Cane et al.

Implementation Science 2012, 7:37

http://www.implementationscience.com/content/7/1/37
Implementation
Science

RESEARCH Open Access

Validation of the theoretical domains framework

for use in behaviour change and implementation
research
James Cane1, Denise O’Connor2 and Susan Michie3*

Abstract
Background: An integrative theoretical framework, developed for cross-disciplinary implementation and other
behaviour change research, has been applied across a wide range of clinical situations. This study tests the validity
of this framework.
Methods: Validity was investigated by behavioural experts sorting 112 unique theoretical constructs using closed
and open sort tasks. The extent of replication was tested by Discriminant Content Validation and Fuzzy Cluster
Analysis.
Results: There was good support for a refinement of the framework comprising 14 domains of theoretical
constructs (average silhouette value 0.29): ‘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’, ‘Emotions’, and ‘Behavioural
Regulation’.
Conclusions: The refined Theoretical Domains Framework has a strengthened empirical base and provides a
method for theoretically assessing implementation problems, as well as professional and other health-related
behaviours as a basis for intervention development.
Keywords: Theoretical domains framework, Behaviour, Change, Implementation, Validation, Theory

Background motivation, the ability to keep up with current

Behaviour change is key to improving healthcare and changes, clarity of roles and practice, and the culture
health outcomes. Behaviours may be those of healthcare of specific healthcare practices [3,4].
workers, such as implementation of evidence-based prac- Improving implementation of evidence-based practice
tice, of patients, such as medication adherence, or of the by healthcare workers depends on changing multiple
general population, such as smoking cessation and in- behaviours of multiple types of people (e.g., health pro-
creasing physical activity. Despite high-level recommen- fessionals, managers, administrators) [5]. Changing be-
dations to improve implementation of evidence-based haviour is not easy, but is more effective if interventions
practice [1,2] and a rapidly developing field of implemen- are based on evidence-based principles of behaviour
tation science, implementation remains variable, with nu- change [6]. These principles form part of many theories
merous organisational and individual factors influencing of behaviour change, but are seldom drawn on in
healthcare workers’ behaviour. These factors include the designing and evaluating implementation interventions.
availability of evidence, its relevance to practice, the There is some evidence that behaviour change interven-
dissemination of evidence and guidelines, individual tions informed by theory are more effective than those
that are not [7,8]. Designing interventions on the basis
* Correspondence: [email protected] of practitioner or researcher intuition rather than theory
3
Research Department of Clinical, Centre for Outcomes Research and
Effectiveness (CORE), Education and Health Psychology, University College
precludes the possibility of understanding the behaviour
London, 1-19 Torrington Place, London, WC1E 7HB, UK change processes that underlie effective interventions
Full list of author information is available at the end of the article

© 2012 Cane 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.
Cane et al. Implementation Science 2012, 7:37 Page 2 of 17
http://www.implementationscience.com/content/7/1/37

and of applying this knowledge to inform the design of based guidelines for acute low back pain [14,15] and de-
future interventions. This is also the case where theory is velop theory-informed behaviour change interventions
cited but poorly applied to intervention development [9]. [16]. In the UK, examples include studies of the barriers
In a review of 235 guideline development and imple- and levers related to hand hygiene [17]; the assessment
mentation studies, only 22.5% were judged to have of theoretical domains relevant to blood transfusion
used theories of behaviour change, and 16.6% of stud- practice across different contexts including neonatal and
ies using a single theory [10]. A further 4.3% used only adult intensive care units [18,19]; and identifying difficul-
selected constructs from theories; across the majority ties in implementing guidelines relating to schizophrenia
of studies there was no clear rationale for theory use. [20]. In Denmark, it has been used to understand behav-
While use of a single theory may be appropriate and iour in the implementation of tobacco use prevention
lends itself to theory testing, in many cases the selec- and counselling guidelines amongst dental providers
tion has not been justified and the theory is not tested [21]. Most of this research has used interviews and focus
[9]. If theory selection is not informed by a compre- groups that are resource intensive; a questionnaire meas-
hensive theoretical assessment of the implementation ure is currently being developed by the authors. This will
or other behavioural problem, there is a risk of miss- facilitate research investigating prediction of implemen-
ing relevant theoretical constructs or including irrele- tation and other types of behaviour change.
vant ones. A second problem in applying theory to This article is one in a series of articles documenting
intervention design stems from basing interventions on the development and use of the TDF to advance the
several theories with overlapping theoretical constructs science of implementation research. To inform future
[11,12]. This makes it difficult to identify the specific use of the TDF, we conducted the current study to pro-
processes underlying successful behaviour change. vide a more thorough test of the validity of the frame-
To overcome such problems, an integrative frame- work than was carried out in the original research. The
work of theories of behaviour change was developed overall objective of the study was to examine the
by 18 psychological theorists in collaboration with 16 content validity of the TDF. Specifically, we wanted to
health service researchers and 30 health psychologists confirm the optimal domain structure (number of
[13]. The aim of the Theoretical Domains Framework domains), domain content (component constructs in
(TDF) was to simplify and integrate a plethora of be- each domain), and domain labels (most appropriate
haviour change theories and make theory more access- names that best reflected the content of the validated
ible to, and usable by, other disciplines. The group domain structure). Card sorting methodology was used
identified 33 theories and 128 key theoretical con- to conduct the validation of the TDF in this study. By
structs related to behaviour change and synthesised building on the validation process undertaken by Michie
them into a single framework to assess implementation et al. [13] the present study aimed to improve the em-
and other behavioural problems and inform interven- pirical basis of this framework.
tion design. They used a six stage consensus approach:
identifying theories and theoretical constructs relevant to Method
behaviour change, where a theoretical construct was Design
defined as ‘a concept specially devised to be part of a the- The study used a cross-sectional design.
ory’ [13]; simplifying these resulting constructs into over-
arching theoretical domains, where a theoretical domain Participants
was defined as ‘a group of related theoretical con- Eligible participants possessed a good understanding of
structs’ [13]; evaluating the importance of the theoret- behaviour change theory and were unaware of the original
ical domains; conducting an interdisciplinary evaluation framework reported in Michie et al. [13]. Potentially eli-
and synthesis of the domains and constructs; validating gible participants were identified by systematically search-
the domain list; and piloting interview questions rele- ing five online electronic journal databases (Web of
vant to the constructs and domains. This resulted in Science, PsychInfo, CINAHL Plus, Ingenta Connect, JStor)
12 theoretical domains and exemplar questions for using terms ‘behaviour change’ AND ‘theory’ from 1990 to
each to use in interviews or focus groups to provide a 2011, by sending email invitations through membership
comprehensive theoretical assessment of implementa- mailing lists for the European Health Psychology Society,
tion problems. the American Psychological Association Division of
This framework has been used by research teams Health Psychology, the USA’s National Institute of Health’s
across several healthcare systems to explain implementa- Behaviour Change Consortium, the Midlands Health
tion problems and inform implementation interventions. Psychology Network in the UK, and by searching through
For example, in Australia it has been used to identify the delegate lists from the 2008 to 2010 annual conferences of
barriers and enablers to the implementation of evidence- the UK Society for Behavioural Medicine and British
Cane et al. Implementation Science 2012, 7:37 Page 3 of 17
http://www.implementationscience.com/content/7/1/37

Psychological Society’s Division of Health Psychology. Evaluating the framework

The contact details of all individuals identified as To evaluate the original framework, a three step method
authors on papers identified through the electronic was used:
database searches were located via publically available
sources (e.g., searches of university and other organisa- Step one: Identify the optimal number of domains by
tion websites). sort task methods.
Of 101 individuals who asked for full information about Step two: Establish domain content by identifying the
the study, 61 expressed an interest in taking part and were most suitable construct allocation to each of the domains.
sent links to one of the online tasks; 37 of these (61%) Step three: Finalise domain labels by identifying the most
completed their assigned task. The majority were from the appropriate labels for new domains (labels for domains
UK (16), with the remaining participants being from the that replicated the original ones were retained).
Netherlands (8), USA (2), Ireland (2), Australia (2), Italy
(2), Portugal (1), South Africa (1), Greece (1), Germany Sort task methodology
(1), and Switzerland (1). The 27 women and 10 men had a Two types of sort task were used: an open sort task and
mean age of 36.54 years (range 22 to 62). a closed sort task (see Figure 1). In the open sort task,
The sample size for the tasks was based on estimates participants were asked to sort constructs into groups of
of between six and 36 participants shown as sufficient their choice and label these groups according to their
for sort and cluster analysis tasks [22-28]. For content- content. The optimal grouping of constructs into
validation tasks, such as those proposed in the closed domains was identified using Fuzzy Cluster Analysis [34],
sort task, two to 24 participants have been shown to be whereby sorting patterns across individual participants
sufficient [29-32], with more than five participants redu- could be aggregated into clusters. This cluster technique
cing the influence of rater outliers [33]. has the benefit over the more commonly used k-means

Figure 1 Steps taken to validate the Theoretical Domains Framework.

Cane et al. Implementation Science 2012, 7:37 Page 4 of 17
http://www.implementationscience.com/content/7/1/37

and k-medoid cluster analysis, and other grouping meth- be given. For the closed sort task, 12 labeled boxes were
ods, in that it allocates a membership value (in the form displayed, each described by a single domain label from
of a probability value) for each possible construct-cluster the original framework. In both tasks, individual con-
pairing rather than simply assigning a construct to a single structs could be assigned to multiple boxes and for every
cluster, thereby the membership of items to more than allocation a confidence rating was requested using a
one group could be assessed. The results obtained from drop-down menu (from 1 – ‘not at all confident’ to 10 –
the open sort task and Fuzzy Cluster Analysis were used ‘extremely confident’). Constructs were presented in ran-
to identify the optimal domain structure (step one), the dom order that was determined by the online program
content of new domains (step two), and the most appro- for each participant. Definitions for each construct (open
priate domain labels, based on the group names given and closed sort tasks) and domain (for closed sort task
by participants (step three). In the closed sort task, par- only) were available when the participant hovered over
ticipants were asked to sort constructs into the domains the word with their mouse. Participants were asked,
defined in the original framework and rate their confi- through open-ended questions, to record the length of
dence in their allocation of each construct to a domain. time they had been involved in using behaviour change
The extent to which participants believed each con- theories, the context in which they used them (e.g., teach-
struct belonged to the original 12 domains was assessed ing, research, etc.) and their expertise in behaviour
by Discriminant Content Validation (DCV) methods. change theory and in using behaviour change interven-
DCV methods are able to examine the confidence of tions (1 – ‘A great deal’, 2 – ‘quite a lot’, 3 – ’some’, 4 – ‘a
relationship between a single item and a particular little’, 5 – ’none’).
domain [35]. The results from the closed sort task were
used to identify any domains containing constructs with Procedure
high confidence ratings and good agreement between Invitations were emailed to potentially eligible partici-
participants (step one), and the constructs allocated to pants giving a brief overview of the study and inquiring
these domains (step two). Both types of sort task as to their expertise. If they considered themselves to
informed step one because it was considered important have expertise in behaviour change theory and reported
to include domains that developed naturally from the not knowing about the original framework, they were
construct groupings (as informed by the open sort task), invited to participate and emailed the relevant web link
and to include domains to which there was good agree- to the task they were allocated to. Eligible participants
ment across participants in the confidence of construct were alternately allocated to an open or closed sort task
allocation to these domains (as informed by the closed based on the order in which they contacted the
sort task). To achieve this, the open sort task results researchers. To avoid contamination of results across
were used to identify the domains based on the clusters tasks, each participant was allocated to, and completed,
formed in the open sort task; the closed sort results only the closed sort task or the open sort task. For both
were then used to identify any additional domains for tasks, an information screen gave a brief background to
which there was good agreement and confidence in the study and asked for consent to take part. Participants
assigning constructs to these domains. were given detailed instructions on how to complete
their task (see Additional file 2) before completing the
Materials sort task they were assigned to. There was no time limit.
There were 112 unique constructs (see Additional file 1), In both tasks, participants were asked to familiarise
after 12 duplicates from the original framework were themselves with the construct definitions and, in the
removed (participants had the opportunity to sort each closed sort task only, the domain definitions. In the open
construct to multiple domains). Definitions for the sort task, participants were asked to sort the constructs
domains and constructs were selected or constructed into groups based on their semantic similarity using as
from dictionaries, (e.g., American Psychological Associa- many groups as they wanted to (up to 24) and were
tion Dictionary of Psychology [36]), and internet sources asked to provide a label for each group created. Partici-
(e.g., www.oed.com). Each definition was evaluated by pants could also provide a description for each group if
the authors of the original framework and definitions they felt it was necessary. In the closed sort task, partici-
were agreed by consensus. The sort tasks were delivered pants were asked to assign each construct to one or
via an online computer program with constructs dis- more of the 12 labelled domain boxes that they thought
played at the top of the computer screen. For the open were most appropriate. Across both tasks, participants
sort task, 24 unlabelled boxes were displayed below the were asked to give confidence ratings for each assign-
construct item window into which the participants could ment; if an item was not allocated to a domain it auto-
sort the constructs. Above each box a space was given so matically received a confidence rating of 0. For both
that labels and descriptions for each group created could tasks, participants were made aware that they could
Cane et al. Implementation Science 2012, 7:37 Page 5 of 17
http://www.implementationscience.com/content/7/1/37

allocate each construct to multiple groups. After assign- k-means and k-medoid cluster methods). The fit of con-
ing all constructs, participants were asked to review their structs within the clusters was calculated by silhouette
construct allocations and to change any allocations if values (s(i)) [42]. Silhouette values are calculated for
they wished to. On completion, participants were given each construct and range from +1, indicating strong as-
further information about the project. sociation with a cluster and distance from neighboring
clusters, through 0, indicating no distinct association
Data analysis with clusters, to −1, indicating that a construct is prob-
Data were collected using MySQL databases. For the ably assigned to the wrong cluster and should be consid-
open sort task, data were the construct-group alloca- ered as belonging to the neighbouring cluster [42]. The
tions, confidence ratings, and group labels allocated by average silhouette values (ave s(i)) across construct items
the participant. For the closed sort task data were the within a cluster indicates how well a cluster is defined,
construct-group allocations and confidence ratings. and the overall average of silhouette values across clus-
ters can be used to compare cluster solutions of different
Open sort sizes.
To examine the optimal clustering of constructs (step The optimal outcome of the cluster analysis is to
one: identify domains), the open sort data were first achieve the highest average silhouette value with the
organised into a dissimilarity matrix for each participant. fewest clusters. It has been argued that average cluster
Construct pairs, consisting of all possible construct-by- silhouette values greater than 0.70 indicate a strong
construct combinations, were assigned 0 if they were structure, whilst average silhouette values below 0.50 in-
placed in the same group and 1 if they were placed in a dicate weak structures and silhouette values <0.25 indi-
different group. Agreement across these individual cate that there is little evidence for any reliable structure
matrices was assessed using Mantel Correlations and [34]. Informed by these cutoff values, we considered that
Kendall's Coefficient of Concordance, W [37] using a construct with a silhouette value <0.25 in relation to a
CADM.global and CADM.post from the ‘ape’ package cluster did not belong to that cluster.
[38] in the R statistics program [39]. Mantel Correlations In addition to identifying the optimal domain struc-
determine the extent to which an individual participants’ ture, the open sort results were used to identify the ex-
matrix correlates with other participants’ matrices and tent to which the clusters replicated the construct
were used to identify any potential outlying sort patterns allocation in the original framework when domain labels
that should be excluded from subsequent analysis. An were not provided (step two: establish domain content).
individual’s matrix is considered to be an outlier when it Congruence was quantified as the percentage of con-
negatively correlates with the other participants’ matri- structs from the original framework domain remaining
ces [40]. Kendall’s Coefficient of Concordance provides in a cluster solution (e.g., if domain m contained con-
an indication of the overall concordance across all parti- structs x, y, and z and the cluster contained only x and
cipants’ sort patterns, Kendall’s W ranges from 1 to 0 z, then congruence was 67%). If the structure of the
[37], where 1 equals complete agreement in sorting pat- domains identified in the Fuzzy Cluster Analysis was
terns and 0 equals no agreement across sorting patterns. considerably different from that of the original frame-
To identify the clusters formed through these sorting work, confidence ratings would be used for secondary
patterns, means were calculated for each construct analysis to infer construct allocation to the new domains
pairing across individual matrices to form a single, formed.
aggregated dissimilarity matrix. Fuzzy Cluster Analysis The group labels given by participants in the open sort
of this matrix, using the FANNY algorithm [34,41] in the task were organised according to their similarity and the
R statistics program, led to a membership value assigned frequency that they occurred across participants noted.
to each construct-cluster pairing. These membership Those labels that occurred frequently and were related
values, converted into percentages, serve as an indica- to the content of the newly-formed domains were used
tion of the extent to which a construct belongs to a to inform newly-formed domain labels (step three: final-
particular cluster. Values near 100% indicate a high ise domain labels).
probability of association with a cluster and values near
0% indicate a low probability of association. Using these Closed sort
values, construct membership to multiple domains can To identify pre-existing domains that might also be con-
be assessed (e.g., construct x might have 53% member- sidered for inclusion in the framework (step one: identify
ship to cluster y and 47% membership to cluster z). domains), the strength and agreement of construct allo-
Constructs were then allocated to the cluster with cations to pre-existing domains from the closed sort task
which it has the highest membership value (known as a were examined. Confidence ratings for each construct x
‘hard’ cluster solution and comparable to outputs of the domain pairing, excluding those that had no confidence
Cane et al. Implementation Science 2012, 7:37 Page 6 of 17
http://www.implementationscience.com/content/7/1/37

ratings, were applied to a table. To examine the agree- behaviour change theory for a mean of 9.74 (SD = 9.14)
ment of these construct x domain ratings and construct years and rated both their expertise in behaviour change
assignment across participants, two-way intraclass cor- theory and in delivering behaviour change interventions
relation coefficient (ICC) measures of consistency were as 1.97 (SD = 0.64) and 2.46 (SD = 0.90), respectively, as
used within each domain [43]. In line with previous re- measured on five-point scales (lower score indicates
search we classified ICC values <0.21 as indicating poor more expertise).
agreement, values between 0.21 to 0.40 as fair agree-
ment, values between 0.41 to 0.60 as moderate agree- Sample size suitability and open sort pattern
ment, and values of ≥0.61 as good to excellent concordance
agreement [44]. ICC values were used as an indication of Post-hoc power analysis for the closed sort task revealed
the agreement in assignments and ratings across partici- that there was sufficient power (82%) with the final sam-
pants, but were not used to influence the final domain ple size of 18 to detect a mean rating of 1.53 (SD = 2.42,
content. d = 0.63) as significant within a one-tailed one-sample
To identify the strength of construct assignment to t-test with α = 0.05. The mean rating used in the power
particular domains, DCV methods were used with one- analysis was based on the mean of confidence ratings
sample t-tests on the participants’ confidence ratings across all variables included in the closed sort analyses.
against the value zero. A construct was considered as For the open sort task, Mantel Correlation analysis
belonging to a domain if its mean confidence rating indicated that all participants’ matrices were positively
across participants was significantly greater than zero correlated, with aggregated Mantel correlation values
(p < 0.05) following the adoption of Hochberg’s correc- for each participant ranging from 0.14 to 0.25 (see
tion [45] (see [29,35] for similar methods). Hochberg’s Additional file 3). Therefore none of the participants’
correction was used to control for the family-wise sort patterns were considered as outliers, and matrices
error rate given the number of tests used. Whilst this from all 19 participants were included in the final ana-
approach may not be considered a conventional use of lysis. The overall concordance of sorting patterns was
one-sample t-tests, it provides a suitable criterion for W = 0.22, p = 0.01, reflecting the unconstrained nature
inclusion and exclusion of constructs to a particular of this task and its high number of variables.
domain over and above the use of a subjective cut-off
value. To ensure that domains with highly-rated, rele- Step one: identify domains
vant constructs assigned to them were considered for In the open sort task, participants created on average
inclusion in the framework, domains containing two or 13.59 (SD = 3.61) groups. To identify the optimal fit for
more constructs with ratings significantly greater than the cluster patterns based on the groups created by the
zero were considered. These constructs were also used participants, silhouette values for solutions of minimum
to inform construct allocation to pre-existing domains two and maximum 18 clusters were examined. Analysis
(step two: establish domain content). The allocation of revealed the 13-cluster solution to be the most appro-
constructs to domains in the closed sort task was priate fit, achieving the highest overall average silhou-
compared with construct allocation in the original ette value of 0.29 (Figure 2 shows the relative overall
framework to identify the extent of congruence be- silhouette values plotted for each cluster solution). The
tween assigned constructs when domain labels were construct allocation within the ‘hard’ version of the
available. Here congruence was quantified as the per- 13-cluster solution, whereby each construct is allocated
centage of constructs from the original framework do- to only one domain, is presented in Table 1 next to
main that were also in that domain within this study. the domains they most closely represent (see ‘Open
sort task construct clusters’ and see Additional file 4
Ethical approval for related silhouette values).
The study was approved by University College London’s Within the 13 cluster solution, three of the original
Psychology Department Ethics Committee [STF/2007/ domains, ‘Beliefs about Capabilities’, ‘Beliefs about Conse-
003], and each participant gave full informed consent quences’, and ‘Motivation and Goals’, formed two clusters
prior to participating. each. Four of the 13 clusters showed low average silhou-
ette values (<0.25), one of the clusters arising from the
Results ‘Motivation and Goals’ domain, also ‘Memory, Attention,
Eighteen participants completed the closed sort task and and Decision Processes’, ‘Environmental Context and
19 the open sort task. All participants indicated that they Resources’, and ‘Behavioural Regulation’. This was due to
had experience of behaviour change theory through ei- the inclusion of a number of constructs that had low (or
ther research, clinical practice, or teaching (or a combin- negative) silhouette values, indicating that these con-
ation of these). Participants reported working with structs were not closely grouped with the other
Cane et al. Implementation Science 2012, 7:37 Page 7 of 17
http://www.implementationscience.com/content/7/1/37

0.25
Mean silhouette width

0.20
0.15

5 10 15
Number of clusters

Figure 2 Comparison of fit across 2–18 cluster solutions.

constructs within these clusters. To examine the impact ‘Procedural Knowledge’ attained a silhouette value equal
of these low value constructs, they were removed and the to or greater than 0.25 (all other knowledge-related con-
average silhouette values of the clusters were recalcu- structs <0.23). In contrast, within the closed sort task
lated. After removal, 10 clusters had average silhouette the confidence ratings of three knowledge-related con-
values greater than 0.25 (see Additional file 4, column 7) structs, ‘Knowledge’, ‘Knowledge of Task Environment’
with the average silhouette value across these 10 domains and ‘Procedural Knowledge’ indicated that knowledge
equal to 0.47 and the concordance across sorting pat- might form a separate domain if the label ‘Knowledge’
terns increasing to W = 0.34 (p = 0.01). Three clusters was available (confidence ratings >6.32 across these
remained with silhouettes below 0.25, ‘Environmental three constructs). Therefore, it was considered that the
Context and Resources’, ‘Memory, Attention, and Deci- ‘Knowledge’ should be included when it was thought to
sion Processes’, and ‘Behavioural Regulation’. Whilst be important in the specific context.
these clusters showed relatively weak cluster formations Based on the results across both tasks, 14 domains
in the open sort, the confidence ratings in the closed sort were specified through this first step. Eight domains were
indicated that when the domain labels were apparent the similar to the original framework domains: ‘Knowledge’,
confidence ratings of allocated constructs were sufficient ‘Skills’, ‘Social/Professional Role and Identity’, ‘Memory,
to form domains. Therefore, these three domains were Attention and Decision Processes’, ‘Environmental Con-
considered important to retain in the framework. Also, text and Resources’, ‘Social Influences’, ‘Emotion’, and ‘Be-
there was no cluster indicative of the domain of ‘Know- havioural Regulation’. The domains ‘Beliefs about
ledge’ in the 13 cluster solution, with all constructs from Capabilities’, ‘Beliefs about Consequences’, and ‘Motiv-
the original ‘Knowledge’ domain allocated to alternative ation and Goals’ were retained but were divided into six
clusters; the constructs ‘Knowledge’ and ‘Knowledge of new clusters. The domain of ‘Nature of the Behaviours’
task environment’ were allocated to the ‘Environmental was removed because it was not represented in the open
Context and Resources’ cluster, ‘Mindsets’ was allocated sort by any single cluster solution and only had one con-
to one of the clusters arising from the ‘Beliefs about struct assigned to it in the closed sort task.
Capabilities’ domain, ‘Schemas’ was allocated to the
‘Memory, Attention and Decision Processes’ cluster and Step two: establish domain content
‘Procedural knowledge’ was allocated to the ‘Skills’ clus- The mean confidence ratings and ICCs for the construct
ter. However, within these construct reassignments only allocation to domains given in the closed sort task are
Cane et al. Implementation Science 2012, 7:37 Page 8 of 17
http://www.implementationscience.com/content/7/1/37

Table 1 Comparison of the refined framework, closed sort task, and open sort task groupings
Refined framework domain Closed Sort Task construct groups Open Sort Task construct clusters
name and constructs (* = new domain) (constructs achieving p < .05a; in (constructs in order of s(i) values
order of confidence rating high – low) decreasing; italics = constructs with
silhouette value < .25)
1. Knowledge
Knowledge (including knowledge Knowledge (including knowledge of -No cluster representing Knowledge-
of condition /scientific rationale) condition /scientific rationale)
Procedural knowledge Procedural knowledge
Knowledge of task environment Knowledge of task environment
2. Skills
Skills Skills Competence
Skills development Skills development Skills
Competence Competence Skill assessment
Ability Ability Ability
Interpersonal skills Interpersonal skills Interpersonal skills
Practice Practice Skills development
Skill assessment Skill assessment Procedural knowledge
3. Social/ Professional Role and
Identity
Professional identity Professional identity Organisational development
Professional role Professional role Organisational culture / climate
Social identity Social identity Management commitment
Identity Identity Professional role
Professional boundaries Professional boundaries Crew resource management
Professional confidence Professional confidence Leadership
Group identity Leadership Change management
Leadership Group identity Professional boundaries
Organisational commitment Organisational commitment Organisational commitment
Supervision
Professional identity
Project management
Champions / To champion
Team working
Power
Hierarchy
4. Beliefs about Capabilities
Self-confidence Self-confidence Self-efficacy
Perceived competence Perceived competence Perceived competence
Self-efficacy Self-efficacy Self-confidence
Perceived behavioural control Perceived behavioural control Perceived behavioural control
Beliefs Self-esteem Professional confidence
Self-esteem Beliefs Self-esteem
Empowerment Empowerment
Professional confidence Professional confidence
5. Optimism*
Optimism Optimism
Pessimism Pessimism
Cane et al. Implementation Science 2012, 7:37 Page 9 of 17
http://www.implementationscience.com/content/7/1/37

Table 1 Comparison of the refined framework, closed sort task, and open sort task groupings (Continued)
Unrealistic optimism Unrealistic optimism
Identity Identity
Mindsets
6. Beliefs about Consequences
Outcome expectancies Outcome expectancies Beliefs
b b
Chars. of outcome expectancies Chars. of outcome expectancies Attitudes
Beliefs Beliefs Outcome expectancies
b
Anticipated regret Anticipated regret Chars. of outcome expectancies
Consequents Consequents Illness representations
7. Reinforcement *
Rewards (proximal/distal, valued/not valued, Rewards (proximal/distal, valued/not
probable/improbable) valued, probable/improbable)
Incentives Incentives
Punishment Punishment
Consequents Sanctions
Reinforcement Contingencies
Contingencies Reinforcement
Sanctions Consequents
8. Intentions*
Stability of intentions Goals (autonomous, controlled) Stability of intentions
Stages of change model Intrinsic motivation Stages of change model
Trans. model/stages of change b Goal target /setting Trans. model/stages of change b
Distal and proximal goals Certainty of intentions
Goal priority Intention
Intention Commitment
Stability of intentions Intrinsic motivation
Certainty of intentions Mods. of the intention-behaviour gap b
9. Goals*
Goals (distal / proximal) Goal target/ setting
Goal priority Goals (distal / proximal)
Goal / target setting Goal priority
Goals (autonomous / controlled) Goals (autonomous / controlled)
Action planning Action planning
Implementation intention Implementation intention
Representation of tasks
10. Memory, Attention and Decision Processes
Memory Memory Memory
Attention Attention Attention control
Attention control Attention control Attention
Decision making Decision making Decision making
Cognitive overload / tiredness Cognitive overload / tiredness Appraisal
Schemas
Cognitive overload / tiredness
11. Environmental Context and Resources
Environmental stressors Environmental stressors Conflict-comp. demands, conf. roles b
Resources / material resources Resources / material resources Barriers and facilitators
Cane et al. Implementation Science 2012, 7:37 Page 10 of 17
http://www.implementationscience.com/content/7/1/37

Table 1 Comparison of the refined framework, closed sort task, and open sort task groupings (Continued)
Barriers and facilitators Barriers and facilitators Environmental stressors
Organisational culture /climate Organisational culture climate Knowledge of task environment
Person x environment interaction Person x environment interaction Person x environment interaction
Salient events / critical incidents Salient events / critical incidents Control of behaviour, material and
social environment
Knowledge
Empowerment
Negotiation
Anticipated regret
Threat
Past behaviour
12. Social Influences
Social pressure Social pressure Group norms
Social norms Social norms Group conformity
Group conformity Group conformity Group identity
Social comparisons Social comparisons Social pressure
Group norms Group norms Social norms
Social support Social support Social support
Intergroup conflict Intergroup conflict Alienation
Power Power Social comparisons
Group identity Group identity Intergroup conflict
Alienation Alienation Social identity
Modelling Modelling
13. Emotion
Anxiety Anxiety Anxiety
Fear Fear Depression
Affect Affect Positive / negative affect
Stress Stress Stress
Depression Depression Fear
Positive / negative affect Positive / negative affect Affect
Burn-out Burn-out Burn-out
14. Behavioural Regulation
Self-monitoring Self monitoring Learning
Breaking habit Breaking habit Review
Action planning Action planning Breaking habit
Direct experience
Self-monitoring
Evaluation
a b
Key: = after applying Hochberg’s correction for multiple comparisons within each domain, - Chars. of outcome expect. = Characteristics of outcome
expectancies; Conflict-comp. demands, conf. roles = Conflict - competing demands, conflicting roles; Mods. of the intention-behaviour gap = Moderators of the
intention-behaviour gap; Trans. model/stages of change = Transtheoretical model and stages of change.

shown in Additional file 4. In the closed sort task, the constructs listed in the domains of the original frame-
content of domains for ‘Emotion’, ‘Skills’, ‘Motivation and work (>69%) and fair ICCs (0.31 to 0.40). The domains
Goals’, ‘Social/Professional Role and Identity’, ‘Beliefs of ‘Knowledge’, ‘Environmental Context and Resources’
about Capabilities’, and ‘Memory, Attention and Decision and ‘Social Influences’ showed lower congruence with
Processes’ all showed good congruence with the the constructs listed in the original domains (27% to
Cane et al. Implementation Science 2012, 7:37 Page 11 of 17
http://www.implementationscience.com/content/7/1/37

50%) and fair ICCs (0.26 to 0.39). The domains of ‘Be- Regulation), ‘Beliefs’ (Beliefs about Consequences and
havioural Regulation’, ‘Nature of the Behaviours’, and Beliefs about Capabilities), ‘Consequents’ (Beliefs about
‘Beliefs about Consequences’ showed both low congru- Consequences and Reinforcement), ‘Group identity’ (So-
ence between the original constructs and those assigned cial/Professional Role and Identity and Social Influences),
to these domains (<27%) and low ICCs (0.07 to 0.25). ‘Identity’ (Social/Professional Role and Identity and Opti-
This was due in part to the low number of constructs mism), and ‘Professional confidence’ (Social/Professional
assigned to these domains. ‘Behavioural Regulation’ only Role and Identity and Beliefs about Capabilities).
had two constructs out of the original ten (Self-monitor-
ing, and Action planning) that were rated as belonging
Step three: finalise domain labels
to the domain. ‘Nature of the Behaviours’ only had one
Fifteen of the 19 open sort participants provided labels
construct (Routine/automatic/habit) included from the
for the groups they created. The majority of labels were
original six constructs. ‘Beliefs about Consequences’ only
similar to those in the original framework: (number of
had five of the constructs included from the original
participants giving that label shown in parenthesis):
framework. Only constructs that achieved significance in
Knowledge (4), Skills (5), Intentions (7), Goals (6), Emo-
the closed sort after Hochberg correction were allocated
tion (9), Cognitive-related (8), Beliefs (5), Beliefs about
to these pre-existing domains from the original
Capabilities (7), Outcomes (6), Environment-related (6),
framework.
Organisational (7), Models / Theories (8), Learning /
For the newly formed clusters arising from ‘Motivation
Reinforcement (7), Self-Regulation (3), Consequences
and Goals’ (two clusters), ‘Beliefs about Capabilities’ (one
(3), Social / Group (14), and Planning (2). Examples of
cluster), and ‘Beliefs about Consequences’ (one cluster),
other labels that could not be categorized (i.e., labels
construct allocation was informed by the constructs
given by only one participant) included ‘Techniques’, ‘Bar-
assigned to these clusters in the open sort task that
riers’, ‘Awareness’, ‘Reviewing’, and ‘Persistence’. Given the
achieved individual construct silhouette values greater
similarity between the labels provided in the open sort
than 0.25.
task and the labels used in the original framework, those
To identify if any constructs should be considered for
domains that were retained with only minor modification
multiple allocation to domains the membership values
were allocated their respective label used in the original
from the Fuzzy Cluster Analysis were examined. This
framework. The labels for the newly developed domains
revealed that the majority of constructs (74/112) were
were based on the frequency of labels and the domain
strongly associated with only one cluster (i.e., showed
content: these were Intentions, Goals, Reinforcement,
membership values over 80% to one specific cluster, see
and Optimism. The domain label of ‘Emotion’ was plura-
Additional file 5). A further 32 constructs showed mod-
lised to ‘Emotions’ to bring in line with the other domain
erately high associations with one cluster (memberships
labels and to ensure that it clearly represented the range
values between 28% to 79%), with the remaining propor-
of emotions that were included as component constructs.
tion of memberships for these constructs spread over
Therefore, the final labels chosen to represent the 14
other clusters. Only eight constructs had the greatest
domains were: ‘Knowledge’, ‘Skills’, ‘Social/Professional
proportion of their membership values split across at
Role and Identity’, ‘Beliefs about Capabilities’, ‘Optimism’,
least two clusters, indicating possible multiple domain
‘Beliefs about Consequences’, ‘Reinforcement’, ‘Intentions’,
memberships, these were ‘Knowledge’, ‘Coping strategies’,
‘Goals’, ‘Memory, Attention and Decision Processes’, ‘En-
‘Empowerment’, ‘Anticipated regret’, ‘Negotiation’, ‘Mod-
vironmental Context and Resources’, ‘Social Influences’,
erators of the intention-behaviour gap’, ‘Routine/auto-
‘Emotions’, and ‘Behavioural Regulation’.
matic/habit’, and ‘Past behaviour’. However, none of the
multiple memberships indicated in the open sort results
were replicated in the closed sort task where three dif- The refined framework
ferent constructs, ‘Professional confidence’, ‘Beliefs’, and The refined framework contains 14 domains and 84
‘Group identity’, were allocated to multiple domains. component constructs (the number of component con-
Given lack of agreement across the two tasks, only the structs in each domain is defined in brackets): ‘Know-
multiple allocations shown in the closed sort task or mul- ledge’ (3), ‘Skills’ (7), ‘Social/Professional Role and
tiple allocations that occurred through the construct se- Identity’ (9), ‘Beliefs about Capabilities’ (8), ‘Optimism’
lection process (i.e., using the closed sort for predefined (4), ‘Beliefs about Consequences’ (5), ‘Reinforcement’ (7),
domains and using the open sort for new domains) were ‘Intentions’ (3), ‘Goals’ (6), ‘Memory, Attention and Deci-
used in the final framework. Using this approach, six sion Processes’ (5), ‘Environmental Context and
constructs were allocated to more than one domain: (the Resources’ (6), ‘Social Influences’ (11), ‘Emotions’ (7), and
domains that constructs are allocated to are shown in ‘Behavioural Regulation’ (3). The full version of the new
parenthesis) ‘Action planning’ (Goals and Behavioural framework is shown in Table 2.
Cane et al. Implementation Science 2012, 7:37 Page 12 of 17
http://www.implementationscience.com/content/7/1/37

Discussion behaviour. The separation of ‘Beliefs about Conse-

This validation study, using open and closed sort tasks, quences’ into two domains, one retaining the original
has shown good support for the basic structure of the name and one termed ‘Reinforcement’, made psycho-
TDF and led to refinements producing 14 domains: logical sense. The former refers to beliefs whereas the
‘Knowledge’, ‘Skills’, ‘Social/Professional Role and Identity’, latter refers to constructs of associative learning. There
‘Beliefs about Capabilities’, ‘Optimism’, ‘Beliefs about Con- was also a separation within the ‘Beliefs about Capabil-
sequences’, ‘Reinforcement’, ‘Intentions’, ‘Goals’, ‘Memory, ities’ domain with a separate ‘Optimism’ domain being
Attention and Decision Processes’, ‘Environmental Con- formed. This separation makes psychological sense in
text and Resources’, ‘Social Influences’, ‘Emotions’, and that the constructs in the optimism cluster concern gen-
‘Behavioural Regulation’. There are three key advantages eral disposition rather than specific capabilities required
of this framework. First, there is comprehensive coverage to achieve an outcome. The domain ‘Behavioural Regula-
of possible influences on behavior. Second, there is clar- tion’ is clearer in the refined framework where it refers
ity about each kind of influence, as a result of each do- to self-regulatory processes rather than including a mix-
main being specified by component constructs. Third, ture of self-regulation and goal-related constructs, as
the framework makes links between theories of behav- was the case in the original TDF.
iour change and techniques of behaviour change to ad- Second, the ‘Nature of the Behaviours’ domain was
dress implementation problems. The framework can be dropped in the new framework, because its original
applied by gathering either qualitative data (interviews component constructs were not assigned to the domain
or focus groups) or quantitative data (e.g., by question- in the closed sort, and there was no cluster representing
naires). The findings have strengthened the evidence for the ‘Nature of the Behaviours’ in the open sort. This
the structure and content of the domains, increasing strengthens the coherence of the new TDF because the
confidence in the usefulness of the TDF as an approach domain did not sit easily in the original TDF. It was
to assessing implementation and other behaviour pro- defined as the ‘Essential characteristics of the behaviour’,
blems, and laying the foundation for theoretically had constructs relating to habit and experiences/past
informed interventions. behaviours, and constituted an outcome, or dependent
To the authors’ knowledge, Fuzzy Cluster Analysis and variable, rather than an independent variable. Whilst
Discriminant Content Validity have not been used in understanding the nature of behaviours is absolutely key
combination to determine the validity of a framework to analyzing implementation and other behavioural pro-
structure. By combining these methods, we have investi- blems, analysing the nature of behaviour is a different
gated the validity of the original framework both when task than analysing influences on behaviour. A comple-
the original domain labels were, and were not, pre- mentary theoretical approach to analyzing behaviour as
sented. The results from both the open and closed sort a basis for intervention design has been recently devel-
tasks generally replicated the original framework, which oped, as part of the ‘Behaviour Change Wheel’ [46]. Pre-
adds confidence to the validity of the framework’s vious studies that have adopted the TDF framework
structure. have seldom used the ‘Nature of the Behaviours’ domain
The study findings pointed to some changes in the [17]. Furthermore, where the domain has been used, in
framework, which had good face validity. First, there relation to changing transfusion practice, it was noted
was a separation and clarification of a number of exist- that when participants were asked questions relating to
ing domains. The separation of ‘Motivation and Goals’ the ‘Nature of the Behaviours’ domain they often
into two domains of ‘Intentions’ and ‘Goals’ was indi- repeated answers that were previously given in response
cated by both the closed and open sort task results and to questions relating to the ‘Behavioural Regulation’ do-
was particularly apparent in the labels provided by the main [19], therefore making responses in respect to
participants, with labels relating to ‘intentions’ and ‘goals’ ‘Nature of the Behaviours’ redundant. This along with
amongst the most frequently used. The APA dictionary empirical evidence shown in the present study shows a
of psychology defines a goal as ‘the end state toward clear indication that the ‘Nature of the Behaviours’ do-
which a human or non-human animal is striving: the main should be considered differently to the compo-
purpose of an activity or endeavour.’ [36] and defines nents of the TDF.
intention as ‘a conscious decision to perform a behav- In designing interventions, the TDF fits well with the
iour; a resolve to act in a certain way or an impulse for Behaviour Change Wheel (BCW) [46] referred to above.
purposeful action. In experiments, intention is often The BCW characterises the target behavior in terms of
equated with goals defined by the task instructions.’ [36]. Capability, Opportunity and Motivation (the COM-B sys-
Therefore ‘Goals’ tends to refer to an end state that can tem in the Behaviour Change Wheel), with Capability
be seen as a preferred outcome, whereas ‘Intentions’ is divided into psychological and physical capability, Oppor-
concerned with the resolve to initiate or terminate a tunity divided into social and physical opportunity and
Cane et al. Implementation Science 2012, 7:37 Page 13 of 17
http://www.implementationscience.com/content/7/1/37

Table 2 The refined framework based on results of the open and closed sort tasks
Domain (definition1) Constructs
1. Knowledge Knowledge (including knowledge of condition /scientific rationale)
(An awareness of the existence of something)
Procedural knowledge
Knowledge of task environment
2. Skills Skills
(An ability or proficiency acquired through practice)
Skills development
Competence
Ability
Interpersonal skills
Practice
Skill assessment
3. Social/Professional Role and Identity Professional identity
(A coherent set of behaviours and displayed personal qualities of an
Professional role
individual in a social or work setting)
Social identity
Identity
Professional boundaries
Professional confidence
Group identity
Leadership
Organisational commitment
4. Beliefs about Capabilities Self-confidence
(Acceptance of the truth, reality, or validity about an ability, talent, or Perceived competence
facility that a person can put to constructive use)
Self-efficacy
Perceived behavioural control
Beliefs
Self-esteem
Empowerment
Professional confidence
5. Optimism Optimism
(The confidence that things will happen for the best or that desired Pessimism
goals will be attained)
Unrealistic optimism
Identity
6. Beliefs about Consequences Beliefs
(Acceptance of the truth, reality, or validity about outcomes of a Outcome expectancies
behaviour in a given situation)
Characteristics of outcome expectancies
Anticipated regret
Consequents
7. Reinforcement Rewards (proximal / distal, valued / not valued, probable / improbable)
(Increasing the probability of a Incentives
response by arranging a dependent relationship, or contingency,
Punishment
between the response and a given stimulus)
Consequents
Reinforcement
Contingencies
Sanctions
Cane et al. Implementation Science 2012, 7:37 Page 14 of 17
http://www.implementationscience.com/content/7/1/37

Table 2 The refined framework based on results of the open and closed sort tasks (Continued)
8. Intentions Stability of intentions
(A conscious decision to perform a behaviour or a resolve to act in Stages of change model
a certain way)
Transtheoretical model and stages of change
9. Goals Goals (distal / proximal)
(Mental representations of outcomes or end states that an Goal priority
individual wants to achieve)
Goal / target setting
Goals (autonomous / controlled)
Action planning
Implementation intention
10. Memory, Attention and Decision Processes Memory
(The ability to retain information, focus selectively on aspects Attention
of the environment and choose between two or more alternatives)
Attention control
Decision making
Cognitive overload / tiredness
11. Environmental Context and Resources Environmental stressors
(Any circumstance of a person's situation or environment that Resources / material resources
discourages or encourages the development of skills and
abilities, independence, social competence, and adaptive behaviour) Organisational culture /climate
Salient events / critical incidents
Person x environment interaction
Barriers and facilitators
12. Social influences Social pressure
(Those interpersonal processes that can cause individuals to Social norms
change their thoughts, feelings, or behaviours)
Group conformity
Social comparisons
Group norms
Social support
Power
Intergroup conflict
Alienation
Group identity
Modelling
13. Emotion Fear
(A complex reaction pattern, involving experiential, behavioural, Anxiety
and physiological elements, by which the individual attempts to deal
with a personally significant matter or event) Affect
Stress
Depression
Positive / negative affect
Burn-out
14. Behavioural Regulation Self-monitoring
(Anything aimed at managing or changing objectively observed Breaking habit
or measured actions)
Action planning
1
All definitions are based on definitions from the American Psychological Associations’ Dictionary of Psychology [36].

Motivation divided into reflective and automatic motiv- experts in behavior change, with 100% agreement (Table 3).
ation. The domains from the refined framework have been Use of the COM-B may help identify the TDF domains that
independently mapped onto the COM-B segments by three are likely to be important in changing behaviour. By
Cane et al. Implementation Science 2012, 7:37 Page 15 of 17
http://www.implementationscience.com/content/7/1/37

starting with a behavioural analysis such as this, interven- instruments [53], their exclusion from the refined frame-
tion designers can be selective about the domains they in- work seems warranted. The remaining constructs stand
vestigate to inform the nature of the intervention. as a more defined, focused set of constructs that are
Research using the TDF has identified lack of know- more relevant to behaviour change theory and more pre-
ledge as a potential barrier to a number of professional cisely partitioned into domains. Within these remaining
health behaviours, including hand hygiene [17], changing constructs, there are also a number of constructs that
transfusion practice [19], and the adoption of tobacco appear in more than one domain. Such allocations indi-
use cessation counseling in dental practices [21]. How- cate the relevance of constructs across different domain
ever, for most health-related behaviours that are the contexts. For example, ‘Action Planning’ appears in both
target of theoretically-based behaviour change interven- the ‘Goals’ domain and the ‘Behavioural Regulation’ do-
tions (e.g., smoking, healthy eating, physical activity), main and can be considered as being influential in
knowledge is not an important source of variance [47- achieving a particular goal (e.g. I plan to achieve goal
52]. This may be why participants did not identify a x through specific actions) and also in regulating be-
separate domain for knowledge, but that it has been haviour (e.g. in a certain situation I plan to behave in
identified as an important influence on some health pro- a particular way).
fessional behaviours. We therefore recommend that Two domains showed weak clustering: ‘Environmental
knowledge be assessed along with the other TDF Context and Resources’ and ‘Behavioural Regulation’.
domains. However, these domains, alongside the domain of
Of the original 112 unique constructs in the TDF, 34 ‘Knowledge’, were comprised of constructs consistently
have been removed. They appear to be a mixture of assigned to them when the original domain labels were
rather vague constructs (e.g., Mindsets), very general presented in the closed sort task. This suggests that
constructs (e.g., Review), ambiguous constructs (e.g., people are clear about the constructs within these
Commitment), and infrequently used constructs in domains when the domain labels are present. A second
behaviour change theory (e.g., Generating alternatives). limitation is that the refined framework is limited to the
Because constructs that are ‘poorly defined’, ‘undifferenti- constructs identified in the original framework. Whilst
ated’, and ‘imprecisely partitioned’ have previously been the current range of component constructs is quite ex-
found to influence the content validity of assessment tensive, it does not cover all theories of behaviour
change [54], and future research is likely to identify
others that are important to behaviour change. Just as
Table 3 Mapping of the Behaviour Change Wheel’s the current framework is an advance on the 2005
COM-B system to the TDF Domains version, so future work is likely to improve it further.
COM-B TDF Domain The issue of how to evaluate appropriateness and quality
component
of theories in given contexts is an under-researched area,
Capability Psychological Knowledge
but one that is beginning to be addressed [54].
Skills
Memory, Attention and Decision
Processes Conclusions
Behavioural Regulation Through a three-step validation process, the present
Physical Skills research has identified a refined version of the original
Opportunity Social Social Influences TDF. This refined framework contains 14 domains and
Physical Environmental Context and 84 component constructs. The strength of the frame-
Resources work validation stems from the methods used. Both the
Motivation Reflective Social/Professional Role & Identity closed and open sort task methods alongside DCV and
Beliefs about Capabilities Fuzzy Cluster Analysis have provided complementary
methods for examining the structure of the original
Optimism
framework. DCV methods assessed the confidence of
Beliefs about Consequences
allocation of constructs to the described domains, and
Intentions the Fuzzy Cluster Analysis led to a refinement of the
Goals structure of the framework. The TDF has proved useful
Automatic Social/Professional Role & Identity across a number of healthcare systems and this empiric-
Optimism ally-based refinement lays the basis for stronger explana-
tory and predictive power, and therefore increased
Reinforcement
usefulness in informing interventions to improve imple-
Emotion
mentation and bring about other behaviour change.
Cane et al. Implementation Science 2012, 7:37 Page 16 of 17
http://www.implementationscience.com/content/7/1/37

Additional files 9. Michie S, Prestwich A (2010) Are interventions theory-based?

Development of a theory coding scheme. Heal Psychol 29:1–8
10. Davies P, Walker AE, Grimshaw J (2010) A systematic review of the use of
Additional file 1: Constructs from the original Theoretical Domains
theory in the design and implementation strategies and interpretation
Framework and associated definitions [55-68].
of the results of rigorous evaluations. Implement Sci 5:1–6
Additional file 2: Instructions and additional questions given to 11. Weinstein N (1993) Testing four competing theories of health-protective
participants. Instructions, consent information and additional questions behaviour. Heal Psychol 12:324–333
given to participants. 12. Nigg CR, Allegrante JP, Ory M (2002) Theory-comparison and multiple-
Additional file 3: Mantel correlation values by participant. Mantel behaviour research: common themes advancing health behaviour
correlation coefficients for each participant. research. Health Ed Res: Theory Prac 17:670–679
13. Michie S, Johnston M, Abraham C, Lawton R, Parker D, Walker A, on behalf
Additional file 4: Comparison of the refined framework, closed sort task,
of the ‘Psychological Theory’ Group (2005) Making psychological theory
and open sort task groupings with related mean confidence ratings,
useful for implementing evidence based practice: a consensus approach.
Intra-Class Correlation (ICC) values and silhouette values.
Qual Saf Health Care 14:26–33
Additional file 5: Membership values (%) for each construct to each 14. McKenzie JE, French SD, O’Connor DA, Grimshaw J, Mortimer D, Michie S,
domain cluster defined in the cluster analysis. Francis J, Spike N, Schattner P, Kent PM et al (2008) Implementing a clinical
practice guideline for acute low back pain evidence-based management
in general practice (IMPLEMENT): cluster randomised controlled trial
Competing interests study protocol. Implement Sci 3:11
SM and DOC are both Associate Editors of Implementation Science. 15. McKenzie JE, O’Connor DA, Page MJ, Mortimer D, French SD, Walker BF,
Keating JL, Grimshaw JM, Michie S, Francis JJ, Green SE (2010) Improving
Acknowledgements the care for people with acute low-back pain by allied health
We would like to thank the MRC Population Health Sciences Research professionals (the ALIGN trial): a cluster randomised trial protocol.
Network for funding this work, Daniel and Matthew West (www. Implement Sci 5:86
britishwebsites.net) for website design, Henry Potts (University College 16. French SD, Green SE, O’Connor D, McKenzie JE, Francis JJ, Michie S,
London) for statistical advice and Jill Francis (University of Aberdeen) and Buchbinder R, Schattner P, Spike N, Grimshaw JM: Developing theory-
Lou Atkins (Univerisity College London) for helpful comments on an earlier informed behaviour change interventions to implement evidence into
draft. DOC is supported by an Australian National Health and Medical practice: a systematic approach using the Theoretical Domains
Research Council Public Health Fellowship. Framework. For submission to Implementation Science in this series 2011.
17. Dyson J, Lawton R, Jackson C, Cheater F (2010) Does the use of a
Author details theoretical approach tell us more about hand hygiene behaviour? The
1
School of Psychology, Keynes College, University of Kent, Canterbury, Kent barriers and levers to hand hygiene. J Infection Prev 12:17
CT2 7NP, UK. 2School of Public Health and Preventive Medicine, Faculty of 18. Francis J, Stockton C, Eccles MP, Johnston M, Cuthbertson BH, Grimshaw
Medicine, Nursing and Health Sciences, Monash University, The Alfred JM, Hyde C, Tinmouth A, Stanworth SJ (2009) Evidence-based selection of
Centre, 99 Commercial Road, Melbourne, VIC 3004, Australia. 3Research theories for designing behaviour change interventions: Using methods
Department of Clinical, Centre for Outcomes Research and Effectiveness based on theoretical construct domains to understand clinicians’ blood
(CORE), Education and Health Psychology, University College London, 1-19 transfusion behaviour. Br J Heal Psychol 14:625–646
Torrington Place, London, WC1E 7HB, UK. 19. Francis J, Tinmouth A, Stanworth SJ (2009) Using theories of behaviour
change to understand transfusion prescribing three clinical contexts in
two countries: development work for and implementation trial.
Authors’ contributions Implement Sci 4:70
JC conducted preparation of materials, data collection, data analysis, and drafted 20. Michie S, Pilling S, Garety P, Whitty P, Eccles MP, Johnston M, Simmons J
the manuscript. DOC and SM commented on and aided in the drafting of the (2007) Difficulties implementing a mental health guideline: an
manuscript. All authors read and approved the final manuscript. exploratory investigation using psychological theory. Implement Sci 2:8
21. Amemori M, Korhonen T, Kinnunen T, Michie S, Murtomaa H (2011)
Received: 1 September 2011 Accepted: 24 April 2012 Enhancing implementation of tobacco use prevention and cessation
Published: 24 April 2012 counselling guideline among dental providers: a cluster randomised
controlled trial. Implement Sci 6:13
References 22. Paul CL (2008) A modified approach to a new card sorting methodology.
1. Clinical Effectiveness Research Agenda Group (CERAG): An Implementation J Usability Stud 4:7–30
Research Agenda: A report prepared for the High Level Group on Clinical 23. Whaley AL, Longoria RA (2009) Preparing card sort data for
Effectiveness. London: CERAG; 2008. multidimensional scaling analysis in social psychological research: a
2. World Health Authority (WHA): World Health Assembly Resolution 51.12: methodological approach. J Soc Psychol 149:105–115
Health Promotion. 2010, Accessed from [http://www.who.int/ 24. Trochim WMK, Cook JA, Setze RJ (1994) Using concept mapping to
healthpromotion/wha51-12/en/], 27th July 2010. develop a conceptual framework of staff's views of a supported
3. Newman M, Papadopoulous L, Sigsworth J (1998) Barriers to evidence- employment program for individuals with severe mental illness. J
based practice. Clin Eff Nurs 2:11–18 Consult Clin Psychol 62:766–775
4. McKenna H, Ashton S, Keeney S (2004) Barriers to evidence based practice 25. Morrow D, Leirer V, Altieri P, Tanke E (1991) Elders’ schema for taking medication:
in primary care: a review of the literature. Int J Nurs Stud 41:369–378 implications for instruction design. J Gerontol 46:378–385
5. Grol R, Grimshaw J (2003) From evidence to best practice: effective 26. Niven K, Totterdell P, Holman D (2009) A classification of controlled
implementation of change in patients’ care. Lancet 362:1225–1230 interpersonal affect regulation strategies. Emotion 9:498–509
6. Abraham C, Kelly MP, West R, Michie S (2009) The UK national institute for 27. Chollet S, Lelièvre M, Abdi H, Valentin D (2011) Sort and Beer: Everything
health and clinical excellence public health guidance on behavior you wanted to know about the sorting task but did not dare to ask.
change: a brief introduction. Psychology Health Med 14:1–8 Food Quality and Preference 22:507–520
7. Albarracin D, Gillette JG, Earl AN, Glasman LR, Durantini MR, Ho MH (2005) A 28. Reed CL, McGoldrick JE, Shackelford R, Fidopiastis CM (2004) Are human
test of major assumptions about behavior change: a comprehensive bodies represented differently from other objects? Experience shapes
look at the effects of passive and active HIV-prevention interventions object representations. Vis Cogn 11:523–550
since the beginning of the epidemic. Psychol Bull 131:856–897 29. Dixon D, Pollard B, Johnston M (2007) What does the chronic pain grade
8. Noar SM, Zimmerman RS (2005) Health behavior theory and cumulative questionnaire measure? Pain 130:249–253
knowledge regarding health behaviours: are we moving in the right 30. Lynn MR (1986) Determination and quantification of content validity.
direction? Health Ed Res: Theory Pract 20:275–290 Nurs Res 35:382
Cane et al. Implementation Science 2012, 7:37 Page 17 of 17
http://www.implementationscience.com/content/7/1/37

31. Waltz CW, Strickland OL, Lenz ER (1991) Measurement in nursing research. F. 58. Oxford English Dictionary Online [www.oed.com]
A. Davis Co, Philadelphia, PA 59. Wikipedia [www.en.wikipedia.org/wiki]
32. Dixon D, Johnston M, McQueen M, Court-Brown C (2008) The Disabilities 60. Ajzen I (1991) The Theory of Planned Behavior. Organizational Behavior
of the Arm, Shoulder and Hand Questionnaire (DASH) can measure the and Human Decision Processes 50:179–211
impairment, activity limitations and participation restriction constructs 61. Ogden J (2000) Health Psychology: a textbook (2nd Edition). Open University
from the International Classification of Functioning, Disability, and Press, Buckingham
Health (ICF). BMC Musculoskelet Disord 9:114 62. World Health Organization (WHO) [www.who.int/mental_health/
33. Carmines EG, Zeller RA (1979) Reliability and Validity Assessment. Sage management/depression/definition/en/]
Publications, Newbury Park, CA 63. Dictionary.com [http://dictionary.reference.com/]
34. Kaufman L, Rousseeuw PJ (1990) Finding groups in data: an introduction to 64. Forsyth DR (2006) Group Dynamics (5th Edition). Wadsworth, Cengage
data analysis. Wiley, New York Learning
35. Pollard B, Johnston M, Dieppe P (2006) What do osteoarthritis heath 65. Alicke MD: Evaluating social comparison targets. In Handbook of Social
outcome instruments measure? Impairment, activity limitation or Comparison: Theory and Research. Edited by Suls JM, Wheeler L. New York:
participation restriction? J Rheumatol 33:757–63 Plenum Publishing; 2000: 271–293.
36. American Psychological Association (APA): APA Dictionary of Psychology. 66. YourDictionary.com [www.yourdictionary.com/]
Washington, DC: American Psychological Association; 2007. 67. The Free Dictionary [www.thefreedictionary.com/]
37. Kendall MG, Babington Smith B (2009) The Problem of m Rankings. Ann 68. Proschaska JO, DiClemente CC: The Transtheoretical Approach: Crossing
Math Stat 10:275–287 Traditional Boundaries of Change. Homewood IL: J.Irwin.
38. Paradis E, Bolker B, Claude J, Sien Cuong H, Desper R, Durand B, Dutheil J,
Gascuel O, Heilbl C, Lawson D, Lefort V, Legrande, Lemon J, Noel Y, doi:10.1186/1748-5908-7-37
Nylander J, Opgen-Rhein R, Popescu A A, Schliep K, Strimmer K, de Vienne Cite this article as: Cane et al.: Validation of the theoretical domains
D: Package ‘ape’: Analysis of Phylogenetics and Evolution. URL [http:// framework for use in behaviour change and implementation research.
cran.r-project.org/web/packages/ape/ape.pdf]. 2011, 1–222. Implementation Science 2012 7:37.
39. R Development Core Team: R: A language and environment for statistical
computing. Vienna, Austria: R Foundation for Statistical Computing;
ISBN 3-900051-07-0; URL [http://www.R-project.org/]; 2011.
40. Legendre P: CADM.global: Congruence among distance matrices. In
Paradis E, Bolker B, Claude J, Sien Cuong H, Desper R, Durand B, Dutheil J,
Gascuel O, Heilbl C, Lawson D, Lefort V, Legrande, Lemon J, Noel Y,
Nylander J, Opgen-Rhein R, Popescu A A, Schliep K, Strimmer K, de Vienne
D: Package ‘ape’: Analysis of Phylogenetics and Evolution. URL [http://
cran.r-project.org/web/packages/ape/ape.pdf]. 2011, 1–222.
41. Maechler M: Package ‘cluster’: Cluster Analysis Extended Rousseeuw et al.
Version 1.14.0. URL [http://cran.r-project.org/web/packages/cluster/cluster.
pdf]. 2011.
42. Rousseeuw PJ (1987) Silhouettes: a graphical aid to the interpretation
and validation of cluster analysis. J Comput Appl Math 20:53–65
43. McGraw K, Wong SP (1996) Forming inferences about some intraclass
correlation coefficients. Psychological Methods 1:30–46
44. Landis JR, Koch GG (1977) The measurement of observer agreement for
categorical data. Biometrics 33:159–174
45. Hochberg Y (1988) A sharper Bonferroni procedure for multiple tests of
significance. Biometrika 75:800–802
46. Michie S, van Stralen M, West R (2011) The Behaviour Change Wheel: a
new method for characterising and designing behaviour change
interventions. Implement Sci 6:42
47. Kenkel DS (1991) Health behaviour, health knowledge, and schooling.
J Polit Econ 99:287–305
48. Klesges RC, Somes G, Pascale RW, Klesges LM, Murphy M, Brown K, Williams
E (1988) Knowledge and beliefs regarding the consequences of cigarette
smoking and their relationships to smoking status in a biracial sample.
Heal Psychol 7:387–401
49. Eckhardt L, Woodruff SI, Elder JP (1994) A longitudinal analysis of
adolescent smoking and its correlates. J Sch Heal 64:67–72
50. Dishman RK, Sallis JF, Orenstein DR (1985) The determinants of physical
activity and exercise. Public health reports 100:158–171
51. Baranowski T, Cullen KW, Baranowski J (1999) Psychosocial correlates of
dietary intake: advancing dietary intervention. Annu Rev Nutr 19:17–40
52. Shapiro J, Radecki S, Charchian AS, Josephson V (1999) Sexual behavior Submit your next manuscript to BioMed Central
and AIDS-related knowledge among community college students in and take full advantage of:
Orange County, California. J Community Health 24:29–43
53. Haynes SN, Richard DCS, Kubany ES (1995) Content validity in • Convenient online submission
psychological assessment: a functional approach to concepts and
• Thorough peer review
methods. Psychol Assess 7:238–247
54. Hobbs L, Campbell R, Hildon Z and Michie S. Behaviour change theories • No space constraints or color figure charges
across Psychology, Sociology, Anthropology and Economics: A • Immediate publication on acceptance
Systematic Review. Psychology and Health 2011, 26:Supplement 1.
55. Taylor SE (2003) Health Psychology (5th Edition). MacGraw-Hill, Boston • Inclusion in PubMed, CAS, Scopus and Google Scholar
56. Sutherland S (1995) The Macmillan Dictionary of Psychology (2nd Edition). • Research which is freely available for redistribution
Palgrave Macmillan, London
57. WordNet [www.wordnet.princeton.edu/perl/webwn]
Submit your manuscript at
www.biomedcentral.com/submit