Post on 29-Apr-2023
Accepted Manuscript
Guideline uptake is influenced by 6 implementability domains for creating andcommunicating guidelines: A Realist Review
Monika Kastner, Onil Bhattacharyya, Leigh Hayden, Julie Makarski, Elizabeth Estey,Lisa Durocher, Ananda Chatterjee, Laure Perrier, Ian D. Graham, Sharon E. Straus,Merrick Zwarenstein, Melissa Brouwers
PII: S0895-4356(15)00007-4
DOI: 10.1016/j.jclinepi.2014.12.013
Reference: JCE 8787
To appear in: Journal of Clinical Epidemiology
Received Date: 7 March 2014
Revised Date: 21 November 2014
Accepted Date: 8 December 2014
Please cite this article as: Kastner M, Bhattacharyya O, Hayden L, Makarski J, Estey E, DurocherL, Chatterjee A, Perrier L, Graham ID, Straus SE, Zwarenstein M, Brouwers M, Guideline uptake isinfluenced by 6 implementability domains for creating and communicating guidelines: A Realist Review,Journal of Clinical Epidemiology (2015), doi: 10.1016/j.jclinepi.2014.12.013.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service toour customers we are providing this early version of the manuscript. The manuscript will undergocopyediting, typesetting, and review of the resulting proof before it is published in its final form. Pleasenote that during the production process errors may be discovered which could affect the content, and alllegal disclaimers that apply to the journal pertain.
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
1
Guideline uptake is influenced by 6 implementability domains for creating
and communicating guidelines: A Realist Review
Monika Kastner
1*, Onil Bhattacharyya
1,2, Leigh Hayden
1, Julie Makarski
3, Elizabeth Estey
4, Lisa
Durocher3, Ananda Chatterjee
1, Laure Perrier
1,5, Ian D Graham
6, Sharon E Straus
1,2,7, Merrick
Zwarenstein8, Melissa Brouwers
3
1Li Ka Shing Knowledge Institute (LKSKI) of St. Michael's Hospital, Toronto, Ontario, Canada
2Keenan Research Centre of LKSKI of St. Michael’s Hospital, Toronto, Ontario, Canada
3Department of Oncology, McMaster University, & Escarpment Cancer Research Institute,
Hamilton, Ontario, Canada 3Strategic Policy, Planning & Initiatives, Health Services, Region of Peel, Toronto, Ontario,
Canada 5Continuing Education and Professional Development, Faculty of Medicine, University of
Toronto, Toronto, Ontario, Canada 6Department of Epidemiology and Community Medicine, University of Ottawa, & Ottawa
Hospital Research Institute, Ottawa, Ontario, Canada 7Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada
8Department of Family Medicine, Schulich School of Medicine and Dentistry, Western
University, London, Ontario, Canada
*Correspondence to:
Monika Kastner
Li Ka Shing Knowledge Institute of St. Michael’s Hospital
209 Victoria Street, Room 722
Toronto, ON, M5B 1W8
Phone: 416-864-6060, x77367
Email:monika.kastner@utoronto.ca
Funding
This study was funded by Knowledge Translation Canada through a Canadian Institutes of
Health Research (CIHR)/Canadian Foundation for Innovation (CFI) grants, and a CIHR
AGREE-A3 grant.
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
2
ABSTRACT
Objective: To identify factors associated with the implementability of clinical practice
guidelines (CPGs), and to determine what characteristics improve their uptake.
Study design and setting: We conducted a Realist Review, which involved searching multiple
sources (e.g., databases, experts) to determine what about guideline implementability works, for
whom, and under what circumstances. Two sets of reviewers independently screened abstracts
and extracted data from 278 included studies. Analysis involved the development of a codebook
of definitions, validation of data, and development of hierarchical narratives to explain guideline
implementability.
Results: We found that guideline implementability is associated with two broad goals in
guideline development: 1) creation of guideline content, which involves addressing the domains
of Stakeholder involvement in CPGs, Evidence synthesis, Considered judgment (e.g., clinical
applicability), and Implementation feasibility; and 2) the effective communication of this content,
which involves domains related to fine-tuning the CPG’s Message (using simple, clear and
persuasive language) and Format.
Conclusions: Our work represents a comprehensive and interdisciplinary effort toward better
understanding which attributes of guidelines have the potential to improve uptake in clinical
practice. We also created codebooks and narratives of key concepts, which can be used to create
tools for developing better guidelines to promote better care.
Keywords: Guideline implementability; Guideline adherence; Clinical practice guideline;
Systematic Review; Realist review
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
3
What is new?
Key finding
• We identified six domains of guideline implementability
hypothesized to affect uptake of recommendations within 2
broad categories: 1) the “creation” of guideline content, which
involves addressing Stakeholder involvement in guidelines,
Evidence synthesis, Considered judgment, and Implementation
feasibility; and 2) the “communication” of this content by fine-
tuning the guideline’s Message and Format
What this
adds to what
is known
• Building on the work of others, this is the first systematic
review to investigate guideline implementability from a
comprehensive and multi-disciplinary perspective (i.e.,
psychology, management, and human factors engineering)
• Our analysis moves beyond the medical and implementation
science literatures (which tend to focus on the creation of
content), and incorporates other disciplinary content, which
expands understanding of the relevant factors, particularly those
related to the communication of content
What is the
implication,
what should
change now?
• Guidelines summarize clinical evidence in order to inform
clinicians’ decision making, but how they are developed and
written influences how often they are used
• Our work represents an important, comprehensive and
interdisciplinary effort toward better understanding of which
attributes of guidelines have the potential to improve uptake in
clinical practice
• We created narratives of key concepts, which can be used to
develop tools to build better guidelines and promote better care
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
4
INTRODUCTION
Recent efforts in the clinical practice guideline (CPG) research enterprise have focused
on identifying factors that can be targeted to increase the uptake of recommendations to improve
patient outcomes and strengthen delivery systems. These efforts have included investigations
into extrinsic (changing the practice setting to facilitate recommendation use) and intrinsic
approaches (changing the guideline itself) to improving CPGs. Both approaches are needed, but
given the costs and context dependence of extrinsic approaches, investigation of intrinsic
approaches may lead to solutions at minimal cost and that may be more broadly applicable and
feasible. Shiffman et al. have referred to ‘implementability’ as the perceived characteristics of
guidelines that predict the relative ease of their implementation1. Gagliardi et al. developed a
framework of guideline implementability, which was tested with 20 different CPGs, and found
that elements related to guideline format (e.g., guidelines that provide summary versions) and
content (e.g., clinical considerations to individualize recommendations) provided the best
opportunities to modify CPGs for improved uptake2. Grol et al. found that guidelines that are
compatible with existing norms among the target group for implementation and those that do not
demand too much change to existing routines, extra resources, or acquisition of new knowledge
and skills were used more3. Michie et al. suggest that clarity, specificity of behavioural
instructions, and specific plans are important to get physicians to follow guidelines but these
factors have largely been overlooked4,5
.
Together, these contributions have served as an excellent foundation and have enabled a
dialogue within the health services research community about guideline implementability.
However, in developing this concept and understanding it further, the field has mostly focused
their efforts on the medical literature. In doing so, it has largely ignored relevant paradigms from
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
5
other disciplines. For example, in the areas of social, cognitive and health psychology, there have
been decades of research developing theory and models to explain behavior change, persuasion,
motivation, and communication styles6-8.
There has been much discussion around improving the rigor of guidelines9, and
consequently, there are a number of tools to increase how the evidence supporting guideline
recommendations is synthesized, analyzed, and presented. These include AGREE II10
,
GRADE11
, GLIA1, ADAPTE
12 and CAN-IMPLEMENT
13. However, these tools are mostly
informed by the medical literature1,2
, and target methodological and reporting concerns.
Currently, no resources take a comprehensive view of all factors relevant to guideline
implementability and investigate this from other disciplines focused on changing human
behavior, such as psychology, marketing, design and human-factors engineering. To better
understand the concept of implementability and the relationship between characteristics of
guidelines and their uptake by clinicians (who represent a primary end users of CPGs), our
primary objective was to identify factors associated with the implemenatbility of CPGs and
recommendations through a comprehensive and multidisciplinary perspective. Our secondary
objective was to determine what characteristics are posited to improve uptake by whom, and
under what circumstances.
METHODS
We conducted a Realist Review14
, which is an explicitly theory-driven approach to the
synthesis of evidence as it seeks to interrogate the underlying mechanisms of the programs or
interventions being studied14
. We report our methods and findings according to the RAMESES
(Realist And Meta-narrative Evidence Synthesis: Evolving Standards) criteria for the publication
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
6
standards of realist reviews16
, and include a flow diagram of our methods for increased clarity
(Figure 1).
[Figure 1 about here]
Search strategy
Consistent with methodological standards, we used a multiple search strategy that
consisted of 5 iterative stages of searching (see Figure 2): Stage 1: We consulted guideline
development and knowledge translation (KT) experts to identify a set of core articles in guideline
implementability supplemented with a background search by an information specialist in
MEDLINE and EMBASE (up to the year 2010) using the search terms: “implement*, “clinical
practice guidelines” and “knowledge translation”. Our candidate theories (i.e., reasons for poor
implementation of guidelines) were identified through consultation with guideline and KT
experts: i) Guidelines are not used in part because of specific perceived guideline characteristics
affecting uptake by clinicians (e.g., too complex or too difficult to follow); ii) There are tradeoffs
between various guideline attributes that facilitate or hinder uptake (e.g., increasing specificity
may reduce applicability). Stage 2: We sought the expertise of seven content experts known
amongst our team in Guideline development and KT (n = 3), Psychology (n = 2), Management
(n = 1), and Human factors engineering (n = 1) to suggest seminal articles or to provide
direction to where to look for articles in these disciplines (a method akin to snowball sampling in
qualitative research). Stage 3: We expanded our search by looking for the related articles of our
core Medicine citations identified in Stage 1 (n = 76). Reviewers independently selected a sub-
set of priority articles (highly cited or written by known researchers in guideline
implementability; n = 22), which were entered into the PubMed ‘related-articles’ interface
resulting in 1677 citations. Two reviewers scanned this yield independently using our inclusion
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
7
criteria of which 131 were deemed potentially relevant. Stage 4: We searched the reference lists
of articles from stages 1 and 2 to determine if we had reached theoretical saturation. Duplicate
review of a random subset of citations did not identify new concepts or understandings, so we
did not pursue this further. Stage 5: We continuously sought out other potentially relevant
articles including the grey literature (e.g., unpublished documents and websites such as the
Guideline International Network). During all stages, we conducted saturation testing iteratively,
by asking at regular intervals of searching whether the latest sample of literature has added
anything new to our understanding and whether further searching is likely to add new
knowledge14
.
[Figure 2 about here]
Article selection process
Two sets of reviewers independently screened abstracts to look for our candidate theories
and to test for relevance. We included all article types written in English, and excluded opinion-
driven reports (editorials, commentaries, and letters) unless authored by an individual identified
by our team as an “expert” in the field of guideline implementability. Our unit of analysis was
not dependent on study design because relevant data could exist in any section of the article (e.g.,
Abstract; Introduction; Discussion), so we did not assess study quality14.
Data extraction
We developed a standardized data extraction form that was pilot tested independently by
two reviewers using five articles. Data was extracted by two primary reviewers and audited by
another two reviewers with disagreements resolved through group consensus on study
characteristics, discipline (e.g., medicine, psychology), guideline attribute name and definition,
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
8
attribute operationalization (how the attribute functions within the discipline or context), attribute
relationship with guideline uptake, and any potential tradeoffs.
Data organization
We identified 1736 guideline attributes, which were initially sorted by the same name or
name root (e.g., valid/validity). Two groups of reviewers took the same list of sorted attributes
and independently clustered them into logical categories to determine how attributes are related
to one another. We then combined groups of similar attributes (including their synonyms and
antonyms) that conceptually “fit” within a larger theme, and created a label and description for
each category. Table 1 describes the operationalization of this process. We compared the
categorizations between the two reviewer groups for agreement by documenting “agreed” and
“divergent” classifications. This resulted in deriving a common set of 27 categories (and their
attributes) grouped into 5 broad domains that we conceptualized as associated with guideline
uptake: Language, Format, Rigor of development, Feasibility, and Decision-making. We
considered this conceptualization as our “raw data” for analysis.
[Table 1 about here]
Analysis
Data analysis involved a 3-level process where data were further interrogated and refined
with each level of synthesis: Level 1: Development of a codebook: We developed (in duplicate)
a codebook of definitions for each of the five identified domains and attributes to determine
recurrent patterns of outcomes (i.e., what makes guidelines implementable) and their associated
mechanisms and contexts. We included attributes that were deemed “modifiable” by guideline
developers (e.g., the actionability of recommendations vs. the self-efficacy of the guideline end-
user); documented its definition, operationalization, and the discipline and context they were
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
9
studied; and noted any reported positive or negative relationships with guideline uptake and any
tradeoffs if existed. We subsequently looked at the data for consistency or divergence of
outcomes as a function of discipline. The various disciplines tended to focus on unique sets of
attributes; the data did not lend to a formal cross-discipline analysis; Level 2: Validation of data:
We determined which attributes were most appropriate for which domain/sub-domain to identify
the hierarchy of groupings and important relationships. We tested the validity of this
organization by administering an online survey to nine experts in KT and guideline development
known amongst our team. They were asked to review the organization, and to rename, rearrange
and condense as they saw fit. Accordingly, we collapsed and renamed some
attributes/categories/domains (e.g., the Language domain was renamed Message), and we added
a 6th domain (Stakeholder involvement) – see Table 2; Level 3: Hierarchical explanatory
narratives: We further expanded the knowledge derived from previous levels of analysis by
building up an explanation for what intrinsic factors about a guidelines promote (facilitator
attributes) or impede (barrier attributes) guideline use and uptake, under what circumstances this
happens (an explanation of situations/contexts/settings), the mechanism by which it happens (its
operationalization), and the outcomes and consequences of this. We did this by deriving
explanations for each of the broad domains and sub-domains from a summation of their smaller
parts.
[Table 2 about here]
RESULTS
Of 2550 potentially relevant articles identified, 350 articles were screened in full text and
278 articles contributed to the analysis from 170 journals across 7 disciplines: medicine (n =
188), psychology (n = 30), management/marketing/business (n = 23), human factors engineering
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
10
(n = 17), information technology/computer science (n = 10), graphic design (n = 7), and
sociology (n = 3). The three most common study designs were narrative reviews (n = 110),
qualitative studies (n = 37) and, observational studies (n = 32). See Figure 3 for the flow of
article selection.
Using our hierarchical narrative approach, we built an explanation for each of the six
domains (comprising 27 attributes) showing that guideline implementability is associated with
two broad goals during the guideline development process: the Creation of guideline content (4
domains) and the Communication of this content (2 domains) – See Table 3 and Appendix A.
[Table 3 about here]
I. Creation of content in guidelines (see Table 4 and Appendix B). This category includes 4
domains:
1) Stakeholder involvement: Findings showed that the credibility of a guideline affects its
likelihood of uptake. The key determinants of credibility are the range of stakeholders
involved17
, disclosure of any conflicts of interest and funding sources19-22
. The literature suggests
that addressing stakeholder involvement entails ensuring the appropriate composition and
relevant and unbiased expertise of the guideline development group, documenting the views and
preferences of the target population (patients, public), and defining target end users10,23; 2)
Evidence synthesis: Synthesizing the appropriate evidence is a complex task for guideline
developers, but appears to be the most developed aspect of guideline development in the
literature, primarily originating from Medicine. Rigorous evidence synthesis is emphasized
across medicine, nursing, health policy, and the approaches reflect a common understanding of
how to conduct this process. The literature reflects the attributes deemed necessary to enhance
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
11
guideline validity and reproducibility and comprises the consistent reporting of the elements that
need to be included in guidelines20,24-26, the execution of these elements26-28, and the currency of
guidelines29,30; 3) Considered judgment: The literature suggests that guideline developers need to
supplement evidence-based formulation of recommendations with considered judgment. This
involves the consideration of the evidence in light of other factors such the complex trade-offs
between competing benefits and harms and risks of different options for managing the disease or
condition, clinical applicability and contexts, the values and preferences of those for whom the
recommendations are intended (providers, patients, and developers), organizational needs, and
costs10,23,31-35
; 4) Implementation Feasibility: Feasibility involves the local applicability of
guidelines36, the consideration of resource constraints to make them more implementable10,37,38,
and the influence of the degree of novelty or familiarity of the interventions1,20. There is
sometimes a tension between feasibility and impact. A recommendation may be highly feasible
but unlikely to make much of a difference (it supports the status quo)39
. In contrast, a
recommendation may require far-reaching changes, and have enormous potential impact, but
have low feasibility10,27
.
II. The Communication of content to guideline end-users (Table 4 and Appendix B): This
category includes our final 2 domains: 5) Message: Studies suggest that to optimize the
messaging of a guideline and in turn its uptake, the language used in recommendations has to be
simple, clear, and persuasive. Literature spanned across all disciplines, with the psychology
literature focusing on how to reduce cognitive load and increase understanding and retention; the
management literature focusing on crafting convincing and salient arguments. It is also
suggested that the level of complexity of guidelines and recommendations is inversely
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
12
proportional to its adoption40-45
and compliance46,47
. The bodies of literature did not contradict,
but rather focus on different aspects of messaging; 6) Format: A number of formatting aspects of
the guideline were identified from the literature, which allows the various elements of guidelines
to become more explicit38
, promoting their use in practice2,49-52
. These include having: i) multiple
versions of guidelines. This refers to the evolution of guidelines, which typically progress from a
research-based tool to a short version for clinical use, and to a lay-language version for
patients53
; ii) specific components within guidelines (e.g., purpose, methods, recommendations);
and iii) the presentation (appearance) of guidelines, which involves layout, structure, and
information visualization. Some of the common design principles for scientific communication
have an empirical foundation, but many are derived from best practices and user preferences.
Since most formatting principles are based on cognitive processes, they are likely to be
generalizable across disciplines and contexts.
[Table 4 about here]
DISCUSSION
Our synthesis of 278 articles from a multidisciplinary body of literature identified six
domains of guideline implementability that affect uptake of recommendations within 2 broad
categories (the creation and communication of guideline content). Building on the work of
Shiffman and Gagliardi, we investigated guideline implementability from a broad perspective.
Shiffman et al developed GLIA by searching the literature for guideline attributes in 5 key
reports describing impact of factors on successful implementation and by examining 3
instruments for the appraisal of guideline quality1. Gagliardi et al focused on 18 healthcare
studies and created an implementability framework comprised of 22 elements organized into 8
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
13
domains (adaptability, usability, validity, applicability, communicability, accommodation,
implementation, and evaluation)2. We drew on a wide range of disciplines to further enhance this
work and involved appropriate experts in the review process. Our analysis moves beyond the
medical and implementation science literatures (which tend to focus on aspects of the Creation
of content), and incorporates other disciplinary content, which expands understanding related to
the Communication of content. Literature from human factors engineering adds information and
direction on how developers can structure guidelines to meet the specific needs of end-users and
mirror their work processes and approach to care. The business/management literature adds
insights and direction regarding the importance of persuasive and clear messaging while the
design literature outlines specific design principles, which are intended to improve the usability
and attractiveness of products. Literature from cognitive psychology alerts to the limitations of
information processing and provides some explicit strategies for developers to ease cognitive
load.
We achieved rigor in our review through transparent documentation of the
methodological approach15
as well as for our search strategy, study selection, analysis and
interpretation, including any deviations from the original method. We also provided a detailed
flow diagram for our study methods (Figure 1) and article selection process (Figure 2). We also
adopted a consensus-based approach to clarify difficult concepts, where smaller groups of
investigators discussed their assumptions related to the meaning of a concept, sought
clarification from discipline-specific experts if needed, and then reached consensus within the
broader study team on the applicability of such concepts in the context of guideline
implementability.
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
14
Our review suggests a few areas for further study. Guideline end-users commonly
complain about their length, and we know that simplifying language and content can improve
information processing and retention and improve perceived ease of use (and thus potentially
intention to use), so simplifying guidelines appears to be a worthy endeavor. However, we do
not know the threshold of simplification that could result in misinterpretation of guideline
recommendations. Determining what type of content ought to be simplified and in what ways
requires additional investigation to provide developers with guidance regarding a balanced
approach to creating and communicating guideline content. In addition, several issues of how to
operationalize the concept of Considered judgment remains unclear. An approach to help
developers facilitate the process is needed. Without practical procedures, it is likely that this step
will be conducted in a cursory fashion, or not at all.
Our work has some limitations. We chose a realist approach to develop theories about
what attributes of guidelines contribute to their implementability, for whom, and under what
circumstances. However, the literature was rarely explicit about context and circumstances of
guideline development and uptake, so it became difficult to develop and interrogate such theories
and to explore Context-Mechanism-Outcome relationships. Additionally, we had anticipated
exploring the similarities and differences between how the literatures understood guideline
implementability, but found few contradictions in the literature. Studies tended to focus on the
problems inherent in guidelines from different perspectives: usability and usefulness (human
factors engineering), information processing and retention (cognitive psychology), document
construction (design), consumer relations (management), and clinical relevance and importance
(medicine). As an alternative, we developed a conceptual framework to synthesize the data,
highlighting contextual issues through the domain narratives, positing relationships rather than
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
15
proving them. Although we sought the expertise of content experts in Psychology, Management
and Human factors engineering, we may have missed other relevant articles in these disciplines
as we did not systematically search these beyond our expert-identified and snowball sampling
techniques. To offset this risk, we performed a “related-articles” search of “core” articles.
Additionally, literature from Medicine may be over-represented as the bulk of our core articles
were identified from this discipline. However, given that guidelines are most heavily discussed
in the health/medicine literature, we thought this was an acceptable risk of bias. There may be
relevant concepts from other disciplines we didn’t include, and there are areas that are not well
covered within the disciplines we studied. However, we incorporated a rigorous, systematic, and
transparent methodology and consulted experts at several key points during the analysis as a
check for external validity.
Conclusions and recommendations
Guidelines summarize clinical evidence to inform clinicians’ decision making, but how
they are developed and written influences how they are used. Our work represents a
comprehensive and interdisciplinary effort toward better understanding which attributes of
guidelines have the potential to improve uptake in clinical practice. However each of these will
need to be empirically tested to identify the best “package” of attributes. We also created
narratives of key concepts, which can be used to develop tools to determine their impact on
building better guidelines aimed at increasing their uptake and promoting better care.
Acknowledgements
We thank the following individuals for providing content expertise: Jamie Brehaut (Psychology);
Jeremy Grimshaw (Knowledge Translation); Stephen Hanna, Holger Schunemann, Francoise
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
16
Cluzeau (Guideline development); Dilip Soman (Management); Mark Chignell (Human Factors
Engineering); Hyun Joo Lee (Design).
Ethical approval:
No ethics approval was required for this study.
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
17
REFERENCES
1. Shiffman R DJ, Brandt C, Essaihi A, Hsiao A, Michel G, O’Connell R. . The Guideline
Implementability Appraisal (GLIA): development of an instrument to identify obstacles to
guideline implementation. BMC Med Info Dec Mak. 2005;5:23.
2. Gagliardi A, Brouwers MC, Palda VA, Lemieux-Charles L, Grimshaw JM. How can we
improve guideline use? A conceptual framework of implementability. Implementation
Science. 2011;6:26.
3. Grol R DJ, Thomas S, Veld C, Rutten G, Mokink H. Attributes of clinical guidelines that
influence use of guidelines in general practice: Observational study. BMJ. 1998;317:858-
861.
4. Michie S, Johnston M. Changing clinical behaviour by making guidelines specific. BMJ.
02/07 2004;328(7435):343-345.
5. Michie S LK. Words that matter: increasing the implementation of clinical guidelines. Qual
Saf Health Care. 2005;14:367-370.
6. Rogers E. Diffusion of Innovations. New York: Free Press; 1995.
7. Rey G BF. The Expertise Reversal Effect: Cognitive Load and Motivational Explanations.
Journal of Experimental Psychology: Applied. 2011;17(1):33-48.
8. Wang A, Dowding T. Effects of visual priming on improving web disclosure to investors.
The Journal of Behavioral Finance. 2010;11(1):11-20.
9. Tricoci P AJ, Kramer JM, Califf RM, Smith SC Jr. . Scientific evidence underlying the
ACC/AHA clinical practice guidelines. JAMA. 2009;301(8):831-841.
10. Brouwers MC KM, Browman GP, Burgers JS, Cluzeau F, Feder G, et al. AGREE II:
Advancing guideline development, reporting and evaluation in health care. CMAJ.
2010;182(18):E839-842.
11. Atkins D, Best D, Briss PA, Eccles M, Falck-Ytter Y, Flottorp S, Guyatt GH, Harbour RT,
Haugh MC, Henry D, Hill S, Jaeschke R, Leng G, Liberati A, Magrini N, Mason J,
Middleton P, Mrukowicz J, O'Connell D, Oxman AD, Phillips B, Schünemann HJ, Edejer T,
Varonen H, Vist GE, Williams JW Jr, Zaza S; GRADE Working Group. Grading quality of
evidence and strength of recommendations. BMJ. 2004;328:1490.
12. Fervers B BJ, Haugh MC, Latreille J, Milka-Cabanne N, Paquet L, et al. Adaptation of
clinical guidelines: literature review and proposition for a framework and procedure. Int J
Qual Health Care. 2006;18(3):167-176.
13. Harrison MB GI, van den Hoek J, Dogherty EJ, Carley ME, Angus V. Guideline adaptation
and implementation planning: a prospective observational study. Implementation Science.
2013;8:49.
14. Pawson R GT, Harvey G, Walshe K. Realist review – a new method of systematic review
designed for complex policy interventions. . J Health Serv Res Policy. 2005;19(Suppl
1):S21-S34.
15. Kastner M EE, Perrier L, Graham ID, Grimshaw J, Straus SE, Zwarenstein M,
Bhattacharyya O. Understanding the relationship between the perceived characteristics of
clinical practice guidelines and their uptake: protocol for a realist review. Implementation
Science. 2011;6:69.
16. Wong G GT, Westhorp G, Buickingham J, Pawson R. RAMESES publication standards:
Realist synthesis. BMC Medicine. 2013;11:21.
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
18
17. Solberg LI, Brekke ML, Fazio CJ, et al. Lessons from experienced guideline implementers:
attend to many factors and use multiple strategies. The Joint Commission journal on quality
improvement. 2000;26(4):171-188.
18. Muhlhauser I. From authority recommendations to fact-sheets--a future for guidelines.
Diabetologia. 2010;53(11):2285-2288.
19. The ADAPTE Process: Resource Toolkit for Guideline Adaptation. Version 2.0. 2009.
Available at: http://www.g-i-n.net/document-store/working-groups-
documents/adaptation/adapte-resource-toolkit-guideline-adaptation-2-0.pdf . Accessed on
September 7, 2014.
20. Rosenfeld RM SR. Clinical practice guideline development manual: A quality driven
approach for translating evidence into action. Otolaryngology Head Neck Surg. 2009;140(6
Suppl 1):S1-S43.
21. American Academy of Pediatrics Steering Committee on Quality Ia, Management.
Classifying recommendations for clinical practice guidelines. Pediatrics. 2004;114(3):874-
877.
22. Tricoci P, Allen JM, Kramer JM, Califf RM, Smith SC, Jr. Scientific Evidence Underlying
the ACC/AHA Clinical Practice Guidelines. JAMA: The Journal of the American Medical
Association. 2009;301(8):831-841.
23. Horvath AR, Kis E, Dobos E. Guidelines for the use of biomarkers: principles, processes
and practical consideration. Scandinavian Journal of Clinical and Laboratory Investigations.
2010;70(1):109-116.
24. Grimshaw JM RI. Achieving health gain through clinical guidelines II: Ensuring guidelines
change medical practice. Quality in Health Care. 1994;3:45-52
25. Wolf M BD, Marbella AM, Casanova JE. . US Family Physicians' Experiences with Practice
Guidelines. Fam Med. 1998;30(2):117-121.
26. 27Hayward RSA WM, Tunis SR, Bass EB, Guyatt G. Users' guide to the medical literature:
VII. How to use clinical practice guidelines. Are the recommendations valid? JAMA
1995;274(7):570-574.
27. McAlister FA vDS, Padwal RS, Johnson JA, Majumdar SR. How evidence-based are the
recommendations in evidence-based guidelines? PLoS Medicine. 2007;4(8):e250.
28. Grimshaw JM, Russell IT. Effect of clinical guidelines on medical practice: a systematic
review of rigorous evaluations. Lancet. 11/27 1993;342(8883):1317-1322.
29. Institute of Medicine (IOM). Clinical Practice Guidelines We can Trust. Washington, DC:
The National Academies Press. Brief report available at:
http://www.iom.edu/Reports/2011/Clinical-Practice-Guidelines-We-Can-Trust.aspx.
Accessed on September 7, 2014.
30. Chou R. Using evidence in pain practice. Part I: Assessing quality of systematic reviews and
clinical practice guidelines. Pain Medicine. 2008;9(5):518-530.
31. Watine J FB, Nagy E, Onody R, Oosterhuis W, Bunting PS, Charet JC, Horvath AR.
Conflict between guideline methodologic quality and recommendation validity: A potential
problem for practitioners. Clinical Chemistry. 2006;52(1):65-72.
32. Oosterhuis WP BD, Watine J, Sandberg S, Horvath AR. Evidence-based guidelines in
laboratory medicine: Principles and methods. Clinical Chemistry. 2004;50(5):806-818.
33. Verkerk K vVH, Severens JL, Hendriks EJM, Burgers JS. Considered judgment in evidence-
based guideline development. International Journal for Quality in Health Care 2006;18(5):
365-369.
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
19
34. Bousquet J, Schunemann HJ, Zuberbier T, et al. Development and implementation of
guidelines in allergic rhinitis - an ARIA-GA2LEN paper. Allergy. 2010;65(10):1212-1221.
35. Schunemann HJ, Fretheim A, Oxman A. Improving the use of research evidence in
guideline development: 10. Integrating values and consumer involvement. Health Research
Policy and Systems. 2006;4(22).
36. Fervers B, Carretier J, Bataillard A. Clinical practice guidelines. Journal of Visceral
Surgery. 2010;147(6):e341-e349.
37. Keeley PW. Clinical guidelines. Palliative Medicine. 2003;17(4):368-374.
38. Schunemann HJ, Oxman A, Brozek J, et al. Grading quality of evidence and strength of
recommendations for diagnostic tests and strategies. BMJ. 2008;336(7653):1106-1110.
39. Redelmeier DA, Ferris LE, Tu JV, Hux JE, Schull MJ. Problems for clinical judgement:
introducing cognitive psychology as one more basic science. Canadian Medical Association
Journal. 2001;164(3):358-360.
40. Dahm P YL, Gallucci M, Simone G, Schunemann HJ. How to use a clinical practice
guideline. The Journal of Urology. 2009;181:472-479.
41. Scott SD, Plotnikoff RC, Karunamuni N, Bize R, Rodgers W. Factors influencing the
adoption of an innovation: an examination of the uptake of the Canadian Heart Health Kit
(HHK). Implementation Science. 2008;3(41).
42. Tornatzky LG, Klein KJ. Innovation Characteristics and Innovation Adoption-
Implementation: A Meta-Analysis of Findings. IEEE Transactions on Engineering
Management. 1982;EM-29(1):28-43.
43. Thompson RL, Higgins CA, Howell JM. Personal Computing: Toward a Conceptual Model
of Utilization. MIS Quarterly. 1991;15(1):125-143.
44. Ball K. Surgical smoke evacuation guidelines: compliance among perioperative nurses.
AORN Journal. 2010;92(2):e1-e23.
45. Cabana MD, Flores G. The role of clinical practice guidelines in enhancing quality and
reducing racial/ethnic disparities in paediatrics. Paediatric Respiratory Review.
2002;3(1):52-58.
46. Rashidian A EM, Russell I. Falling on stony ground? A qualitative study of impelemtation
of clinical guidlines' prescribing recommendations in primary care. Health Policy.
2008;85:148-161.
47. Gurses AP, Marsteller JA, Ozok AA, Xiao Y, Owens S, Pronovost PJ. Using an
interdisciplinary approach to identify factors that affect clinicians' complicance with
evidence-based guidelines. Critical Care Medicine. 2010;38(8 Suppl):S282-S291.
48. Shaneyfelt TM, Mayo-Smith MF, Rothwangl J. Are guidelines following guidelines? The
methodological quality of clinical practice guidelines in the peer-reviewed medical
literature. JAMA, the journal of the American Medical Association. 1999;281(20):1900-
1905.
49. Carlsen B GC, Pope C. Thou shalt versus thou shalt not: A meta-synthesis of GPs' attitudes
to clinical practice guidelines. British Journal of General Practice. 2007;57:971-978.
50. Tong A. Clinical guidelines: can they be effective? Nursing Times. 2001;97(9):III-IV.
51. Grimshaw JM, Russell IT. Effect of clinical guidelines on medical practice: a systematic
review of rigorous evaluations. Obstetrical & Gynecological Survey. 1994;49(7):469.
52. Stone TT, Schweikhart SB, Mantese A, Sonnad SS. Guideline attribute and implementation
preferences among physicians in multiple health systems. Quality Management in Health
Care. 2005;14(3):177-187.
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
20
53. Harris JS. Development, use, and evaluation of clinical practice guidelines. Journal of
Occupational and Environmental Medicine. 1997;39(1):23-34.
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
1
Table 1: Process that we used to categorize attributes using the Message domain as an example
Goal Steps Example
Organize, group,
and
appropriately
label similar or
“like” attributes
1. Group attributes that are antonyms • Complex/Simple
2. Group attributes that are synonyms • Unclear/Confusing
3. Group attributes with the same root • Specific/Specificity
• Validity/Valid
4. Sort database by attribute
Categorize
attributes into
logical domains
5. Are there commonalities among
attributes?
6. Is there a central theme or focus
among groups of attributes?
The following attributes can be grouped
into a category called “Clarity”
• Unambiguous
• Precise
• Specific
Go through each
domain to
determine sense
and fit of
attributes
7. Do the attributes belong within the
same cluster?
8. Can they be collapsed?
9. Use attribute definitions to make
these decisions
The following categories can be
collapsed:
• ”Complexity” with “Information
overload”
• “Actionability” with “Wording”
Develop a
definition for
domains
10. Based on their included attributes
and definitions, define and label the
cluster
The MESSAGE domain can be defined
as: The clarity, simplicity and
persuasiveness of the guideline
language
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
2
Table 2: Final list of attribute categories across 6 domains of guideline implementability
Category (N = 16) Major attributes Domain (N = 6)
Credibility of guideline
development group
Credibility Stakeholder
involvement
Disclosure of conflict
of interest
Conflict of interest; Transparency; Funding
sources; Editorial independence
Reporting of what is
needed
Scope; Patient preferences; Cost and resource
requirements; Outcomes data; Harms and
benefits
Evidence
synthesis
Execution of what is
needed
Evidence-based; Valid and reliable; Transparent
Updating of guidelines Updating; Currency
Clinical applicability Clinical relevance; Appropriateness of patient
population; Considered implementation
Considered
judgment
Values Guideline developer values;
Professional/provider values (clinical judgment,
clinical freedom); Patient values (acceptability,
patient preferences);
Local applicability Adaptation; Application tools Implementation
feasibility Resource constraints Availability of resources; Economic outcomes
Novelty Compatibility; Requires new knowledge and
skills
Simple Information overload; Complexity Message
Clear Actionability (specificity, ambiguity); Effective
writing
Persuasive Framing; Relative advantage
Multiple versions End-users; Versions (flat, dynamic); Document
type
Format
Components Components to include in guidelines (e.g.,
purpose, target audience, methods)
Presentation Layout of full document (placement, length);
Structure within sections (match the system to
the real world, sequential bundling); Information
visualization (information display [e.g.,
algorithms, pictures], information context [e.g.,
framing, vividness])
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
3
Table 3: Summary of hierarchical narratives to describe guideline implementability
GUIDELINE IMPLEMENTABILITY: To facilitate uptake, the process of guideline development has two broad aims: the creation of content and the communication of that content. The four domains of content creation are stakeholder involvement, evidence synthesis, considered judgment in formulating recommendations, and feasibility. The two domains of content communication relate to fine-tuning the message itself and its format.
Creation of Content: The four interrelated domains of content creation are stakeholder involvement (including credibility of the developers and disclosure of conflicts of interest), evidence synthesis (specifying what evidence is needed and how and when it is synthesized), considered judgment (including clinical applicability and values), and feasibility (local applicability, resource constraints, and novelty). These domains may be considered non-sequentially and iteratively.
Communication of content: Communication of guidelines entails fine-tuning both the message of the recommendations (through use of simple, clear, and persuasive language) and their format (through representation in multiple versions, inclusion of specific components, and effective layout and structure).
STAKEHOLDER INVOLVEMENT
EVIDENCE SYNTHESIS
CONSIDERED JUDGMENT IMPLEMENTATION FEASIBILITY
MESSAGE FORMAT
The guideline development group should have appropriate composition; its members should have relevant, unbiased expertise and suitable credibility, and potential conflicts of interest should be disclosed. The target population of end users (patients, the public) should be clearly defined, and their views and preferences considered. Ensuring stakeholder involvement during guideline development facilitates uptake.
To enhance guideline validity and reproducibility, the necessary evidence must be specified, the method of synthesis clearly defined (ideally evidence-based, valid, reliable, and transparent), and the timing of sequential syntheses appropriate (balancing timeliness of the guideline with stability over time).
The guideline development group must supplement evidentiary factors (quality, quantity, and consistency) with considered judgment, making complex trade-offs between the competing benefits and harms, side effects, and risks of various options for managing the disease or condition. They must also consider clinical applicability (whether the guideline responds to variability among patients) and the values and preferences of patients, developers, and care providers (i.e., the relative worth or importance of a health state and consequences such as benefits, harms, and costs of a decision).
Feasibility reflects local applicability (i.e., strategies for adapting recommendations to local conditions), consideration of resource constraints (availability of resources and other economic implications), and the influence of novelty of (or familiarity with) the guidelines, where novelty refers to the degree to which the recommendations propose behaviours considered unconventional by clinicians or patients). Feasible implementation of guidelines allow for flexibility in individual clinical decisions, are in agreement with users’ opinions and skills, and are suitable for routine use in intended settings.
Guideline messages should use simple, clear, and persuasive language. Simplicity can be achieved by limiting the number of elements, the number of steps within each recommendation, or the number of conditional factors influencing performance (to prevent the quantity of information from exceeding available cognitive capacity). Clarity is enhanced by using specific, unambiguous language and by applying a direct writing style, with active voice, suitable punctuation, short sentences, and bullet lists to convey series of points, and without awkward breaks, abbreviations, hyphenation, redundancy, or unnecessary jargon. Finally, the guideline messages should be clinically convincing and should be framed in terms of potential gains, to convey their relative advantage over any previous approach.
Guidelines may be formatted in multiple versions (e.g., research-based, information-gathering, analytical tool; briefer guide for clinical education; short version for point-of-care clinical use; lay-language version for patients). Formatting involves determining which components to include in various versions (e.g., scope and purpose, target audience, guideline development panel, update plan, and implementation considerations), their presentation (e.g., proper placement of visual elements and document length) and structure (i.e., matching the system to the real world and bundling); and the overall visualization of information.
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
4
Table 4: Facilitators and barriers for domains of guideline implementability*
Domain and examples of facilitators and barriers Actions for guideline developers
STAKEHOLDER INVOLVEMENT
Facilitators:
• wide range of stakeholders
• disclosure of conflicts of interest and funding sources
Barriers
• industry contributions
• recommendations based on expert opinion alone
• select unbiased participants with relevant
expertise
• declare competing interests and funding
sources
• document views and preferences of target
population
EVIDENCE SYNTHESIS
Facilitators: • consistent reporting of elements (e.g., phases of illness, resource requirements,
outcomes data, harms and benefits)
• identification of recommendations based on expert judgment or consensus
• maintenance of currency in relation to new evidence
Barrier: excessive frequency of revision
• report type and quality of evidence
• report methods of synthesis
• set timelines for guideline review
CONSIDERED JUDGMENT
Facilitators:
• perceived clinical relevance and appropriateness for patient population
• indicators of relative strength of recommendations
• definition of values that influenced recommendations
Barriers:
• lack of fit between clinicians’ experiences and recommendations
• lack of applicability to all or “typical” patients.
• use of low-quality, weak, or conflicting evidence
• state population(s) to which statements
apply
• disclose appropriateness or applicability of
• clearly state value judgments
• document process for managing dissent
IMPLEMENTATION FEASIBILITY
Facilitators:
• limit on recommendations that require large investment of time or resources
• availability of data regarding resource requirements and cost-effectiveness
• minimization of change required for users and systems
Barriers:
• requirement for new knowledge or skills
• unconventional changes
• inconsistency of changes with existing values, needs, and experiences of adopters
• formulate recommendations in terms of
measurable criteria and targets for quality
improvement
• identify costs and resource requirements
• specify competencies, training, and
technical specifications required
• include economic data
MESSAGE
Facilitators
• actionability
• crisp and persuasive messages
• effective use of language
Barriers
• underlying evidence that is contradictory or complex
• use of evidence that is still evolving or evidence not commonly observed in practice
• ambiguity and vagueness
• poor framing of the guideline
• use conditional statements
• use specific, concrete statements
• justify any deliberate vagueness
• use short sentences with proper punctuation
• keep related information together
• frame recommendations in terms of gains
• focus on errors of omission (not doing the
right thing) rather than errors of commission
(doing the wrong thing)
FORMAT
Facilitators:
• use of multiple formats or alternate versions
• inclusion of components known to be important to implementation
• appropriate placement of graphics and text
• appropriate structure (high-level categorization of recommendations)
• information visualization to shift cognitive load to perceptual system through
graphics and animation (e.g., matching system to the real world, bundling of related
recommendations)
• use of words for procedural information, logical conditions, and abstract concepts
• use of images for spatial structures, location, and detail
• tailor guidelines to intended end users
• highlight key evidence-based features
having the most significance for patient care
• highlight key recommendations with links to
more extensive explications
• choose the most appropriate graphic
according to the type of information to be
conveyed
• use colour appropriately
*Please see Web Appendix for more detailed list of facilitators, barriers and suggested actions for guideline developers with
references.
MA
NU
SC
RIP
T
AC
CE
PTE
D
ACCEPTED MANUSCRIPT
1
Figure 1: Flow of Realist Review Methods
Extraction of data 1736 guideline attributes were extracted in
duplicate from 278 included articles
Organization of data 2 sets of investigators (Group 1; Group 2)
independently clustered 1736 attributes into logical categories
Realist Review Methods
Auditing process
• Primary reviewers’ data extractions audited by second reviewer
• Disagreements resolved through consensus
Category agreement
• Compared “agreed” and “divergent” classifications
Data Extraction
Group 1 = 33 categories
Group 1 = 28 categories
Derived a common set of categories through group
discussion (N = 27) across 5 domains
LEVEL 2: Expert validation of Codebook and domain categorization
Collapsed into 16 categories across 6 domains
Validation with 9 experts
• Identify flaws in categorization
• Sense and fit of categories, attributes and their labels and definitions
Data Organization
LEVEL 1: Codebook of definitions Determine attributes of guideline uptake, their
operationalization, and relationship with uptake
Data Analysis
LEVEL 3: Hierarchical narratives Determined what works for whom, under what
circumstances and why
Search strategy An iterative, multiple search strategy that
consisted of 5 non-linear stages of searching
Article selection Two sets of reviewers (6 in total) independently
screened articles using inclusion criteria
Search strategy
Article selection
Stage 1: Core articles
Stage 2: Expert identified
Stage 3: PubMed Related articles
Stage 5: Other
Stage 4: Bibliography
Abstract level: N = 2044
Full text level: N = 350