. Certainly, further work on the label and/or definition of this BCT is needed.

The coders trained for this study seem typical of those likely to want to use BCT labels and definitions to identify intervention components associated with greater intervention effectiveness, for example when conducting meta-analyses. 20,23,48,56 Although we used data from 40 trained coders, we recommend that future studies continue to use at least two coders to check (and report) intercoder reliability.

We found that reliability was high and similar over the 40 intervention protocols coded. However, as reported above, we used a sample of descriptions included in well-cited journals that publish BCI protocols related to health improvement and this reliability might not be maintained in intervention descriptions of poorer quality.

We compared intercoder reliability assessments generated by the PABAK and AC1 statistics. Reassuringly, our findings indicate that these indices generate very similar patterns of intercoder scores across identified BCTs. We found, however, that for the most frequently indentified BCTs, the two indices differ significantly in their representation of reliability with PABAK generating lower reliability scores than AC1. These frequently identified BCTs were less reliably identified than other BCTs. Thus, our findings suggest that, at least in some contexts, the two statistics may not be directly interchangeable when data similar to the current data are investigated. Guessing may be more likely for frequently occurring BCTs than for rare BCTs meaning that random effects, such as those that AC1 controls for, could emerge. AC1 may therefore be a more appropriate index. It may also be helpful to interpret these statistics alongside other results (e.g. frequency of item observations). The consistently higher AC1 scores suggest that our data may have been influenced by random judgement effects that are controlled in AC1 but not PABAK, with resulting lower scores in PABAK. Nevertheless, even PABAK takes account of the high prevalence of negative codes and of the extent to which coders use similar marginal totals and, therefore, is more appropriate for this type of data than a simple kappa. The use of more than one statistic to assess agreement is a relatively new approach and the consistent difference between the scores indicates that the AC1 takes account of more sources of error in these data and, therefore, is more appropriate for establishing the level of agreement in BCT coding. The high correlation between the measures indicates that the pattern of relative agreement across BCTs is adequately represented by either score, although the PABAK may result in fewer frequently observed BCTs achieving reliability above thresholds such as 0.70, as used here.

Comparison of the pattern of results for frequent and rare BCTs suggests that PABAK is as good as AC1 for the rare but not for the frequent BCTs, while the poor reliabilities (using PABAK) were more common for frequently occurring BCTs. As was suggested by our comparison of confidence ratings across BCT categories, guessing may be more likely for common than for rare BCTs. This would result in occurrence of the random effects that AC1 controls and confirms the need to use AC1 rather than PABAK in studies assessing reliability. Further developments of reliability indices111 controlling for additional sources of error and more recently proposed indices may give an even better indication of the true reliability of coding.

In conclusion, our results show that BCTTv1 can be used by trained coders to reliably identify BCTs included in intervention descriptions. Feedback from our coders highlights several BCT definitions that require further classification and refinement before they can be identified reliably. BCTTv1 provides a uniquely comprehensive list of BCTs. The taxonomy can be used both by developers and implementers of BCIs and by those wishing to identify effective BCTs within complex interventions (e.g. systematic reviews).

Abstract

Objectives: To assess whether or not use of BCTTv1 improves the reporting of observed BCIs.

Methods: 166 participants (‘writers’) watched videos of BCIs and wrote descriptions of the active content delivered. Three different study designs were used so we could assess effects of BCTTv1 provision, with and without training. The first two studies were RCTs: study 5a examined provision of BCTTv1 without training and study 5b examined the effects of training plus provision of BCTTv1 compared with a control group receiving neither the taxonomy nor training. Study 5c used a within-person design to assess change in the quality of descriptions before and after training. Writers provided with BCTTv1 for the task evaluated usability and acceptability of using BCTTv1 to write descriptions. Twelve ‘raters’ (untrained in BCTTv1) assessed description quality in terms of clarity and replicability, and 12 ‘coders’ (trained in BCTTv1) coded the descriptions for BCTs.

Results: Providing BCTTv1 did not improve the reliability or validity of the descriptions written by untrained writers (for all measures p > 0.05; study 5a) and resulted in poorer quality of descriptions (all p < 0.05). Compared with untrained writers, those who were trained and using BCTTv1 wrote descriptions that had no greater quality or validity but which resulted in more reliable BCT identification (all p < 0.05) (study 5b). Following training, writers’ descriptions did not increase in reliability or validity but were rated to be of higher quality (all p < 0.05 except replicability of mode of delivery) than those written before training (study 5c). Writers thought that using BCTTv1 to write intervention descriptions was a difficult task but considered it ‘useful’, ‘good’ and ‘desirable’, and that their descriptions would be clear and replicable if the taxonomy were to be used (all means above mid-point of the scale).

Conclusions: We did not find a clear benefit of using the BCTTv1 to report observed interventions. This may be owing to task demands, the nature of the task evaluated, omission of other intervention components in descriptions, lack of time and experience, or an insufficient training package. Further work is needed to evaluate its usefulness in this area.

Introduction

There is a lack of guidance about how to report intervention descriptions in terms of both the proposed active content (BCTs) and supportive content necessary to engage the participant. 32

It is important to examine whether or not the use of BCTTv1 can improve the reporting of intervention descriptions. Improved reporting of intervention content means specifying the content of all the BCTs delivered in the intervention in a manner that makes them recognisable and replicable. 153,154 We investigated whether or not using BCTTv1 improved communication of the content of observed interventions. We used observed interventions with the aim of providing all the information about the intervention without writing it down and thus providing the words to be used. However, this is a rather artificial scenario as usually interventions are developed in an iterative process of reading and discussion.

This chapter reports three substudies that investigated whether or not BCTTv1 improves the reporting of descriptions of observed BCIs. We focused on whether or not provision of BCTTv1 alone and/or BCTTv1 plus training has an impact on the quality of reporting and recognition of BCTs in written intervention descriptions. The writers of intervention descriptions (‘writers’) reflected a range of potential users of BCTTv1: primary researchers, systematic reviewers, funders, policy-makers and practitioners, and both trained and untrained users of the BCTTv1. Participants watched videos of BCIs as delivered in practice and were asked to write descriptions and rate the usefulness and acceptability of BCTTv1 for this purpose. We assessed intervention descriptions generated by the writers in multiple ways. We assessed their clarity and replicability using independent raters who were not familiar with the BCTTv1 (‘raters’) as they reflect the typical reader of intervention descriptions. Separately, trained coders (‘coders’) were asked to identify BCTs and we assessed how well their coding corresponded with the agreement reached by experienced coders from the BCTTv1 project team (‘expert consensus’). We used trained coders as we showed that training in BCTTv1 improves the identification of BCTs96 (see Chapter 4, Methods).

A series of three substudies addressed the following RQs:

1. Does provision of BCTTv1 or training in its use improve:

   1. the quality (i.e. clarity and replicability) of intervention descriptions?

   2. recognition of BCTs described in intervention descriptions (in terms of intercoder agreement)?

   3. agreement with expert consensus about BCTs described?

2. Do users find BCTTv1 a useful and acceptable tool for reporting intervention descriptions?

Method

Study 5a: randomised controlled trial evaluating the effect of provision of Behaviour Change Technique Taxonomy version 1 on the writing of descriptions

Study 5b: randomised controlled trial evaluating the effect training in, and provision of, Behaviour Change Technique Taxonomy version 1 on the writing of descriptions

Study 5c: within-person study of the effect of training in, and provision of, Behaviour Change Technique Taxonomy version 1 on the writing of descriptions

Discussion

In Study 5a, 5b and 5c we investigated the effect of the provision of BCTTv1 alone and training plus BCTTv1, using both randomised and within-participant designs. The results do not show clear benefit of providing BCTTv1 for reporting observed interventions. This is not simply due to lack of power as even Study 5a which had the largest number of writers did not show the benefit of training except for reliability of BCT identification. As Tables 16 and 17 show, the descriptions produced in the trained and/or BCTTv1 condition were only found to perform better than those of the control condition for the reliability of identification of BCTs (study 5b) and for the rated clarity and replicability of the descriptions (study 5c). Even so, the latter results report only a small effect, one that might not be meaningful in practice and could simply be due to practice effects as the trained writing followed the untrained writing. Furthermore, there might have been more room for some writers to improve their writing ability from baseline ability to after training. Baseline scores in study 5c, for example, were much lower than for the comparable group in study 5b.

In contrast, there was evidence of poorer performance among those who had been trained and had access to BCTTv1 than those who received no training and BCTTv1. Untrained writers produced descriptions that were rated to be of higher quality (clarity and replicability) than untrained writers in study 5b, in which participants were randomised and no practice effects were possible. These results were somewhat unexpected and difficult to explain. Although they appear to have made descriptions easier to recognise in one study, they appear to have done the opposite in another. Given the difficulty and uncertainty as to how best to report interventions, one would have expected that a structured and detailed method would have led to consistently improved report writing. There are a number of possible explanations for the results of study 5b:

1. Trained writers completed the task at the end of a workshop and may have been more tired than untrained writers. If so, then one might expect their descriptions to be shorter and less detailed. However, this result was not found in study 5a, in which both groups completed the writing task before the workshop.

2. Alternatively, trained writers may have written longer or more complex descriptions that made it more difficult for coders to identify the distinguishing elements.

3. Trained coders may have concentrated on reporting BCTs and omitted other descriptive elements such as detail about the delivery procedures, which might be more recognisable.

4. The time constraint may have seemed more restrictive for the trained than the untrained writers with the result that their descriptions were less clear. The demands of the task may have seemed especially high to trained writers, expected to implement new knowledge and skill from their training to use the taxonomy with 93 BCTs in addition to reporting the content of a 9-minute video, which was shown only once.

5. Immediately following training, writers may be unable to clearly report all of their ideas about the intervention and may need more time or experience with BCTTv1 to write good descriptions.

6. The training in reporting intervention descriptions using the BCTTv1 may have been insufficient in terms of time and content. Most of the training focused on identifying BCTs from written descriptions and videos; however, only 30 minutes was devoted to how to write sentences about active ingredients using the BCTTv1 with a single task, and no emphasis was placed on the importance of writing down other intervention components, e.g. delivery procedures, alongside the active ingredients.

These explanations can be tested by examining whether or not controlling for the following variables alters the findings: the number of words (explanations 1 and 2), the intelligibility of the text (explanation 2), the number of BCTs identified (explanation 3) and ratings of time pressure and difficulty (explanation 4). Testing explanation 5 would require a longitudinal study and testing explanation 6 would require the development and evaluation of an optimised training package for using the BCTTv1 to report intervention content. These will be investigated in future research.

There were limitations in the design, the measures and the writing task. In terms of design, these studies confounded training condition with time of day, which was inevitable given the resources of the study. This might have been improved by asking participants to write the descriptions at a later date at the same time of day, although this introduces the possibility of further confounders such as additional reading by either group.

In terms of measures, we were seeking quality measures that reflected the likelihood that readers of the descriptions would be able to envisage the intervention in such a way that they felt confident about being able to replicate it. Although there are guidelines which aid the assessment of whether or not intervention components are reported, the standardised measures for assessing the quality of reporting of active ingredients and mode of delivery in intervention descriptions is still developing. 155 Hence, we developed measures of clarity and replicability specifically for studies 5a, 5b and 5c and ensured that active content was rated separately from mode of delivery of the intervention. However, further work needs to be done to establish their psychometric properties. While the use of indices of reliability of BCT identification is an accepted method, assessing validity of BCT identification is less well established. We used consensus following independent judgements by experienced, ‘expert’, coders as our best approximation to a criterion of validity against which judgements could be evaluated.

In terms of the writing task, this was an unusual task and was outside the normal activities involved in intervention development and reporting. The normal practice would not be to describe an intervention after observing it on a video, it would be to describe it after reading and discussion, without a strict time pressure. The task was constrained by the time available within the workshop format and it was also difficult to decide on the appropriate time allowed for completing the task. Nevertheless, participants did not express dissatisfaction with this task in their feedback. However, in retrospect, a better test of BCTTv1 for this purpose would be to investigate whether use of BCTTv1 enhances existing intervention descriptions rather than using it to create descriptions simply from observation. For example, authors of published intervention protocols might be invited to rewrite the intervention description using BCTTv1 and research participants asked to choose the ‘better’ description from each pair of old and new descriptions.

The finding of unexpected negative effects of training is consistent with some of the counterintuitive findings reported in the psychology literature. For example, there is evidence of a paradoxical effect of incentives on skilled performance, described as a ‘choking effect’. 156 Participants under high mental load and in a stressful situation displayed greater levels of physiological stress reaction when asked to relax than participants who were not asked to relax. 157 Furthermore, therapists who participated in a training programme showed unexpected deterioration in skills after training. 158 The authors proposed that this was evidence of a ‘post-training phase in which [participants’] performance actually declined in certain ways as they struggled to naturally integrate new techniques into their existing styles and approaches’. 158 It is possible that the participants in our study similarly responded to the intervention writing task in ways that parallel a choking effect or the post-training effect described in these examples. Hence, it may be that measurable benefits of training occur after a period of integration and practice. Further follow-up research could explore this possibility.

In conclusion, we did not find a clear benefit of providing BCTTv1 for reporting observed interventions. This may be owing to task demands, the nature of the task evaluated, omission of other intervention components, lack of time and experience, or an insufficient training package. Further work is needed to improve its utility in this area. This includes further examination of the data to identify potential reasons for the unexpected findings, testing the utility of BCTTv1 in enhancing existing intervention descriptions, and potential improvements of the training package, such as more detailed guidance for how to use BCTTv1 to report the active content of interventions alongside mode of delivery and a wider range of interactive tasks to train users in reporting good quality descriptions. Further work is also needed to train users in how to use BCTTv1 to report descriptions of newly developed interventions.

Disseminating Behaviour Change Technique Taxonomy version 1 development and project progress

A static website was developed during the BCTTv1 project to serve as an informative repository for taxonomy users: www.ucl.ac.uk/health-psychology/bcttaxonomy. 97 Google (Google Inc., Mountain View, CA, USA) analytic statistics from 2010 to 2013: 820 page views, 173 visits, 127 unique visits. Supplementary to regular newsletters sent to contacts via e-mail, the website provided users with updates on project progress, featured related publications, useful links and advertised forthcoming training opportunities.

Project progress was widely disseminated at national and international conferences. The ‘Theories and Techniques of Behaviour Change’ special interest group was established in 2012 at the Society for Behavioral Medicine annual conference to provide a platform for developers and users to discuss application, usability and future development of BCTTv1. A similar group was formed at the European Health Psychology Conference in 2013. Individual studies from the BCTTv1 project have been published as peer-reviewed journal articles (see Publications). 40,73,96,107

Providing training in use of Behaviour Change Technique Taxonomy version 1

Studies 3 and 4 showed that BCTTv1 offers a generally reliable method for specifying the active ingredients of BCIs when used by trained coders. We recommend that coders already trained in the use of BCTTv1 regularly review training materials and check their own reliability. In order to give coders further practice and experience in applying relevant learning principles to coding BCTs, and in order to train new coders using BCTTv1, we have developed an interactive online training portal: www.bct-taxonomy.com. 103 It is based on the tutorial training model. This includes practice coding tasks with associated feedback, the possibility of structured discussion guided by an expert tutor and access to a social support network to foster continued learning. It trains coders on a greater number of frequently used BCTs than those taught and evaluated in the current study.

Demand for face-to-face training remains high. Since completion of the project, the team have facilitated a further 100 workshops, training approximately 3000 researchers, practitioners and policy-makers.

Increasing accessibility and usability of Behaviour Change Technique Taxonomy version 1

To further increase usability of the taxonomy, a digital version of BCTTv1 as a smartphone and tablet application has been developed. The application is available for download on all mobile platforms (see www.ucl.ac.uk/health-psychology/bcttaxonomy97 for more information). Since launching in early 2014, the app has received 500 downloads via the Google Play™ store and Apple App^SM store (Apple Inc., Cupertino, CA, USA).

Informing future development of Behaviour Change Technique Taxonomy version 1

An interactive system is now being developed within the website to receive feedback from users which will inform development of BCTTv2. Discussions are under way to develop an international consortium to receive, synthesize and disseminate feedback from the use of BCTTv1 across populations, settings, behaviours and countries, and to develop a methodology for BCTTv2.

Creating a structure for Behaviour Change Technique Taxonomy version 1

The recognised need by the BCCTv1 study team and the IAB to produce a structure for BCTTv1 to make it more usable and BCTs easier to find raised several issues. The first was whether to create a theory-driven, top-down structure or whether to create a more pragmatic, bottom-up structure that reflected a shared way of thinking about BCTs among users with different theoretical perspectives and/or disciplinary backgrounds. The general view of the IAB and the study team was that it would be more important to start by adopting a pragmatic, bottom-up structure as it would enable use of the structured taxonomy by users with very different backgrounds and would form a basis for consensus on its use and future development.

The second issue concerned the nature of the groupings produced. Despite not explicitly drawing on theory, the BCT groupings appeared to make theoretical sense. Therefore, we explored the relationship between the obtained structure and an alternative structure generated by people using a simple integrated classification of theoretical constructs. We found moderate overlap between the two structures, suggesting that apparently a theoretical use of BCTs may involve implicit theory.

To increase the applications and impact of BCTTv1, work has begun to establish linkage between BCTs and behaviour change theory in order to increase the BCTTv1’s usefulness for designing theory-based interventions,45 as well as understanding effective interventions in terms of theoretical mechanisms21,24,49,159 testing theory. A large number of theories have been used to explain behaviour change, with a recent review identifying 83. 160,161 Just as developing a consensus list of BCT labels and definitions required intensive work to develop a methodology, so developing a consensus about how BCTs link to theoretical mechanisms will require intensive work to develop a methodology. This work has begun, funded by the UK’s MRC. 138

Coding of behaviour change techniques as a trainable competence

Our experience has shown that a large number of users and potential users wish to undergo formal training in the use of BCTTv1. We showed that both workshop and distance group tutorial training formats increased validity of coding BCTs in intervention descriptions. Although there is room for improvement, we showed that training by either method increased the percentage of trainees achieving an acceptable standard of coding competence as indicated by agreement with expert consensus (75% for tutorials and 45% for workshops). In study 4 (see Chapter 5), a subgroup of trained coders then went on to successfully apply their knowledge and skills to a more extensive coding exercise, achieving good reliability at baseline and at 1 month later. The likely future identification of further BCTs in addition to the already extensive nature of BCTTv1 mean that maintaining reliability and accuracy in using the taxonomy becomes more challenging. Data from the training evaluation and trainee feedback have informed the development of an open-access, online assessment and training course based on our tutorial training model96 (see Chapter 4, Methods; see www.ucl.ac.uk/health-psychology/bcttaxonomy/97). As face-to-face training, delivered either in a 1-day workshop or distance group tutorial format, is labour intensive for the trainers, the study team has been unable to meet the demand from users wanting to be trained. The online training course will allow new users to develop BCT coding competence and will allow experienced users to upgrade their skills and assess their level and pattern of competence. This remains to be evaluated and we are not yet in a position to offer training or assessment of skills in using BCTTv1 to report intervention descriptions.

Reporting interventions

We expected that training in BCTTv1 would lead to better, that is clearer, more readily replicable and more recognisable intervention descriptions. This is not what we found. Our training was as likely to make the descriptions worse as it was to improve them, although it was limited to the context of reporting interventions for which the only source of information was watching a brief video. Therefore, although the findings of our studies did not support the added value of using BCTTv1 in reporting observed interventions, the nature of the task was rather artificial and so probably not a good test. Therefore, we can make no recommendations for using the BCTTv1 at this stage and further research to investigate this is urgently needed.

However, we continue to think that this is an important objective and, as indicated in Chapter 6, there are certain avenues that we would wish to explore to achieve a better understanding of the results we obtained. There are issues to do with the design, the assessment measures and the training package we used, but perhaps a key issue is the nature of the reporting task we used. While the task involved reporting an intervention, it is not a task that would normally be undertaken in either a practice or a research setting. We would not normally produce a description of an intervention after observing it, unless it is an existing intervention in clinical practice or policy settings and no manual or protocols are available. We would be much more likely to write an intervention description before its delivery (i.e. for inclusion in the study protocol) and once intervention delivery has been completed (i.e. when reporting for evaluation and publication). In retrospect, a better test of BCTTv1 for this purpose would be to investigate whether use of BCTTv1 enhances existing intervention descriptions rather than using it to create descriptions simply from observation, or whether it enhances intervention reporting after trainees have had the opportunity to integrate the taxonomy into their normal reporting style. For example, authors of published intervention protocols or existing manuals routinely used in clinical practice or policy settings might be invited to rewrite the intervention description using BCTTv1 and research participants asked to choose the ‘better’ description from each pair of old and new descriptions.

Strengths and limitations

The work conducted in this series of studies represents a step-change in the methodology for reporting the content of BCIs. However, it also has limitations in both the methods used and the product. Constraints in terms of both time and financial resources meant that recruitment was conducted via known networks and contacts, with the result that the 400 people involved in the work were predominantly from Europe and the USA, with only a sprinkling from countries beyond this. There was also an unevenness of disciplines involved in the work, the majority being psychologists. Whilst psychologists are likely to be familiar with behaviour change concepts and therefore likely to be relatively easy to train to use the taxonomy, its wide application will require usability across a range of backgrounds and expertise. We engaged one lay participant in the study to check comprehensibility of materials. Future work should extend this to a wider range of lay perspectives and to engagement with dissemination as well as development activities.

The aim of the work was to develop and evaluate the usefulness of the BCT approach to specifying interventions, that is demonstrate ‘proof of principle’ rather than to produce a finished product and training programme. In fact, when we designed the studies we had not envisaged developing a training programme; the need for this emerged as the size of the taxonomy emerged. Given resource constraints, we decided to train participants in 24 of the most common of the 93 BCTs, as this would be likely to be most useful and familiar to them and, therefore, likely to be learned most quickly. It may be that the less-familiar BCTs are more difficult to teach to the competence criteria and would need more intensive and longer training than was provided in this project.

The training was four distance group tutorials, which is substantially less than the training that the authors’ research groups used when applying precursor taxonomies. Although this is sufficient to train the principles underlying the use of the taxonomy and to get an acceptable degree of inter-rater reliability and agreement with expert judgements, our findings show that more training is required to become expert in using the method. To make the methodology as accessible as possible, we developed an e-learning programme based on our tutorial materials and methods. 103 However, we consider that this programme, and 1-day workshops, are good introductions but unlikely, on their own, to bring participants to the level of expertise required for research and implementation. Further, coding BCTs alongside another independent coder will allow discrepancies to be identified and accuracy to be improved.

More widespread use of BCTTv1 is likely to identify ambiguities or uncertainties about labels, definitions and examples, and ways in which they can be improved. Although we are gathering this information where we can via the online training and the study website, we recognise that we require an international consortium to systematically gather and review user feedback, draw in a wider group of users, and oversee the development and refinement of BCTTv2 in a few years. This will include a detailed and transparent log of all changes with their rationales. This approach will allow the international research community to move forward in a co-ordinated fashion, which will maximise communication between groups of users and the efficiency and effectiveness with which new evidence is generated.

As mentioned earlier, there were limitations in the methods used in study 5 that limit our confidence in the findings. We aimed to evaluate BCTTv1 for writing intervention descriptions and used a variety of study designs to do this, gathering data from training workshops. We decided to show participants the intervention in video form so that we did not cue them with the written word. However, taking in detailed information about an intervention by watching an intervention is not something participants were experienced in doing. Additionally, the time pressure of the workshops meant that the videos were very brief and there was not time to show them more than once (which many participants requested). In retrospect, this experimental paradigm was probably too artificial and implemented too hastily for it to have been a good basis for the evaluation. Intervention descriptions usually emerge from reading and discussions, and a better evaluation would be to investigate the impact on this process of using BCTTv1. Given the challenge of control conditions, it is likely that future studies will seek to involve large groups of naive participants, that is those who are new to designing BCIs (e.g. students). An alternative approach, as mentioned earlier, would be to investigate intervention description enhancement than intervention descriptions de novo.

Implications for using Behaviour Change Technique Taxonomy version 1 in evidence synthesis and primary studies

Although BCTTv1 provides a means of characterising intervention content to facilitate intervention implementation, delivery and evaluation, and it is also useful for synthesising evidence. It allows both intervention and control conditions to be coded in terms of BCTs which allows the identification of the BCTs that were used in the intervention over and above those used in the control or usual-care conditions. For example, de Bruin et al. 162 found that by coding control conditions, it was possible not only to explain differences in outcomes between conditions but it was also possible to explain variation in outcomes in control conditions. The key point here is that many active controls such as ‘usual care’ involve implementation of multiple BCTs, overlapping with those that constitute the intervention itself.

Behaviour change technique methodology is a starting point for investigating the effectiveness of single BCTs and the combinations of BCTs that typically characterise interventions in evidence synthesis. It is also a starting point for investigating how other factors, such as mode of delivery, intervention intensity, target behaviour, target population and context, may make BCTs more or less effective. 163–165 However, given the typically small number of studies that have used any particular BCT and potential influence of confounders, there are limits to this methodology. Meta-analyses lack power to test the effectiveness of most BCTs, as was evident in the National Institute for Health and Care Excellence (NICE) 2014 Behaviour Change guidance which used BCT methodology. 166 Therefore, it is important to be careful not to conclude that a BCT is ineffective when the conclusion should be that there is insufficient evidence to test for effectiveness. However, the importance of BCT methodology for advancing the field is reflected in the research recommendations within NICE’s Behaviour Change Guidance. 167 NICE recommend that research should seek to investigate which combinations of BCTs and which modes of delivery are effective and cost-effective in (1) changing behaviour and (2) in maintaining behaviour change. They recommend seeking to determine how effectiveness varies among people from different sociodemographic backgrounds, those with different skill sets, levels of motivation or with access to different information. Finally, research should include studies that build the evidence base on the effectiveness of each BCT, e.g. using experimental and meta-analytic work to clarify which BCTs work when and for whom.

Investigating behaviour change technique combinations in evidence synthesis and experimental studies

There are several examples of meta-analyses which have investigated both individual BCTs and combinations of BCTs that theory predicts would work synergistically together. 21,24,49,168,169 For example, in reviews of interventions to increase physical activity and healthy eating, Michie et al. 24 and Dombrowski et al. 21 investigated a combination of BCTs predicted by control theory170 and found similar results in two different populations. They found that interventions with the combination of self-monitoring, goal-setting and action planning were twice as effective as those that were not. A similar analytic approach was used to investigate audit and feedback interventions168 in a recent Cochrane review,49 finding that adding the BCTs of goal-setting and action planning to interventions increased the effect of feedback. The theoretically based combination of BCTs associated with provision of information, increasing motivation and enhancing behavioural skills171 has been found to reduce the frequency of sexual interactions that increase health risk. 169

An alternative method of combining BCTs is to use CARTs, referred to as Meta-CART,39 to analyse meta-analytic data. This has been used to identify groups of BCTs to predict intervention effects in a reanalysis of data from Michie et al. 23 Results showed that providing information about the links between behaviour and health was effective if combined with either setting goals or with providing information on the consequences of the behaviour and using follow-up prompts.

Very brief interventions offer a pragmatic context in which to test whether specific BCTs add to or dilute intervention effectiveness as they tend to include only a small number of BCTs. Such work is being conducted as part of a research programme into very brief interventions to promote physical activity in primary care. 172 A similar programme of research, supported by a meta-analysis,173 has identified BCTs in very brief advice interventions to promote smoking cessation. 174

With increasingly sophisticated designs, experimental methods can begin to unpack the ‘black box’ of complex interventions. For example, Collins et al. 175 have proposed an experimental paradigm, known as the MOST approach (the Multiphase Optimisation Strategy approach) that tests combinations of BCTs as a basis for optimising interventions. Using fractional factorial designs, they advocate selecting combinations of BCTs for testing based on both theory and accumulated evidence. These methods have the advantage that they not only provide evidence of effective combinations, but they also efficiently test theoretical propositions about the synergistic effects of constructs.

Behaviour change techniques and theory

Behaviour change techniques are technologies and in themselves are agnostic with regard to the role of theories. In most cases, there is no attempt to link specific BCTs to theoretical mechanisms. For example, a meta-analysis of 190 interventions to increase physical activity and healthy eating interventions found that 56% of studies claimed the interventions were based on a theory. 58 However, 90% of these did not report links between all of the BCTs and specific theoretical constructs. 58 When links are made, this is usually the result of the researchers’ judgement rather than being based on evidence, pilot work or a systematic methodology of establishing this. Establishing links between BCTs and theoretical constructs (their mechanisms of action) is a much needed step to advance the science of behaviour change. It is needed to translate theory into intervention design,43,45,161 to explain effective interventions in terms of their theoretical mechanisms,21,24,49,159 to aid practitioners and policy-makers in selecting effective BCTs to target specific behaviours, and to test theory. A large number of theories have been used to explain behaviour change, with a cross-disciplinary review identifying 83. 160,176 Just as developing a consensus list of BCT labels and definitions required intensive work to develop a methodology, so developing a consensus about how BCTs link to theoretical mechanisms will require intensive work to develop a methodology. Developing a consensus-based methodology to establish BCT theory links is the aim of a follow-up study to the current BCT study, funded by UK’s MRC (Michie S, Carey R, Johnston M, Rothman A, Kelly M, de Bruin M, et al. University College London, 2015, personal communication). This forms the basis of a larger programme of work, developing a Behaviour Change Ontology identifying the links between five levels of intervention characteristic: target behaviours, BCTs, theoretical mechanisms, modes of delivery and contexts. Given the enormity and complexity of the task, this requires collaboration between behavioural and computer scientists. By using computational machine learning methods of literature interrogation and data mining, it is possible to develop a virtual rapid learning environment to (1) optimise BCIs in the real world and (2) create an updatable open source intervention library for users to identify which combinations of intervention characteristics are most relevant to, and likely to be most effective for, their own behavioural targets and contexts.

Behaviour change techniques, replication and implementation

Given that published reports and existing manuals can effectively be coded into BCTs, there are likely to be benefits of using BCT methodology. These include intervention development, evidence synthesis, assessing and improving the replication of interventions within scientific investigation and the faithful delivery of interventions and treatments in practice. Behavioural interventions to increase fidelity of intervention delivery have the potential to bring about as great, or greater, health gains than medical advances on their own. 177 The importance of scientific journals ensuring that intervention content is well specified for replication and implementation has been recognised across the research community. 15 At least two journals have editorial policies stating that trials of interventions will only be published if the interventions are described in sufficient detail for replication and implementation, citing BCT methodology as an example. 10,178 Guidance by funding bodies would also support the adoption of BCT methodology for specifying and reporting BCIs.

There is ample evidence from reviews and primary studies that intervention protocols are often implemented with insufficient fidelity. 12,179–182 For example, in an intervention to increase physical activity among sedentary adults at risk of type 2 diabetes mellitus delivered by trained and quality-assured facilitators, 58% of BCTs specified in the intervention protocol were not faithfully delivered. 182 This was particularly the case for BCTs directed at maintenance of behaviour change. 183 Coding the delivery of BCTs using observations, audio or video recordings, of interventions allows researchers and practitioners to find out how many of the intervention’s proposed active ingredients are actually delivered. Such information can be used to improve the training of those who deliver interventions and their long-term implementation. It also aids the interpretation of trial results, e.g. the extent to which any lack of effects is due to the intervention itself or poor fidelity of delivery. BCTTv1 may also be used to inform the development of measures for assessing delivery of BCTs. 31

Behaviour change techniques and designing interventions

Having access to the 93 BCTs of BCTTv1 allows intervention designers to consider a large range of BCTs that might be effective and appropriate for their target behaviour and population. An example of a step-by-step guide for designing BCIs incorporating BCTs is The Behaviour Change Wheel: A Guide to Designing Interventions. 161 It links BCTs to intervention functions with the behaviour change wheel framework and to the 14 theoretical domains of the TDF. Intervention designers are encouraged to start their selection of BCTs by conducting a ‘behavioural diagnosis’ of the problem at hand. Because 93 BCTs are likely to be too many to easily work with, we recommend starting with a minimum set composed of the 22 frequently occurring BCTs and/or those shown to be effective in the specific area under investigation. These 22 found to frequently occur in 40 intervention descriptions include 18 of the 22 BCTs including in the Abraham and Michie taxonomy20 based on more than 200 intervention descriptions. Nevertheless, coders in study 4 used 80 BCTs to describe the 40 intervention descriptions and it may therefore be important to expand from the initial 22 BCTs to fully specify the intervention. Another issue is whether or not to use the earlier, shorter taxonomies that have been developed for specific behavioural domains,23,27,29,30 given that these BCTs will have high relevance and include the frequently occurring BCTs for that domain. This needs to be balanced with the potential problem that the scientific study of behaviours sometimes occurs in ‘silos’ so that one may miss potentially effective BCTs that are used in other behaviours. For example, ‘behavioural substitution’ was identified from a systematic review of brief alcohol interventions,29 but this effective BCT had not been identified in several reviews of physical activity, healthy eating and smoking cessation interventions, e.g. Carroll et al. 183 A second limitation of sticking with behaviour domain taxonomies is that it limits generalisation of all potential BCTs across behaviours and, therefore, the possibility of generating general models of behaviour change or investigating ways in which BCTs influence different behaviours. Finally, in contrast to earlier taxonomies, BCTTv1 was developed using robust methods and is based on expert consensus, and BCTs in earlier taxonomies (e.g. motivational interviewing) were removed from BCTTv1 as they did not meet the definition of a BCT.

Need for maintenance and updating

The overarching aim of developing BCTTv1 was to maximise the co-ordinated building of evidence and hence the rate at which we can develop more effective BCIs to improve health and well-being, and quality of health care. The aim is to achieve maximum consensus across disciplines, topic areas and countries in order to synthesise and share evidence more efficiently across them. BCTTv1 was termed v1 for a purpose. As BCTTv1 is applied to a wider range of populations, settings and behaviours, adaptations of language and possibly concepts will be needed and new BCTs will be identified. Wider application also involves delivering interventions at different ‘levels’ (e.g. individual, community, organisational, population), as illustrated in Table 18. As BCTTv1 is used to design and specify the content of interventions across wider ranges of delivery modes and contexts, it is likely that additional BCTs will be identified.

Therefore, we need an international consortium, appropriately resourced, to monitor and collate experiences, adaptations and findings so that BCTTv2 can be developed and released. The timing of this will depend on judgement: balancing the needs for stability and accumulation of evidence using a shared method with the weight of evidence of need for refinement and extension. Discussions are under way with international funding agencies to achieve this.

Final conclusion

The BCTTv1 is a technological advance that represents a step change in the translation of behavioural science into practice and in strengthening the science itself. Like all technologies, its usefulness will be determined by its application and dissemination by researchers, those designing and delivering interventions, funding agencies, journal editors and policy-makers.

Recommendations

The findings from this programme of research point to several recommendations for practice and future research. These are listed below in priority order, starting with the highest priority.

Recommendations for practice:

• Characterising interventions – to facilitate accuracy, ease and speed of applying BCTTv1 to characterising intervention content, we recommend that users start their coding or other task (e.g. evidence synthesis, reporting interventions) with the list of the 22 most frequent BCTs (see Table 12). If behaviour-specific taxonomies are used, we suggest supplementing them with additional BCTs from BCTTv1.

• Intervention design – we recommend considering the full range of BCTs in BCTTv1 in the design process, with selection of BCTs guided by theoretical and pragmatic criteria (see the behaviour change wheel guide161).

• Evidence synthesis – BCTTv1 should be used to specify BCTs in both the active and the control arms of the evaluation trial. The effects of combinations of BCTs can be investigated, e.g. by theoretically informed metaregression and/or Meta-CART analyses.

• Implementation – to assess and improve implementation (i.e. delivery that is faithful to the protocol), those reporting interventions should specify the content in terms of BCTs and assess delivery, using reliable methodology.

Recommendations for research:

• Structure of BCTTv1 – a consensus concerning the linkages between BCTs and behaviour change theory is needed to increase BCTTv1’s usefulness for designing theory-based interventions, understanding interventions and testing theory. Work is already under way to address this need (Michie S, Carey R, Johnston M, Rothman A, Kelly M, de Bruin M, et al. University College London, 2015).

• Reporting interventions – further research is needed to understand the potential usefulness of using BCTTv1 to report BCIs.

• Understanding effects of BCTs – we recommend that systematic reviews with meta-analyses are developed for each BCT (for those which this has not already been done), starting with the most frequent BCTs.

• User training – we recommend that further research is needed to evaluate user training, for all 93 BCTs in BCTTv1. This will help to increase understanding as to which BCTs require more/less training and will inform improvement of future training programmes.

• BCTTv2 – precise documentation of the adaptations of BCTTv1 for specific settings, behaviours and populations is required to inform the development of BCTTv2.

Contributions of authors

Susan Michie (Professor of Health Psychology and principal investigator) and Marie Johnston (Professor of Health Psychology) led the development of the research proposal and protocol.

Susan Michie, Caroline E Wood and Marie Johnston wrote the first draft of this monograph.

Marie Johnston, Caroline E Wood and Charles Abraham conducted statistical analyses for studies 3, 4 and 5a, 5b, 5c.

Wendy Hardeman and Jill Francis contributed to the development of the study proposal, overseeing the studies and writing the final monograph.

All authors contributed to the development of the study proposal, overseeing of the studies and the writing of the final monograph.

Contribution of others

Statistical analyses for each of the studies were carried out as follows: study 1 by Michelle Richardson (Senior Research Fellow); study 2 by Ruhina Lahda (Research Assistant), James Cane (Lecturer; Senior Research Fellow) and Michelle Richardson; study 3 by Caroline Wood (Senior Research Associate); study 4 by Caroline Wood and Charles Abraham (Professor of Health Psychology); and studies 5a, 5b and 5c by Caroline Wood and Marie Johnston. Statistical guidance was provided by Beth Pollard (Statistical Expert).

Publications

Michie S, Richardson M, Johnston M, Abraham C, Francis J, Hardeman W, et al. The behavior change technique taxonomy (v1) of 93 hierarchically clustered techniques: building an international consensus for the reporting of behaviour change interventions. Ann Behav Med 2013;46:81–95.

Cane J, Richardson M, Johnston M, Lahda R, Michie S. From lists of behaviour change techniques (BCTs) to structured hierarchies: comparison of two methods of developing a hierarchy of BCTs. Br J Health Psychol 2015;20:130–50.

Wood CE, Richardson M, Johnston M, Abraham C, Francis J, Hardeman W, et al. Applying the Behaviour Change Technique (BCT) Taxonomy v1: a study of user training. Transl Behav Med 2014;5:134–48.

Abraham C, Wood CE, Johnston M, Francis J, Hardeman W, Richardson M, et al. Reliability of identification of behavior change techniques in intervention descriptions. Ann Behav Med 2015 [published online ahead of print 20 August 2015].

Data sharing statement

All available data can be obtained from the corresponding author of this report.

Disclaimers

This report presents independent research funded under a MRC–NIHR partnership. The views and opinions expressed by authors in this publication are those of the authors and do not necessarily reflect those of the NHS, the NIHR, the MRC, NETSCC, the HTA programme or the Department of Health. If there are verbatim quotations included in this publication the views and opinions expressed by the interviewees are those of the interviewees and do not necessarily reflect those of the authors, those of the NHS, the NIHR, the MRC, NETSCC, the HTA programme or the Department of Health.

Intervention content

The telephone intervention script is designed to help parents modify their home food environments through addressing three key domains listed in Table B. The first column of the table lists each domain at the point at which it appears in the schedule of support calls, while the second column lists the specific topics that are used to explore each of the given domains. Each domain has been associated with increased fruit and vegetable consumption in children as described below.

Intervention resources

Based on evidence indicating telephone-based dietary interventions are more effective when used in conjunction with print and other resources, all intervention participants will be mailed resource kits following completion of the baseline survey. The kit comprises a participant workbook containing information and activities, a pad of meal planners, and a cookbook including recipes high in fruit and vegetables. The resources will be used to facilitate participant engagement in the telephone support calls and assist participants to complete intervention activities between telephone contacts.

Conceptual model

The telephone-based intervention accords with the model of family-based intervention proposed by Golan et al. (2001) in the treatment and prevention of childhood obesity. Their model, which draws upon socioecological theory, focuses on introducing new familial norms associated with healthy eating. This is achieved through making changes within the home food environment, providing positive parental role-modelling and increasing parenting- and nutrition-related knowledge and skills. Interventions based on such a model have been shown to be effective in bringing about environmental changes in participants’ homes to support healthy eating and in reducing poor eating habits of overweight and obese children of participants.

The intervention utilises a number of specific behaviour change techniques to initiate the change process as described in Table A. The third column lists the behaviour change techniques used and the fourth column links each technique to its application in the context of the topic listed in column 2. These behaviour change techniques include prompting intention formation, barrier identification, specific goal-setting and the reviewing of such goals, self-monitoring of behaviour and identification as a role-model, teaching to use prompts or cues, and providing general encouragement, as described in the taxonomy proposed by Abraham and Michie (2008).

Excerpt from: Spijkers et al. 122 © 2010 Spijkers et al. ; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution and reproduction in any medium, provided the original work is properly cited.

Intervention

The intervention to be evaluated is Triple P level 3. Triple P is a multilevel system of family intervention which provides five levels of intervention of increasing strength. Triple P intervention at level 3 (Primary Care Triple P) is a brief, narrow-focus parent programme that is aimed at parents with specific concerns about their child’s behaviour or development. It combines advice, rehearsal and self-evaluation to teach parents to manage discrete child problem behaviour during four individual consultations of 20–30 minutes with the parents and their child (Table C).

Extract from: Colagiuri et al. 124 © 2010 Colagiuri et al. ; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution and reproduction in any medium, provided the original work is properly cited.

Intervention – Lifestyle Modification Program

The five aims of the lifestyle modifications are:

1. At least 30 min/day of moderate to vigorous intensity physical activity, including aerobic exercise 3 or more days/week plus strength training at least twice/week (210 min/week total structured exercise)

2. Reduction in the intake of energy from total fat to less than 30%

3. Reduction in the intake of energy from saturated fat to less than 10%

4. Fibre intake of at least 15 g/1000 kcal

5. Achievement of a 5% reduction in body weight at 12 months.

In addition to the 210 min/week structured exercise goal, participants are encouraged to increase incidental physical activity in ways which would enhance both cardiovascular and musculoskeletal fitness.

These five goals are entirely concordant with the Finnish DPS [Diabetes Prevention Service] which was one of the most successful diabetes prevention trials. The physical activity goal, which has been modified slightly from the Finnish DPS is based on a review of the physical activity prescriptions utilised in relation to outcomes achieved in all of the successful trials of diabetes prevention, considerations of cost and feasibility in this translational setting, as well as other literature regarding modality, volume, and intensity of exercise required to improve metabolic risk and body composition in similar cohorts. Both the Finnish DPS and the US DPP [Diabetes Prevention Programme] included resistance training (strength training) in their supervised exercise sessions and is explicitly specified within the physical activity goal of the SDPP [Sydney Diabetes Prevention Programme]. Resistance training is an anabolic form of exercise, differing substantially from aerobic exercise in its ability to induce muscle hypertrophy and associated metabolic and functional changes. It improves insulin sensitivity, glucose homeostasis, blood pressure, dyslipidaemia, markers of inflammation and catabolism, and visceral obesity, thus addressing the key metabolic abnormalities in adults at high risk of type 2 diabetes. Importantly, resistance training (but not aerobic exercise) attenuates or prevents the loss of lean tissue (muscle and bone) accompanying weight loss diets such as those prescribed in this study.

The behavioural components are based on stages of change and social cognitive theories. The intervention is delivered by dedicated program lifestyle officers from a variety of health backgrounds including dietetics, nursing, psychology and exercise physiology. The lifestyle officers undergo specific training in health coaching, group program delivery and standardized data collection used for evaluation. The health coaching approach incorporates principles from self-management, removing psychological blocks to change and confidence.

High risk individuals agreeing to participate in the lifestyle modification program complete an initial computer assisted telephone interview (CATI) survey. This survey includes socio-economic and demographic information, physical activity habits, quality of life and self-efficacy, as well as recent health service utilisation and current medication use. Participants are then scheduled to attend an individual consultation with a lifestyle officer. At this consultation, the lifestyle officers measure height, weight and waist circumference using calibrated stadiometers, scales and tape measures, following a standardized anthropometric protocol as specified by the International Society for the Advancement of Kinanthropometry (ISAK). The individual consultation includes a general discussion about diabetes risk and prevention, an overview of the program, and uses motivational interviewing techniques to assist participants to set goals and develop tools to self-monitor. Following this session, arrangements are made for participants to attend three two-hour group programs held over a six to eight week period. Lifestyle officers conduct these group sessions of approximately 10 people, which cover theoretical, behavioural and practical aspects of diet and physical activity. The overall program motto is: ‘Eat better and move more’. Those who are not able to or do not want to attend a group program are offered the option of three individual health coaching sessions by telephone, covering the same material. The intervention delivered to indigenous participants will be slightly modified to take account of cultural issues.

Follow up telephone calls are made by the lifestyle officers to each participant at 3, 6 and 9 months to enquire about progress, assist with behaviour change and offer participants additional support as required. In addition, participants are provided with details of local community-based lifestyle programs which have been evaluated by the research staff and found to be consistent with the goals of the SDPP. Participants have the option of enrolling in such programs as one way to assist in achieving the SDPP physical activity and dietary goals.

At 12 months the CATI survey is repeated and participants undergo an individual assessment with the lifestyle officer and their general practitioner.

Extract from: Johnston et al. 129 © 2010 Johnston et al. ; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution and reproduction in any medium, provided the original work is properly cited.

Treatment group

The intervention program will (i) provide information and education about the health effects of ETS [environmental tobacco smoke] exposure and use behavioural ‘coaching’ techniques to help mothers/caregivers and family members implement strategies to reduce the infant’s ETS exposure, as well as (ii) identify the smokers among other household members and deliver culturally appropriate smoking cessation advice, counseling and treatment options as requested. An eight weeks supply of free nicotine replacement therapy (NRT) (patches or gum) will be available to participants and other household members for whom such drug therapy is indicated (i.e. they are motivated to quit, are nicotine dependent and have no contraindications to taking NRT). NRT will be provided by the IHW [indigenous health worker] with appropriate counseling and follow-up. Furthermore, for those that are interested a fax referral to Quitline will be offered, with proactive call back by Quitline.

The intervention program will be delivered during three face-to-face home visits (of approximately 45–60 minutes) conducted over the first three months of the infant’s life. Culturally appropriate resources (e.g. flip charts, ‘No Smoking’ stickers, posters, etc.) will be used to assist in both education and behavioural ‘coaching’. These resources will be obtained from relevant health groups in each country who hold a repository of such resources (e.g. QUIT Victoria, the Northern Territory Department of Health and Families, Auckland Regional Public Health Service). IHWs will deliver the program after appropriate training, and will complete standardized progress reports after each program session, which will be used at a weekly team meeting with the health workers and study personnel for discussion and ongoing training.

Extract from: Taylor et al. 130 © 2010 Taylor et al. ; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution and reproduction in any medium, provided the original work is properly cited.

Phase 2: Treatment

Study procedures

Intervention

The TECNOB clinical program has a total duration of 13 months and consists of two stepped down phases: inpatient (1 month) and out-patient (the following 12 months). During the in-patient phase, participants undergo an intensive four-week hospital-based and medically-managed program for weight reduction and rehabilitation. Along this period, participants live in a medical hospital-like environment located on a mountain highland and far away from towns and cities. Visits from parents are allowed only in the afternoon. All patients are placed on a hypocaloric nutritionally balanced diet tailored to the individual after consultation with a dietitian (energy intake around 80% of the basal energy expenditure estimated according to the Harris-Benedict equation and a macronutrient composition of about 16% proteins, 25% fat and 59% carbohydrates). Furthermore, they receive nutritional counseling provided by a dietitian, psychological counseling provided by a clinical psychologist and have physical activity training provided by a physiotherapist.

Nutritional rehabilitation program aims to improve and promote change in eating habits and consists of both individual sessions (dietary assessment, evaluation of nutrient intake and adequacy, nutritional status, anthropometric, eating patterns, history of overweight, readiness to adopt change) and group sessions (45 minutes each twice a week) including: information on obesity and related health risks, setting of realistic goals for weight loss, healthy eating in general, general nutrition and core food groups, weight management and behavior change strategies for preventing relapse).

Psychological counseling is provided once a week both individually and in group setting. Individual sessions, lasting 45 minutes each, are mainly based on the cognitive-behavioral approach described by Cooper and Fairburn (1990) and emphasize the techniques of self-monitoring, goal setting, time management, prompting and cueing, problem solving, cognitive restructuring, stress management and relapse prevention. Group sessions (‘closed’ groups of 5/6 persons), lasting 1 hour each, focus on issues such as motivation, assertiveness, self-esteem, self-efficacy and coping.

Physical activity takes place once a day except for weekend and consists of group programs (20 subjects) based on postural gymnastics, aerobic activity and walks in the open. Inpatients with specific orthopedic complications carry out individual activities planned by physiotherapists and articulated in programs of physical therapy, assisted passive and active mobilization and isokinetic exercise.

Low to moderate weight losses are expected at the end of the in-patient phase, but it is important to note that weight loss is not the primary goal of the in-patient program and each patient is made clear about this point at the very beginning of the treatment. Beyond the medical management of metabolic risk factors for health such as type 2 diabetes, developing a sense of autonomy and competence are the primary purposes of the in-hospital interventions. Patients are afforded the skills and tools for change and are supported in assigning positive values to healthy behaviours and also in aligning them with personal values and lifestyle patterns.

In the last week, just before discharge from hospital, participants are instructed for the out-patient phase of the program. They receive a multisensory armband (SenseWear® Pro2 Armband), an electronic tool that enables automated monitoring of total energy expenditure (calories burned), active energy expenditure, physical activity duration and levels (METs) and sleep/wake states duration. Patients are instructed to wear this device on the back of the upper arm and to record data for 36 hours every two weeks in a free-living context. The Armband holds up to 12 days of continuous data which the outpatients are instructed to download into their personal computer and to transmit online to a web-site specifically designed for data storing. Outpatients are also told that they can review their progress using the InnerView® Software which analyzes and organizes data into graphs and reports. Participants are then instructed to use the TECNOB platform, an interactive web-site developed by TELBIOS S.P.A. The TECNOB web-platform supports several functions and delivers many utilities, such as questionnaires, an animated food record diary, an agenda and a videoconference virtual room. In the ‘questionnaires’ section, patients fill in the Outcome Questionnaire and submit data concerning weight and glycated hemoglobin. In the ‘food record diary’ participants submit actual food intake day by day through the selection of food images from a comprehensive visual database provided by METEDA S.P.A. The same procedure is also possible through a software program called METADIETA (Meteda s.p.a.) previously installed on the outpatients’ mobile phones before discharge. Through the mobile phones outpatients maintain the contact with the dietitian who regularly sends them SMS containing syntax codes that METADIETA, the software previously installed into the outpatients’ mobile phones, uses in order to visually display the food choices (frequency and portions) outpatients have to adhere according to diet prescriptions. By this way, outpatients can keep a food record diary allowing comparisons between current eating and the recommended hypocaloric diet along the whole duration of the program. The ‘agenda’ allows the patients to remember the videoconference appointments with the clinicians and the days when to fill in the questionnaires. Moreover, the patients can use the ‘memo’ space to note down any important event occurred to him/her in the previous week/month. Indeed, some research indicates that changes in behaviour (eating and exercise) often follow discrete moments which have been variably described as life events, life crises, teachable moments or epiphanies. Life events can lead to weight loss but also to weight gain and qualitative research shows that it is not the event per se that results in behaviour change but the ways in which this event is appraised and interpreted by the individual. The clinical psychologist has thus the opportunity to discuss with the outpatients about the significant events reported in the ‘memo’ space during the videoconference sessions and cognitively reconstruct dysfunctional appraisals in functional ways. Finally, outpatients are instructed to use the videoconference tool. Thanks to this medium, they receive nutritional and cognitive-behavioural tele-counseling with the dietitian and the clinical psychologist who attended the patients inside the hospital. In particular, just after discharge, participants have 6 videoconference contacts with both clinicians along 3 months. From the 3rd to the 6th month sessions are scheduled every 30 days and then even more spaced up to an interval of 60 days. During tele-sessions, clinicians (psychologist and dietitian) test the outpatients’ progress, their mood, the maintenance of the ‘good alimentary and physical activity habits’, the loss/increase of weight and ask about critical moments, especially those ones reported on the ‘memo’ web-space. In particular, tele-sessions with the clinical psychologist aim to consolidate strategies and abilities acquired during the in-patient phase, to improve self-esteem and self-efficacy, to support motivation, to prevent relapse and to provide problem-solving and crisis counseling. On the other hand, dietitian assesses adherence and compliance to dietary therapy with a special focus on normal eating behaviour, sufficient fluid intake, hunger and fullness regulation, appropriate eating/etiquette (pace and timing of meals), slow rate of eating, and addresses critical points such as plateau in weight loss or lack of readiness to improve dietary habits.

In addition to videoconference, outpatients can further contact clinicians by e-mail. Indeed, each patient is given the possibility to join his clinician beyond the established videoconference contacts in case of urgency or emergency. According to the e-message’s content, clinicians choose the most appropriate format for delivering feedback among e-mail or telephone. In order to avoid excessive dependence and to contain costs, a maximum number of 1 non-scheduled contact per week is established a priori.

As described, in the outpatient phase of the TECNOB program great relevance is given to the clinicians–patient relationship as an important medium and vehicle of change. After discharge, out-patients begin to experience the autonomy and competence to change they develop during the in-patient phase and inevitably face resistances and barriers. Thanks to videoconferences, out-patients are supported by the clinicians who attended them during the in-hospital phase in exploring resistances and barriers they experience and in finding functional pathways to cope. Furthermore, out-patients are helped to experience mastery in terms of the health behaviour change that needs to be engaged.

Extract from: Wyers et al. 117 © 2010 Wyers et al. ; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution and reproduction in any medium, provided the original work is properly cited.

Nutritional intervention

The nutritional intervention is a combination of dietetic counseling and oral nutritional supplements for three months. The intervention starts during hospital admission and continues after discharge during the stay at the rehabilitation clinic or at the patient’s home. During hospitalization, the study dietician visits the patient twice. At the first visit, two to five days after surgery and immediately after baseline measurements, the dietician interviews the patient regarding medical and social status, and pre-fracture mobility. The dietician also performs a 24-hour recall and takes a general dietary history of the patients’ diet before hospitalization. Next, the patient receives the nutritional supplement, a milk-based supplement providing 21 kJ (500 kcal) and 40 g of protein. The dietician advises the patient on the consumption of the supplement and arranges extra care or services to optimize the food intake if necessary. Before hospital discharge, the dietician visits the patient for the second time. During this visit, a 24-hour recall is performed and the consumption of the nutritional supplement is evaluated. Furthermore, arrangements are made to continue the dietetic advice and the consumption of the nutritional supplement at home or during the stay at the rehabilitation clinic. At home or during the stay in a rehabilitation clinic, the dietician visits the patient three times (one week, two weeks and six weeks after discharge) and makes five telephone calls with the patient (three, four, five, eight and ten weeks after discharge). During these visits, food intake and supplement use is assessed by a 24-hour dietary recall, and tailor-made dietetic advice is given to optimize the amount and composition of the diet. As soon as the patient meets nutritional requirements with a normal diet, the use of the nutritional supplement is stopped. Compliance with the nutritional supplement is evaluated by the 24-hour dietary recalls, patients’ registration of the consumption in a diary and by collecting the capsules of the cans of the nutritional supplement during the home visits.

Extract from: Siddiqi et al. 120 © 2010 Siddiqi et al. ; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution and reproduction in any medium, provided the original work is properly cited.

The Intervention and its key components

FIVE STEPS TO QUIT – The Intervention Model: This is based on the evidence-based recommendations for treatment of tobacco addiction published by WHO in 2001. The same approach is being advocated by IUATLD, National US guidelines and NICE in the UK. This model relies on assessing personal motivation to quit tobacco use and uses it as the basis for assessing suitability for the different therapeutic options for tobacco dependence. Thus, the approach maximizes the efficient use of nicotine replacement therapy (NRT) and bupropion.

Key Components: We would like to develop components of the ‘Five Steps to Quit’ intervention model on the basis of the following principles:

• Based on best available evidence and following an approach recommended by international agencies such as WHO and IUATLD.

• Can be delivered in primary care setting by non-specialist health care staff (doctors or other non-medical personnel) integrated in their routine healthcare provision.

The intervention model consists of five key steps

1. Asking about the status of nicotine use;

2. Advising about the benefits of stopping nicotine use;

3. Assessing the motivation to stop its use;

4. Assisting in stop attempts through various therapeutic options; and

5. Arranging follow-up

Examples of the specific components for each step of the intervention are given in Appendix A; this is a general guide only and modifications will be made during the development of the intervention.

Delivery of the intervention

We propose to use a systematic, standardised approach to deliver ‘five steps to quit’ to make it effective and equitable. It is envisaged that the intervention will be primary delivered by the TB DOTS facilitator based in the diagnostic centre with the help and under the supervision of the primary care doctor (Table E). He will assess patients’ eligibility for the study and send them to the TB DOTS facilitator for further assessment. All eligible patients will be provided with verbal and written information about the study and invited to participate. Patients who agree to take part will be taken through different components of the ‘five steps to quit’ programme in two appointments. Patients in intervention arm 1 will also be offered therapeutic option (Bupropion) and such patients will be referred to the primary care doctor for assessing suitability and prescribing Bupropion.

In the intervention arm 1, the primary care doctors, in addition to supervising TB DOTS facilitators, will be responsible for assessing and prescribing therapeutics (Bupropion) to patients who wish to quit and are being referred by the TB DOTS facilitators. All patients on treatment will have an additional follow up visit at week 8. In intervention arm 2, doctors will be only responsible for supervising TB DOTS facilitators. They will be provided with the appropriate training and relevant materials.

Apart from the relevant training and materials, health professionals will not be provided with any other incentive, financial or otherwise. However, treatments packs will be provided from the research budget. The research officer will also organise regular supervisory visits to oversee patient management and data collection procedures. Monthly cluster meetings of the health professionals involved will also be organised to discuss progress and potential problems.

Appendix A

Extract from: Mc Namara et al. 121 © 2010 Mc Namara et al. ; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution and reproduction in any medium, provided the original work is properly cited.

Intervention part two: community pharmacist facilitating patient change

Community pharmacists will be trained to deliver their interventions in accordance with the Health Action Process Approach (HAPA) to behaviour change over five counselling sessions conducted at monthly intervals. The emphasis in counselling progresses from change motivation initially (via improved self-efficacy, belief in the need for change and belief that change will generate positive outcomes), through to change maintenance and relapse prevention strategies. Written, achievable goals will be encouraged.

The first session with the patient will prioritise basic health education regarding individual CVD risk and the benefits of potential treatments. It also establishes acceptable goals for the treatment process through patient consultation, and how these might be achieved. Finally, the pharmacist will discuss with the patient any specific medication changes identified in the baseline report that are recommended to improve adherence to CVD guidelines. Community pharmacists will not be trained or asked to make interventions related specifically to diabetes or mental health issues, but will be alerted to any suboptimal assessment results in these areas and asked to discuss with the patient the potential need for GP input. Such issues will also be identified in the baseline assessment summary provided to the GP.

If a patient’s overall 5-year CVD risk score is 5% or less (considered very low risk) they will be advised to discuss with their pharmacist whether they are likely to benefit from continuing with the intervention. The decision to continue will be left to the pharmacist and the patient. Pharmacists will be expected to assess and document patient motivation to undertake various medication and lifestyle changes. Following discussion with each patient, the pharmacist will then forward the clinical summary to the patient’s GP with any additional comments considered relevant.

Subsequent sessions will involve: ensuring necessary changes to medicines have been made; monitoring of medicines adherence especially for new medicines; linking patients with local health and other services that provide relevant patient support; initiating lifestyle change and supporting maintenance and relapse prevention. Throughout these sessions, patient progress towards goals will be continually reassessed, as will be the goals themselves. GP input to patient treatment plans will also be invited.

Extract from: McAlister et al. 125 © 2010 McAlister et al. ; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Experimental arms

Pharmacist Care Group (PCG)

The PCG participants receive a package of interventions from the pharmacist for enhancing their antihypertensive medication adherence, which includes:

• A home BP monitor (Omron®T9IT) with the capacity to store and download BP readings to be used for discussion at three- and six-month follow-ups;

• Training by the pharmacist on self-monitoring of BP;

• Motivational interviewing and education by the pharmacist to help patients improve their medication adherence and achieve target BP;

• Pharmacist-initiated home medicines review (HMR), dose administration aid (DAA) and/or patient medication profile (PMP), where necessary;

• Medication use review (MUR) to identify and resolve possible medication-related hypertension (e.g. due to non-steroidal anti-inflammatory drugs, cold preparations, complementary medicines, etc.);

• Referral to a GP when needed (e.g. very high blood pressure); and

• Refill reminders (by either SMS, telephone or mail) from their pharmacist at a chosen number of days before their antihypertensive medication dispensing is due.

HMRs are designed to assist consumers living at home to maximise the benefits of their medication regimen and prevent medication related problems. The review involves the consumer’s general practitioner (GP) and preferred community pharmacy, and in some cases other relevant members of the healthcare team. The GP refers the consumer to the community pharmacy and an accredited pharmacist visits the consumer at home, reviews their medication regimen, and provides the GP with a report. The GP and consumer then agree on a medication management plan. A DAA is a device developed to assist patients in better managing their medicines by arranging their medicines into individual doses according to the prescribed dose schedule throughout the day. The aim of the DAA Program is to reduce medication-related hospitalisation and adverse events through improving medication management and adherence for people in the community. A PMP is a comprehensive written summary of all regular medicines taken by a patient that assists them in understanding and managing their medicines by informing them how, when and why to take their medicines. The aim of the PMP Program is to reduce the risk of medication-related adverse events by assisting people to understand and manage their medications, including prescription, over-the-counter and complementary medicines. A MUR usually takes place in the pharmacy and it involves the pharmacist checking the patient’s medication, making sure that the patient knows how and why they should be taking their medication, as well as identifying any problems. It provides the patient with an opportunity to ask questions and the pharmacist an opportunity to improve the patient’s medication understanding and adherence, as well as being able to highlight problems and provide appropriate solution.