Tools developed and disseminated by guideline producers to promote the uptake of their guidelines
Abstract
Background
The uptake of clinical practice guidelines (CPGs) is inconsistent, despite their potential to improve the quality of health care and patient outcomes. Some guideline producers have addressed this problem by developing tools to encourage faster adoption of new guidelines. This review focuses on the effectiveness of tools developed and disseminated by guideline producers to improve the uptake of their CPGs.
Objectives
To evaluate the effectiveness of implementation tools developed and disseminated by guideline producers, which accompany or follow the publication of a CPG, to promote uptake. A secondary objective is to determine which approaches to guideline implementation are most effective.
Search methods
We searched the Cochrane Effective Practice and Organisation of Care (EPOC) Group Specialised Register, Cochrane Central Register of Controlled Trials (CENTRAL); NHS Economic Evaluation Database, HTA Database; MEDLINE and MEDLINE In‐Process and other non‐indexed citations; Embase; PsycINFO; CINAHL; Dissertations and Theses, ProQuest; Index to Theses; Science Citation Index Expanded, ISI Web of Knowledge; Conference Proceedings Citation Index ‐ Science, ISI Web of Knowledge; Health Management Information Consortium (HMIC), and NHS Evidence up to February 2016. We also searched trials registers, reference lists of included studies and relevant websites.
Selection criteria
We included randomised controlled trials (RCTs) and cluster‐RCTs, controlled before‐and‐after studies (CBAs) and interrupted time series (ITS) studies evaluating the effects of guideline implementation tools developed by recognised guideline producers to improve the uptake of their own guidelines. The guideline could target any clinical area.
Data collection and analysis
Two review authors independently extracted data and assessed the risk of bias of each included study using the Cochrane 'Risk of bias' criteria. We graded our confidence in the evidence using the approach recommended by the GRADE working group. The clinical conditions targeted and the implementation tools used were too heterogenous to combine data for meta‐analysis. We report the median absolute risk difference (ARD) and interquartile range (IQR) for the main outcome of adherence to guidelines.
Main results
We included four cluster‐RCTs that were conducted in the Netherlands, France, the USA and Canada. These studies evaluated the effects of tools developed by national guideline producers to implement their CPGs. The implementation tools evaluated targeted healthcare professionals; none targeted healthcare organisations or patients.
One study used two short educational workshops tailored to barriers. In three studies the intervention consisted of the provision of paper‐based educational materials, order forms or reminders, or both. The clinical condition, type of healthcare professional, and behaviour targeted by the CPG varied across studies.
Two of the four included studies reported data on healthcare professionals' adherence to guidelines. A guideline tool developed by the producers of a guideline probably leads to increased adherence to the guidelines; median ARD (IQR) was 0.135 (0.115 and 0.159 for the two studies respectively) at an average four‐week follow‐up (moderate certainty evidence), which indicates a median 13.5% greater adherence to guidelines in the intervention group. Providing healthcare professionals with a tool to improve implementation of a guideline may lead to little or no difference in costs to the health service.
Authors' conclusions
Implementation tools developed by recognised guideline producers probably lead to improved healthcare professionals' adherence to guidelines in the management of non‐specific low back pain and ordering thyroid‐function tests. There are limited data on the relative costs of implementing these interventions.There are no studies evaluating the effectiveness of interventions targeting the organisation of care (e.g. benchmarking tools, costing templates, etc.), or for mass media interventions. We could not draw any conclusions about our second objective, the comparative effectiveness of implementation tools, due to the small number of studies, the heterogeneity between interventions, and the clinical conditions that were targeted.
PICOs
Plain language summary
Effectiveness of tools developed and disseminated by guideline producers to improve uptake of their guidelines
Background
Clinical practice guidelines (CPGs) are evidence‐based recommendations for healthcare professionals about the care of patients with specific conditions. The uptake of CPGs by healthcare professionals is inconsistent, despite their potential to improve the quality of health care and patient outcomes. Some guideline producers have addressed this problem by developing tools to encourage the adoption of new guidelines.This review focuses on the effectiveness of tools developed and distributed by recognised guideline producers to improve the uptake of their CPGs.
Characteristics of included studies
Researchers from Cochrane searched the literature up to February 2016 and identified four randomised studies evaluating the effects of tools developed by recognised guideline producers to implement their guidelines. These were developed by guideline producers in France, the Netherlands and in the USA and Canada. In all four studies the interventions targeted the healthcare professional. None of the tools specifically targeted the organisation of care or the patient. The clinical conditions, and the healthcare professionals' behaviour targeted by the CPG, varied across studies, as did the tools used to improve guideline implementation.
Key results
Two of the four included studies reported on how well healthcare professionals stick to guideline recommendations when providing care to their patients, depending on whether they received a CPG with a tool aimed at improving the use of the CPG, or if they received the CPG only. The results of this review show that healthcare professionals who received a guideline tool together with the CPG on the management of non‐specific low back pain or ordering thyroid‐function tests probably stick more closely to the recommendations, compared with those who received the CPG only. A guideline tool aimed at improving the use of a guideline, may lead to little or no difference in cost to the health service.
Certainty of the included evidence
The included evidence was from randomised controlled trials, which is considered the highest level of evidence. However, due to high risk of bias in the included studies our confidence in the effect on observing guideline recommendations was moderate. Our confidence in the evidence for cost effectiveness was low, since only a single study provided evidence for this comparison.
Authors' conclusions
Summary of findings
| Clinical practice guideline (CPG) + implementation tool compared with CPG only for adherence to guidelines | ||||
| Patient or population: Healthcare professionals (physiotherapists, hospital physicians) providing care for people with different clinical conditions (patients with non specific low back pain, patients with symptoms indicating a need for a thyroid function test) Setting: Private physiotherapy clinics in the Netherlands, general hospitals in France Intervention: CPG + guideline implementation tool (e.g. training workshops, paper‐based materials and order forms, reminders, web‐based tools) | ||||
| Outcomes | Median ARD (Absolute risk difference) | No of Participants | Certainty of the evidence | Comments |
| Adherence to guidelines | Guideline tools provided to healthcare professionals as an aid to improve the use of a CPG probably lead to improved adherence with the CPG, as compared to guidelines only. Median ARD: 0.135 (0.115 to 0.15.9) at mean 4 weeks follow‐up | 68 physio‐ therapy practices; and 6 hospitals (2 C‐RCTs) | ⊕⊕⊕⊝ | 2 of the 4 included studies reported a proxy measure for adherence to guidelines, and results from these studies could therefore not be included in the ARD calculation |
| Costs | Guideline tools aimed at improving the use of guidelines may lead to little or no difference in healthcare costs | 68 physio‐ therapy clinics (1 C‐RCT) | ⊕⊕⊝⊝ low2 | 1 trial reported no difference in mean direct annual cost* per patient between intervention and control groups. 1 French paper belonging to 1 of the included trials (6 hospitals) and reporting on costs awaits translation |
| * Direct costs included costs of the dissemination of the guideline and healthcare resource use by the patient GRADE Working Group grades of evidence | ||||
| 1We downgraded the certainty of the evidence one step due to high risk of bias. 2We downgraded the certainty of the evidence two steps due to imprecision. | ||||
Background
The uptake of clinical practice guidelines (CPGs) is inconsistent, despite their potential to improve the quality of health care and patient outcomes (Grol 2003; Schuster 1998; Seddon 2001). A variety of tools to improve the uptake of CPGs have been developed, but not always by the guideline producers themselves. This review focuses on the effectiveness of tools developed and distributed by guideline producers to improve the uptake of their CPGs. These tools are either embedded within a guideline, for example tailoring a guideline recommendation for a particular user group, or they accompany the CPG, for example interactive learning modules developed to support the use of a CPG.
Description of the condition
CPGs have the potential to improve healthcare delivery and outcomes, but the adoption of guidelines by healthcare professionals and health system managers is inconsistent, and gaps remain between recommended care and clinical practice. Previous systematic reviews have identified a range of interventions to support the implementation of guidelines (Grimshaw 2004). However, most of these have been developed independently of the producers of guidelines. In response to this some guideline producers have developed tools to improve the uptake of their CPGs. Some of these interventions focus on improving the skills needed to apply evidence to practice and others aim to integrate the content of a CPG into a local healthcare system. The value of these tools has been questioned by the UK National Institute for Health and Care Excellence (NICE), (New Reference, Leng 2013 [pers comm]) as they are an additional investment for the guideline producer, and the evidence of the effectiveness of this approach is uncertain.
Description of the intervention
Interventions developed and disseminated by guideline producers to improve the consistent use of CPGs by health professionals and health system managers include learning modules (which may be accredited with Continuing Medical Education (CME) points), education outreach visits (for example, academic detailing), communication tools (for example, press releases following the publication of CPGs) or tailored formatting (for example, the wording of recommendations adapted for a target audience or local health system). Learning modules are a popular approach to supporting the use of CPGs; for example, NICE has developed a range of online educational tools (NICE 2012b) in collaboration with BMJ Learning, the Nursing Times and e‐Learning for Health (for example the eVTE online educational tool to reduce the risk of venous thromboembolism (eVTE 2013)). The goal is to enable users of CPGs to be more aware of recent evidence as summarised in the relevant NICE guidance and to apply the newly acquired knowledge in their practice and address any potential barriers. Examples of CPG producers working within health systems to improve the uptake of their CPGs include: NICE working within the National Health Service (NHS) in England and Wales by providing commissioners with quality standards (NICE 2016); the Scottish Intercollegiate Guidelines Network (SIGN) which provides problem‐based small‐group learning modules (SIGN 2012); the American College of Cardiology providing a guideline clinical app and running the Guidelines in Practice (GAP) project to provide customised, guideline‐specific implementation tools (ACC 2016; Mehta 2002); the Veterans Health Administration adapting their CPGs for colorectal screening to local health organisations; Kaiser Permanente's healthcare system which has developed and implemented their Proactive Officer Encounter Programme to provide clinical decision support to increase the uptake of their own and other CPGs (Kanter 2010); and the National Prescribing Centre in the UK that set up the 'communities of practice' (the NHS Medicines and prescribing communities of practice). Data from NHS Evidence show that 92% (33/36) of guideline producers submitting their CPGs for accreditation by NHS Evidence publish support tools intended to aid implementation of their guidance (NHS 2012a). Many guideline producers are working on transforming their narrative CPGs into electronic format, as this may improve uptake through the implementation of CPGs in computer‐based decision‐support systems (Peleg 2010).
How the intervention might work
Producers of CPGs who develop implementation interventions to support their use have focused on the information needs of different user groups. Interventions are aimed at improving awareness of CPGs, strengthening the skills needed to understand and implement a CPG, and supporting the use of a CPG within the context of a local healthcare system (Greenhalgh 2005). Tailoring the implementation of interventions to facilitate practice change (to promote a CPG) typically involves identification of the determinants of healthcare practice. This can include discussions with healthcare professionals about potential barriers and systems requiring change (Baker 2015), identification of ways to facilitate change and designing, applying and assessing appropriate interventions (Wensing 2011). The Guideline Implementability Appraisal (GLIA) instrument may be used by producers of guidelines to identify barriers to implementation during the design phase of a CPG and enable modifications prior to publication (Shiffman 2005). For example, templates may be developed for users of CPGs to populate with local data in order to assess the applicability and impact of a CPG. The tailoring process is also important in engaging clinicians in the implementation process (Horbar 2004; Titler 2009). Findings from a recent survey of more than 300 NHS commissioning staff, who use CPGs to guide decision making, confirm the importance of these approaches. Local public health intelligence, expert advice and examples of best practice appear to be the most sought‐after types of evidence, and in order for knowledge to be used it has to be translated into a practical resource (Gkeredakis 2011). Finally, if a guideline producer has authority and works within the health system, or is perceived to be influential, the uptake of CPGs may be improved (Rogers 1995).
Other determinants of the effective implementation of all CPGs are that they are clearly written, specific to a population and context, easy to use and that there is research evidence of its effectiveness for a particular end‐user's work context (Titler 2001). Guideline development is usually carried out by a multidisciplinary, nationally‐representative group, who conduct a systematic review to identify and critically appraise the evidence, and who ensure that the guideline recommendations are explicitly linked to the supporting evidence. Expert opinions are also used in CPGs where research evidence is not available. Producers of guidelines can also use the AGREE tool by which the quality of a guideline may be evaluated, thus allowing end‐users to decide how well a guideline has been developed and whether it will be applicable to the setting in which they are working (AGREE 2010).
The format used to communicate the content of a CPG can also influence its adoption (Greenhalgh 2005; Rogers 1995). While CPGs are frequently written as text documents (Peleg 2010), studies have shown that clinicians usually do not use written guidelines during the actual care process (Wang 2002). Instead, patient‐specific advice, particularly if delivered during patient encounters, is suggested to be more effective in changing clinician behaviour (Shea 1996). Thus, implementing CPGs in computer‐based decision‐support systems may improve the acceptance and application of guidelines in daily practice, particularly if the actions and observations of healthcare workers are monitored and advice is generated whenever a guideline is not followed (Wang 2002). One example of guideline producers who have provided healthcare professionals with clinical decision support to increase the uptake of CPGs is the Kaiser Permanente healthcare system with their Proactive Officer Encounter Programme (Kanter 2010).
Gagliardi 2011 identified eight features of CPGs that are desired by users of CPGs, or are associated with their use:
-
Usability: the structure of the CPG has been modified to facilitate access, for example by providing a one‐page summary of the recommendations;
-
Adaptability: the CPG is available in different formats for different users or purposes, for example, print and electronic format, versions of the CPG are available for patients and caregivers;
-
Validity: using a standardised system to grade the quality of evidence supporting each recommendation, for example GRADE;
-
Applicability: the wording of the CPG recommendation has been tailored for different target audiences to support application of the guidance to local circumstances; this may include clinical and contextual information;
-
Communicability: information to supplement the CPG, for example, educational resources for patients and information to support patient involvement;
-
Accommodation: the addition of information on costs and resources, for example, the costing templates provided by NICE, and information on competencies and training required to implement the recommendations;
-
Implementation: information on potential barriers and strategies for facilitating implementation, for example, a clinical assessment using a point‐of‐care template;
-
Evaluation: performance measures or quality indicators for audit and monitoring.
Why it is important to do this review
CPGs can improve healthcare delivery and outcomes, but the adoption of guidelines by clinicians and healthcare managers is inconsistent. Previous Cochrane Reviews have described the effectiveness of a range of interventions to support the implementation of guidelines (Akl 2013; Flodgren 2011; Flodgren 2013a; Forsetlund 2009; Giguère 2012; Grilli 2002; Jamtvedt 2006; O'Brien 2007; Shojania 2009). However, most of these have been developed independently of the producers of guidelines. Responding to continued concern about the inconsistent use of CPGs, some national guideline producers have developed and implemented tools to support the uptake of their CPGs. This is an additional investment for the guideline producer and the effectiveness of this approach is not known. The focus of this review is to assess the effectiveness of implementation tools, developed and disseminated by guideline producers, on the uptake of their guidelines. These tools may require changes to the presentation of the CPG (e.g. tailoring of a CPG), or to be published alongside CPGs (e.g. online learning modules).
Objectives
To evaluate the effectiveness of implementation tools developed and disseminated by guideline producers, which accompany or follow the publication of a CPG, to promote uptake.
A secondary objective is to determine which approaches to guideline implementation are most effective.
Methods
Criteria for considering studies for this review
Types of studies
We aimed to include randomised controlled trials (RCTs), cluster‐randomised trials (C‐RCTs), controlled before‐and‐after studies (CBAs) and interrupted time series (ITS) studies evaluating the effects of guideline implementation tools developed by recognised guideline producers to improve the uptake of their own guidelines. CBAs were eligible for inclusion if they involved at least two intervention and two control sites, and ITS studies were eligible if they had at least three data points before and three data points after the intervention.
Types of participants
We included all qualified healthcare professionals, health system managers and policy makers.
We excluded studies involving trainees or medical students.
Types of interventions
We included any interventions developed by producers of CPGs to improve guideline implementation. Guideline producers include, for example, the World Health Organisation (WHO), NICE, and SIGN. As guidelines may be produced for a specific jurisdiction, health system, or group of healthcare professionals, interventions to improve the implementation of these CPGs may be distributed to organisations but targeted at individuals within the organisation, or they may be targeted at entire organisations. We used the definition of a CPG developed by the USA Institute of Medicine: "clinical guidelines are systematically developed statements to assist health care professional and patient decisions about appropriate health care for specific clinical circumstances" (Field 1990).
Using the EPOC taxonomy (EPOC taxonomy 2002) as a guide, we developed the following classification to organise and define interventions as below:
1. Tools targeting the healthcare professional
i) Tailoring
-
Tailoring of CPGs for different users to improve usability and applicability: examples include using different wording, varying the content, incorporating case studies of patients' experiences in the form of vignettes or narratives which contextualise the recommendations.
-
Different CPG formats adapted for different users/purposes, e.g. electronic (for use on a Personal Digital Assistant), paper, multimedia versions, summaries, the inclusion of algorithms.
ii) Education
-
Learning modules (to include interactive learning modules) which may be accredited with Continuing Medical Education (CME) points, or to support the use of audit by junior doctors.
-
Instructions/templates, e.g. instructions, tools or templates to tailor guidelines/recommendations for local context (may also be used at the organisational level); point‐of‐care templates/forms (clinical assessment, standard orders).
-
Decision‐support systems, e.g. electronic guidelines with built‐in decision‐support systems.
2. Tools targeting the patient
-
Producing versions of CPG recommendations for the public to improve provider‐patient communication about guideline recommendations.
3. Tools targeting the organisation of care
-
Benchmarking tools, e.g. measures of gaps in performance to be used by those monitoring the implementation of CPGs (may also be used by individual healthcare professionals).
-
Costing templates as a budgetary aid (may also be used by individual healthcare professionals) to assess the resources required to implement the CPG.
-
Programme evaluation, audit tools, performance measures and quality indicators to evaluate the implementation of the CPG.
4. Mass media interventions
-
Press releases following the publication of a CPG.
The comparisons are as follows:
-
Tools developed by a guideline producer versus a tool developed by another organisation or a guideline user (i.e. tools developed independently of the CPG producer).
-
Tools developed by a guideline producer versus no tool (i.e. CPGs alone).
We excluded the following types of studies/interventions:
-
Tools developed by groups of researchers, guideline groups on commission (no longer existing).
-
Studies describing tools developed by guideline producers to improve guideline uptake without providing objective measurements of the effect of these interventions on professional practice or patient outcomes.
-
Surveys of barriers/facilitators to the uptake of guidelines.
Types of outcome measures
We included studies reporting the following outcome measures:
Main outcomes
Objective measures of healthcare professional performance, healthcare resource use or patient outcomes.
Secondary outcomes
Self‐reported measures of healthcare professional performance and healthcare manager performance, including knowledge or use of CPGs, and costs.
We excluded studies that only included self‐reported outcomes.
Search methods for identification of studies
Electronic searches
Information specialist Nia Roberts (NR) developed the search strategy for MEDLINE in consultation with the review authors, and searched the Cochrane Database of Systematic Reviews and the Database of Abstracts of Reviews of Effects (DARE) up to February 2016 for related systematic reviews, and the following databases for primary studies.
-
Cochrane Effective Practice and Organisation of Care (EPOC) Group Specialised Register, Reference Manager
-
Cochrane Central Register of Controlled Trials (CENTRAL) (Cochrane Library), Wiley (3 February 2016)
-
NHS Economic Evaluation Database, HTA Database (Cochrane Library), Wiley (3 February 2016)
-
MEDLINE and MEDLINE In‐Process and other non‐indexed citations, OvidSP (1946 to 3 February 2016)
-
Embase, OvidSP (1974 to 3 February 2016)
-
PsycINFO, OVIDSP (1967 to 3 February 2016)
-
CINAHL (Cumulative Index to Nursing and Allied Health Literature), EbscoHost (1982 to 3 February 2016)
-
Dissertations and Theses, ProQuest (3 February 2016)
-
Index to Theses (up to 3 February 2016)
-
Science Citation Index Expanded, ISI Web of Knowledge (1945 to 3 February 2016)
-
Conference Proceedings Citation Index ‐ Science, ISI Web of Knowledge (1990 to 3 February 2016)
-
Health Management Information Consortium (HMIC), NHS Evidence (1979 to 3 February 2016)
The MEDLINE search strategy (Appendix 1) was translated for other databases using appropriate syntax and vocabulary for those databases. The strategy included Medical Subject Headings (MeSH) and synonyms for guidelines and implementation. Results were limited by two methodological filters: the Cochrane Highly Sensitive Search Strategy (sensitivity‐ and precision‐maximising version, 2008 revision) to identify randomised trials, and an EPOC methodology filter to identify non‐RCT designs. We did not apply language or publication status restrictions. Search strategies for the other databases are found in Appendix 2; Appendix 3; Appendix 4; Appendix 5; Appendix 6; Appendix 7; Appendix 8; Appendix 9.
Searching other resources
We searched the following additional sources:
Grey literature
We conducted a 'grey literature' search to identify studies not indexed in the databases listed above. Sources included the sites listed in Appendix 10. We document guideline websites searched in Appendix 11.
Trial registries
We searched the following registries:
-
International Clinical Trials Registry Platform (ICTRP), Word Health Organization (WHO) (www.who.int/ictrp/en/)
-
ClinicalTrials.gov, US National Institutes of Health (NIH) (clinicaltrials.gov/)
We also :
-
reviewed reference lists of all included studies, relevant systematic reviews/primary studies/other publications;
-
contacted authors of relevant studies or reviews to clarify reported published information/seek unpublished results/data;
-
contacted researchers with expertise relevant to the review topic/EPOC interventions, as well as guideline‐producing bodies regarding any further published or unpublished research;
-
conducted cited reference searches for all included studies in ISI Web of Knowledge.
Data collection and analysis
Selection of studies
We downloaded all titles and abstracts retrieved by electronic searching to the reference management database Endnote (EndNote X7) and removed duplicates. Four review authors (from GF, AH, LG, SS) and an additional systematic reviewer independently examined the remaining references, excluded those studies which clearly did not meet the inclusion criteria, and produced a long list of possible included studies. Two review authors (from GF, AH and SS) scrutinised these citations, obtained full‐text copies of potentially relevant references, and independently assessed the eligibility of the retrieved full‐text papers. We resolved disagreements by discussion among review authors.
Data extraction and management
Two review authors (from GF, AH and LG) independently extracted data from each study into a modified EPOC data extraction form. We resolved disagreements by discussion among review authors. We extracted the following information: setting; location; characteristics of healthcare professionals; type of healthcare organisation; intended population of guideline; type and target of intervention; the components of the intervention; the comparison intervention; any information about the time (and resources) needed to implement or use the tool, or both; costs and outcomes reported.
We also extracted data on any collaborative effort between producers and users of guidelines aiming to improve the development of implementation tools, e.g. engagement of individual healthcare professionals or the organisation of care or both in the development; assessment of barriers/facilitators to CPG adoption at the provider level or at the organisational level, or both; or assessment of the healthcare professionals' or the organisation of care's readiness to change.
We used the Review Manager 5 software developed by Cochrane (Review Manager 2014) to structure the content of the review when writing it up for publication.
Assessment of risk of bias in included studies
Two review authors (from GF, AH, and LG) independently assessed the risk of bias of each included study using the Cochrane's 'Risk of bias' tool (Higgins 2011) on six standard criteria:
-
Adequate sequence generation;
-
Adequate concealment of allocation;
-
Blinded or objective assessment of main outcome(s);
-
Adequately addressed incomplete outcome data;
-
Free from selective outcome reporting;
-
Free from other potential sources of bias.
We used four additional criteria specified by EPOC (EPOC 2015):
-
Similar baseline characteristics;
-
Similar baseline outcome measures;
-
Reliable main outcome measures;
-
Adequate protection against contamination.
We resolved disagreements by discussion among review authors.
We assigned an overall assessment of the risk of bias (high, moderate or low risk of bias) to each of the included studies using the approach suggested in Chapter 8 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). We considered studies with low risk of bias for all key domains, or where it seems unlikely that bias seriously alter the results, to have a low risk of bias. We rated studies as high risk of bias if at least one domain was unclear or studies were judged to have some bias that could plausibly raise doubts about the conclusions as being at moderate risk of bias. We considered studies with a high risk of bias in at least one domain or judged to have serious bias that decreases the certainty of the conclusions, to have a high risk of bias.
Measures of treatment effect
For each study, we reported data in natural units. Where baseline results were available from RCTs we reported pre‐intervention and postintervention means or proportions for both study and control groups. We also calculated the absolute risk difference (ARD) for each reported dichotomous outcome, using baseline data when available.
Unit of analysis issues
There were no unit of analysis issues, all studies adjusted for clustering (Bekkering 2005; Daucourt 2003; Fine 2003; Shah 2014).
Dealing with missing data
We did not contact authors to request missing data, for example, when the main outcome was graphically presented without numerical data.
Assessment of heterogeneity
Due to the heterogeneity between studies in terms of populations, clinical conditions/targeted behaviour, and implementation tools used, meta‐analysis was not feasible, and we therefore did not assess statistical heterogeneity.
Assessment of reporting biases
As meta‐analysis of main outcomes was not feasible, we did not assess publication bias through a funnel plot. However, our search for studies to include was extensive and covered a number of guideline web sites, Guideline Clearing Houses and professional associations.
Data synthesis
As we did not find sufficiently homogeneous studies to permit meta‐analysis, we reported, for dichotomous outcomes, the median of medians of effect sizes across studies, as has previously been done in other reviews (Flodgren 2011: Grimshaw 2004; Shojania 2009). When multiple adherence outcomes were reported within individual studies, we first calculated the median effect size (range) across all outcomes reported in each study, and then calculated the median of medians and interquartile range (IQR) across studies.
Two review authors used the GRADE tool (www.gradeworkinggroup.org/) to judge the overall certainty of the evidence for each outcome, using the following domains: risk of bias, inconsistency, imprecision, indirectness and publication bias. We downgraded the evidence for serious concerns about each of these domains. We resolved disagreements through discussions among the review authors. We presented the grading of the evidence in summary of findings Table for the main comparison.
Subgroup analysis and investigation of heterogeneity
As all of the included studies used implementation tools that targeted the healthcare professional, and only one study targeted both the healthcare professional and the patient, we did not undertake any subgroup analyses.
Sensitivity analysis
We did not perform a sensitivity analysis, as no meta‐analysis was conducted.
Results
Description of studies
Results of the search
See Characteristics of included studies; Characteristics of ongoing studies and Characteristics of studies awaiting classification tables.
The electronic searches yielded 47,181 citations, down to 26,384 after removal of duplicates. Additional sources searched (including websites and reference lists) yielded 473 citations. Of the these 26,857 citations, we excluded 25,801 irrelevant studies and retrieved and scrutinised 1,056 studies. Of these 1,056 double‐screened studies we excluded 1030 studies and added 14 to the excluded studies table. We listed one study protocol and one conference abstract under 'Ongoing studies' (Salbach 2014; Te Boveldt 2011), and two studies under 'Studies awaiting classification' (Maximov 2012; Van Driel 2007). We judged four studies to be eligible for inclusion in the review. See study flowchart Figure 1.
Study flow diagram.
Included studies
We identified four eligible studies of cluster‐RCTs (Bekkering 2005; Daucourt 2003; Fine 2003; Shah 2014) for inclusion in this review, of which one (Shah 2014) consisted of two separate cluster‐RCTs: one a population‐based C‐RCT including all family practices in Ontario, Canada, and the other an embedded C‐RCT including a subsample of the family practices from the larger study.
Populations
Healthcare professionals
In Bekkering 2005 the participants were physiotherapists (n = 113); and in two studies (Daucourt 2003; Fine 2003) the participants were physicians other than general practitioners (GPs) (n = 1913), or family physicians (number not reported), and in one study the intervention was targeted at family physicians (Shah 2014). None of the studies targeted patients, health system managers or policy makers.
Patients
The number and clinical condition of participants in the included studies were as follows: participants (n = 500) with non‐specific low back pain (Bekkering 2005); participants (n = 608) with hospital‐acquired pneumonia (Fine 2003); an unknown number of patients who required a thyroid‐function test (Daucourt 2003); and people with diabetes > 40 years old (n = 933,769) in Ontario (administrative study) and a subgroup of people with diabetes (n = 1592) at high risk of cardiovascular disease (clinical study) (Shah 2014).
Settings
Bekkering 2005 was set in private physiotherapy practices (n = 68); two studies (Daucourt 2003; Fine 2003) were set in hospitals (n = 13), and Shah 2014 was set in family practices (n = 4007 and n = 80 respectively). The studies were conducted in the Netherlands (Bekkering 2005), France (Daucourt 2003), the USA (Fine 2003) and in Canada (Shah 2014) .
Targeted behaviour
The clinical conditions/behaviours targeted by the CPG were as follows: care for people with non‐specific low back pain (Bekkering 2005); appropriate thyroid‐test ordering (Daucourt 2003); timely conversion (and discharge) from intravenous antibiotic therapy to oral antibiotics for people with pneumonia (Fine 2003); and improved cardiovascular risk screening and risk reduction in people with diabetes (Shah 2014).
The guideline recommendations (n = 4) that were implemented were described in one of the studies (Bekkering 2005).
Guideline producers
See Table 1 for details on the guideline development process.
| Author Year Targeted behaviour | Guideline developers | Literature review | Critical appraisal | Consensus processes | Key stakeholder involvement | Barriers/facilitator assessment |
| Targeted behaviour: management of non‐specific low back pain Number of recommendations: 4 main recommendations | The Royal Dutch Physiotherapy Association. | CPGs2 were constructed on the basis of the phases of the physiotherapy process, using the Dutch method of developing physiotherapy guidelines, and evidence from systematic reviews were identified through searching electronic databases | Not mentioned but probably included in the Dutch method of developing CPGs | Based on scientific evidence. If no evidence was available, consensus between experts was obtained | The CPGs were pilot‐tested among 100 physiotherapists and reviewed by an external multidisciplinary panel | Barriers to change were assessed through a survey as part of the CPG development process |
| Targeted behaviour: appropriate thyroid function testing Number of recommendations: not reported | The Committee for Co‐ordinating Clinical Evaluation and Quality in | The CPG developers conducted a comprehensive review of the literature | ‐ | CPG3 development involved an expert consensus process. | ‐ | ‐ |
| Targeted behaviour: appropriate duration of intravenous antibiotic therapy for treatment of pneumonia Number of recommendations: a 2‐step recommendation | Researchers who were part of the Pneumonia Patient Outcomes Research Team (PORT) | The CPG was based on a review of the medical literature, and empiric evidence on time to reach clinical stability | ‐ | The CPG development process involved the consensus of an 8‐member national guideline panel | The guideline was reviewed by clinical opinion leaders at each study site, and was approved for local use by the relevant utilisation management department. | ‐ |
| Targeted behaviour: management of cardiovascular risk Number of recommendations: no information | Canadian Diabetes Association (and Expert Committee members) | Expert Committee members evaluated the relevant literature, and guidelines were developed and initially reviewed | After formulating new recommendations or modifying existing | Based on scientific evidence/review of the literature | A draft document was circulated nationally and | ‐ |
1Khunti 1998. Development of evidence‐based review criteria for the management of patients with depression in general practice. No published version of the guideline found.
2Bekkering 2003. Dutch physiotherapy guidelines for low back pain.
3Saillour Glénisson 2001. Guidelines for thyroid function tests in adults.
4Shah 2014. Canadian Diabetes Association Clinical Practice Guidelines Expert Committee: Canadian Diabetes Association 2008 clinical practice guidelines for the prevention and management of diabetes in Canada.
In Bekkering 2005 the Royal Dutch Physiotherapy Association developed the guidelines; in Daucourt 2003 the Committee for Co‐ordinating Clinical Evaluation and Quality in Aquitaine (CCECQA) developed the guidelines, together with regional groups and national guideline developers; in Fine 2003 members of the Pneumonia Patient Outcomes Research Team (PORT) project developed the guidelines; and in Shah 2014 the Canadian Diabetes Association (CDA) developed the guidelines.
Description of the intervention
| Author Year | Delivery of the intervention | Theoretical models/ frameworks used | Evidence base | Targeted to barriers | Key stakeholder involvement |
| Type of intervention: simple, but multicomponent; active only Intervention target: the healthcare professional Targeted behaviour: management of nonspecific low back pain | Mode: face‐to‐face Provider: The primary investigator and 1 of 2 additional trainers with adequate clinical experience in the management of low back pain supervised the training sessions | ‐ | "The sessions were based on interventions that have all been shown to be effective, such as interactive education and discussion, feedback, and reminders".1,2,3,4,5 | The content of the strategy was determined on the basis of information about the expected barriers for implementation gathered during the development of the CPGs | 2 experts gave advice on the content of the strategy |
| Type of intervention: multicomponent; passive only Intervention target: the healthcare professional Targeted behaviour: appropriate thyroid function testing | Mode: Paper‐based materials Provider: none | ‐ | "Among the clinical guideline diffusion strategies, the most effective are feedback, reminders, academic detailing and financial incentives1,2,4,5 Administrative procedures such as the implementation of test request forms have also proved effective." | ‐ | ‐ |
| Type of intervention: single; active + passive Intervention target: the healthcare professional Targeted behaviour: appropriate duration of intravenous antibiotic therapy for treatment of pneumonia | Mode: Paper‐based material (detail sheets/treatment recommendations in patient records) and telephone reminder Provider: nurse delivered telephone reminder | ‐ | "The multifaceted guideline dissemination strategy consisted of interventions of proven benefit, including real‐time physician reminders" 1,6,7,8 | ‐ | ‐ |
| Type of intervention: passive Intervention target: family physicians (and diabetes patients at high risk of cardiovascular disease) Targeted behaviour: management of cardiovascular risk factors and outcomes of cardiovascular disease in people with diabetes | Mode: printed educational materials Targeting the family physician: The cardiovascular disease toolkit was a collection of printed educational materials, packaged in a brightly‐coloured box with CDA branding, sent to Canadian family physicians. The contents included an introductory letter from the Chair of the practice Provider: NA13 | The implicit theory behind its | "The literature has demonstrated that the benefits of printed educational interventions are, at best, modest. A systematic review of methods to improve practice guideline adherence demonstrated an absolute improvement of 8% for educational materials. A more recent Cochrane review found that printed educational materials led to a median absolute improvement in performance of only 2% (25). Studies of printed materials specifically tied to clinical practice guidelines also showed modest benefits. A small English study randomised 42 family physicians to receive an algorithm for monitoring and treatment of hypertension of diabetic patients based on practice guidelines, but found no difference in blood pressure control between the intervention and control groups. However, some processes of care were slightly improved: patients in the intervention group were prescribed higher doses of antihypertensive medications, and had more physician visits to monitor blood pressure. In a larger Canadian study, family physicians were randomised to receive a 1‐page summary of a 3‐year‐old practice guideline on anti‐anginal therapy from the local medical governing body. No differences were noted in prescription of β‐blockers, antiplatelet agents, or lipid‐lowering drugs between groups in the 7000 patients reviewed" 9,10,11,12 | ‐ | The toolkit was |
1Bero 1998 Closing the gap between research and practice: an overview of systematic reviews of interventions to promote the implementation of research findings.
2Davis 1995 Changing physician performance. A systematic review of the effect of continuing medical education strategies.
3Wensing 1998 Implementing guidelines and innovations in general practice: which interventions are effective?
4Grimshaw 1995 Developing and implementing clinical practice guidelines.
5Davis 1997. Translating guidelines into practice. A systematic review of theoretic concepts, practical experience and research evidence in the adoption of clinical practice guidelines.
6Murrey 1992 Implementing clinical guidelines: a quality management approach to reminder systems.
7Grimshaw 1993. Effect of clinical guidelines on medical practice: a systematic review of rigorous evaluations.
8Weingarten 2000. Translating practice guidelines into patient care: guidelines at the bedside.
9Grimshaw 2006. Towards evidence‐based quality improvement: evidence (and its limitation) of the effectiveness of guideline dissemination and implementation strategies 1966–1998.
10Giguère 2012. Printed educational materials: effects on professional practice and healthcare outcomes.
11Bebb 2007. A cluster randomised controlled trial of the effect of a treatment algorithm for hypertension in patients with type 2 diabetes.
12Beaulieu 2004. Drug treatment of stable angina pectoris and mass dissemination of therapeutic guidelines: a randomized controlled trial.
13Not applicable
| Author Year | Tailoring | Feedback | Educational outreach/ Academic detailing/ Small group discussions | Reminders (paper, electronic, telephone) | Decision support tools | Other (test order forms, supportive materials etc.) |
| The content of the strategy was determined on the basis of information about the expected barriers for implementation gathered during the development of the clinical guidelines. | ‐ | 2 interactive training sessions, each lasting 2½ hours, for groups of 8 – 12 physiotherapists (including feedback on current management and reminders). For each session a preparation time of 2 hours was recommended. | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | Pocket memorandum card. | ‐ | Test request form. | |
| ‐ | ‐ | ‐ | Paper‐based detail sheet/ treatment recommendations put into the patient's record + real‐time nurse telephone reminder | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | ‐ | Printed educational materials |
i) Interventions targeting the healthcare professional
All four studies evaluated guideline implementation tools targeting the healthcare professional.
Tailored interventions
Bekkering 2005 assessed the effectiveness of two (2½ hours) educational training sessions for groups of 8 to 12 physiotherapists on adherence to CPGs for management of non‐specific low back pain. The sessions were based on interventions reported as being effective in the literature (e.g. interactive education and discussion, feedback, and reminders) and were tailored to barriers found in a survey.
Printed materials
Three studies evaluated the effectiveness of paper‐based educational materials or reminders, or both (Daucourt 2003; Fine 2003; Shah 2014).
Daucourt 2003 evaluated the combined effect of two tools: a memorandum pocket card and a test request form to implement guidelines for appropriate thyroid‐test ordering. Orders were made by checking a box, with boxes corresponding to inappropriate test ordering shaded and therefore making ordering impossible.The physician could overrule this by writing the order at the bottom of the sheet. The pocket card summarised the recommendations according to the various clinical or therapeutic situations requiring a thyroid test.
In Fine 2003 physicians received a multifaceted guideline intervention which included placement of a detail sheet in the patient’s medical record once a patient met guideline criteria for stability when receiving intravenous antibiotic therapy for pneumonia, a follow‐up recommendation to the attending physician, and an offer to arrange follow‐up home nursing care. The three site‐specific detail sheets promoted any of three recommended action(s), i.e. conversion from intravenous to oral antibiotic therapy only, conversion and hospital discharge, or hospital discharge only.
Shah 2014 used a cardiovascular disease toolkit which was a collection of printed educational materials, packaged in a brightly‐coloured box with CDA branding, sent to Canadian family physicians. The contents included an introductory letter from the Chair of the practice guidelines’ Dissemination and Implementation Committee; an eight‐page summary of selected sections of the practice guidelines targeted towards family physicians; a four‐page synopsis of the key guideline elements pertaining to cardiovascular disease risk; a small double‐sided laminated card with a simplified algorithm for cardiovascular risk assessment, vascular protection strategies, and screening for cardiovascular disease; and a pad of tear‐off sheets for patients with a cardiovascular risk self‐assessment tool and a list of recommended risk reduction strategies.
The median duration that an intervention was delivered was 22 weeks (range 4 weeks to 12 months).
ii) Interventions targeting the patient
None of the included studies evaluated interventions that targeted the patient.
iii) Interventions targeting the organisation of care
None of the included studies evaluated interventions that exclusively targeted the organisation of care.
iv) Interventions targeting the healthcare professionals and the patients
None of the included studies evaluated targeted both healthcare professionals and patients.
Assessment of barriers
In one of the three included studies (Bekkering 2005), barriers to guideline implementation were assessed through the means of a survey to inform the shape and content (i.e. tailoring) of the guideline implementation strategies. Another aim of the survey was to retrieve information on the most important discrepancies between current practice and recommendations of the guidelines. A model for changing professionals' behaviour and systematic reviews on the effectiveness of implementation interventions was also used to determine the content of the implementation strategy.
Theory base of interventions
None of the interventions used in the included studies was theory‐based.
Evidence base of interventions
The implementation strategies used in the included studies were all supported by some evidence of their effectiveness and cited high‐quality Cochrane Reviews, systematic reviews or overviews to justify their choice of strategies.
Fidelity
None of the included studies provided information on intervention fidelity.
Delivery of the intervention
Mode of delivery:
In Bekkering 2005 the intervention was delivered face‐to‐face. In two studies (Daucourt 2003; Shah 2014) the paper‐based interventions were provided passively. In Fine 2003 one part of the intervention was delivered over the phone, and the rest passively in the form of paper‐based materials.
Provider delivering the intervention (if not electronic, paper‐based, etc):
In Bekkering 2005 the principal investigator and two additional trainers delivered the intervention. In Fine 2003 a nurse delivered part of the intervention.
Comparison interventions
The comparison intervention in all included studies was passive guideline dissemination. Additional material that was delivered together with the guideline was as follows: in Bekkering 2005 four forms: a self‐evaluation form to assess whether their current management was consistent with the recommendations contained in the clinical guidelines, two forms facilitating discussion with other physiotherapists and general practitioners respectively, a copy of the Quebec Back Pain Disability Scale, and a summary of the CPG. In Fine 2003 a cover letter was sent signed by the hospital’s utilisation management director describing the rationale for the guideline. In Daucourt 2003 all physicians were invited to a local information meeting. In Shah 2014 control participants received the CDA newsletter, which included the revised guideline.
Outcomes
Healthcare professional outcomes
Two of the four included studies reported a measure of healthcare professional adherence to guidelines (Bekkering 2005; Daucourt 2003) at four weeks; these were included in the calculations of the median absolute risk difference (ARD).
Healthcare resource use and costs
Fine 2003 reported length of initial hospital stay and re‐admissions at 30 days after index hospitalisation. Shah 2014 reported (primary outcome in clinical study) the proportion of patients with diabetes at high risk of a cardiovascular event who were prescribed a statin (see Table 4 for details on secondary outcomes reported).
| Author Year | Clinical outcomes; Medical complications | Quality of life Satisfaction with care | Mortality | Healthcare resource use (including medications prescribed) | Costs |
| ‐ | Quality of Life (assessed with the EQ‐5D1), mean (SD): BL: Intervention: 0.6730 (0.2042); Control: 0.6134 (0.2661), P = 0.006. At 6 weeks: Intervention: 0.7778 (0.1978); Control: 0.7497 (0.2316) At 12 weeks: Intervention: 0.8141 (0.1988); Control: 0.7873 (0.2210) Note: results for 26 and 52 weeks reported graphically | ‐ | ‐ | Mean annual cost (Euros) per patient (SD): Direct cost2: Intervention: 374 (437). Control: 449 (572). The costs (Euro) of releasing a new guideline for low back pain to 18,000 physiotherapists: Intervention (active strategy): 87,416 Control (passive strategy): 63,101 | |
| ‐ | ‐ | ‐ | ‐ | Cost paper awaits translation | |
| In‐hospital medical complications, number (%): Intervention:157 (55); Control: 206 (63), P = 0.04 Functional status3 SF‐12 physical health composite score: Intervention:45 ± 7, n = 181; SF‐12 mental health composite score: Intervention: 45 ± 6; Control: 45 ± 7, P = 0.71 | Patient satisfaction with care4 , number (%): Not satisfied with overall care: Intervention: 12 (5.3), n = 228; Control: 11 (4.0), n = 273, P = 0.67 Believed length of stay was too short: Intervention: 59 (26.1); Control: 54 (20.2), P = 0.16 Return to usual activities5,Hazard ratio (95% CI): Nonworkers: 1.09 (0.83 to 1.43), P = 0.55 Workers: 0.85 (0.54 to 1.35); P = 0.49 Return to work (workers) 0.99 (0.63 to 1.58), P = 0.98 | Mortality6 all‐cause, number (%): Intervention: 22 (8), n = 283; Control: 29 (9), n = 325, P = 0.70 Pneumonia‐related mortality, number (%): Intervention: 15 (5); Control: 23 (7), P = 0.44 | Length of index hospital (days) stay, median (IQR): Intervention: 5.0 (3.0 to 7.0); Control: 5.0 (3.0 to 8.0); Hazard ratio (95% CI): 1.16 (0.97 to 1.38), P = 0.11 Rehospitalisation7 number (%): Intervention:37 (14); Control:33 (11), P = 0.42 Duration (days) of intravenous antibiotic therapy, median (IQR): Intervention: 3.0 (2.0 to 5.0),n = 283; Control: 4.0 (2.6 to 6.0), n = 325; Hazard ratio (95% CI): 1.23 (1.00 to 1.52), P = 0.06 | ‐ | |
| Clinical data study: Intervention: n = 40 practices/795 patients; Control: n = 40 practices/797 patients Cardiovascular risk reduction (secondary outcomes): Proportion of participants reaching glycaemic control targets (HbA1c < 7.0%): Intervention: 58.5%; Control: 58.8%; OR 0.93 (0.71 to 1.21), P = 0.58 Proportion of participants reaching blood pressure control targets (< 130/80): Intervention: 52.8%; Control: 63.5%, OR 0.72 (0.53 to 0.98), P = 0.04 Proportion of participants reaching LDL‐cholesterol control targets (< 2.0 mmol/L): Intervention: 59.2%; Control: 61.7% , OR 0.90 (0.68 to 1.18), P = 0.43 Proportion of participants reaching Total to HDL‐cholesterol ratio (< 4.0): Intervention: 74.2%; Control: 76.8%, OR 0.85 (0.63 to1.14), P = 0.27 Clinical (secondary outcomes): When HbA1c > 8.0%: Intervention: 11.8%; Control: 13.0%, OR 0.98 (0.48 to 1.98), P = 0.95 When blood pressure > 140/90: Intervention: 5.6%; Control: 7.2%, OR 0.67 (0.25 to 1.82), P = 0.43 When LDL cholesterol > 3.0 mmol/L: Intervention: 43.5%; Control: 45.2%, OR 0.94 (0.53 to 1.67), P = 0.83 Administrative data study : Intervention: 2008 practices (467,713 participants); Control: 1999 practices (466, 076 participants) Secondary outcomes: | ‐ | Administrative data study : Intervention: 2008 practices (467,713 participants); Control: 1999 practices (466,076 participants) Primary outcome: Death or non‐fatal myocardial infarction: Intervention: 2.5%; Control: 2.5%; OR 1.00 (0.96 to 1.03), P = 0.77 Secondary outcomes: Medication initiation (ACEI/ARB > 1 antihypertensive class, or > 2, or > 3, statin, glucose‐lowering drugs, insulin, nitrate): OR, range: from 0.96 to 1.02, P values from 0.03 to 0.94 | Clinical data study: Primary outcome: Proportion of participants prescribed statins (new or renewed prescription): Intervention: 700 (88.1%); Control: 725 (90.1%); OR 0.73, 95% CI 0.42 to 1.26, P = 0.26 Proportion of participants prescribed an ACEI/ARB: Intervention: Control: Secondary outcome. Administrative data study : Secondary outcomes: CAD assessment (electrocardiogram, cardiac stress test, nuclear imaging, coronary angiography, coronary revascularisation, cardiology or internal medicine visit): OR, range: from 0.96 to 1.00, P values from 0.02 to 0.83 | ‐ |
1EQ‐5D: a standardised instrument for use as a measure of health outcome.The EQ‐5D has five dimensions: mobility, self‐care, usual activity,pain/discomfort and anxiety/depression. Each dimension has three levels, no problems, some problems and serious problems. Hence, EQ‐5D has 243 possible health states. Utility values of the general public for these health states as measured with the time tradeoff technique on a random sample of the adult population of the United Kingdom, the MVH‐A1 tariff, were applied in this study. The scores range from −0.594 (worst situation) to 1.0 (perfect health).
2The direct costs consisted of costs of the dissemination of the guideline and the costs of the healthcare utilisation of the patients. Prices for the year 2002.
3SF‐12 health scores were assessed in all patients able to provide reliable self‐report data during the 30‐day interview, excluding 6 intervention‐arm and 6 control‐arm patients with missing data.
4Patient satisfaction with care was assessed for all patients with a 30‐day interview that was not completed by a paid caregiver, excluding four intervention‐arm and two control‐arm patients with missing data. An additional two patients in the intervention arm and six patients in the control arm were hospitalised for the full 30 days and were not asked about length of hospital stay. SF‐12, 12‐Item Short Form was used.
5Return to usual activities among non‐workers was assessed for 183 intervention arm and 219 control arm patients not employed at baseline who completed a 30‐day interview. Return to usual activities among workers was assessed in 59 intervention‐arm and 59 control‐arm patients employed at baseline. Return to work was assessed among 54 intervention‐arm and 53 control‐arm patients employed at baseline.
6Mortality, medical complications, and return to work and usual activities were adjusted for pneumonia severity risk class.
7Rehospitalisation within 30 days of the index admission was assessed for all patients who were discharged alive from either the index hospitalisation or another acute‐care facility (if transferred to an acute‐care facility from the index hospitalisation).
8Fluid fasting times assessed by local investigator asking the patient about the fasting time, and checking this information against medical notes.
9Cost for designing, editing, reproducing, and posting need when applied to 170 acute trusts.
10Cost of providing 170 acute trusts with implementation support through a web‐based resource championed through opinion leadership. This includes development costs for the tool (which for this project were in‐house costs, in other cases external agencies may have to be used which are likely to be three times higher), publicity materials, training materials and opinion leader time and activity.
Two studies reported on costs (Bekkering 2005; Saillour‐Glénisson 2005 (belonging to Daucourt 2003)). One of the studies reported mean annual cost per patient, total cost for releasing the guideline and cost of active implementation intervention (Bekkering 2005). The other article awaits translation (Saillour‐Glénisson 2005).
Patient outcomes
Bekkering 2005 reported quality‐of‐life measures at four weeks. Fine 2003 reported all‐cause and pneumonia‐related mortality, medical complications, functional status and patient satisfaction with care at 30 days after the initial hospitalisation. Shah 2014 reported (primary outcome in administrative data study) death or non‐fatal myocardial infarction. Daucourt 2003 reported the number of requests for a thyroid function test that complied with the guidelines (Guideline Conformity Rate (GCR)) at 4 weeks. (See Table 4 for details of the secondary outcomes reported).
Excluded studies
After scrutinising the full text we excluded 1030 studies and added 14 to the excluded studies table. See Characteristics of excluded studies table.
Risk of bias in included studies
See 'Risk of bias' tables within the Characteristics of included studies,Figure 2 and Figure 3.
Risk of bias summary: review authors' judgements about each risk of bias item for each included study. White space indicate studies not reporting non‐objective outcomes and for which risk of bias could not be assessed.
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies. White spcace indicate studies not reporting non‐objective outcomes and for which risk of bias could not be assessed.
The randomisation sequence and the allocation concealment were adequate in three studies (Bekkering 2005; Daucourt 2003; Shah 2014), and unclear in Fine 2003. In Shah 2014 blinding was adequate (clinical data study assessed), and in Daucourt 2003 it was unclear whether or not the healthcare professionals were blinded, while in two studies (Bekkering 2005; Fine 2003) it was clear that they were not. In one study the healthcare professional selected a maximum of 10 consecutive patients for the study, and we therefore judged the risk of performance bias to be high (Bekkering 2005). Performance bias was also judged high in Fine 2003 as treatment assignment was not concealed. Blinding of outcome assessment was adequate in two studies (Bekkering 2005; Shah 2014), and unclear in the other two. Baseline characteristics were reported to be similar in one study (Fine 2003), not similar in one study (Shah 2014), and unclear in the other two studies. The outcome data were complete in two studies (Daucourt 2003; Shah 2014), and unclear in the other two (with losses to follow‐up of more than 20%). In Shah 2014 some of the outcomes that were listed in the trial protocol were not in the study report, while in the other three studies the risk of selective reporting was low. Shah 2014 had unclear risk of other bias (contamination), while the other three were at low risk.
Effects of interventions
See: Summary of findings for the main comparison
i) Interventions targeting the health care professionals
Healthcare professional outcomes
See summary of findings Table for the main comparison; Table 4 and Table 5.
| Author Year | Adherence Outcomes | Participants (Settings) | Control Adherence | Intervention Adherence | Median ARD |
| (Hoijenboos 2005) Targeted behaviour: management of non specific low back pain GL tool used: interactive training workshop X2 | Adherence to 4 guideline recommendations: i) Limit number of sessions in normal course back pain ii) Set functional treatment goals iii) Use mainly active interventions iv) Give adequate information Note: an increase was desirable for all outcomes | 113 physiotherapists (68 private physiotherapy practices) | i) Post: 14 (13), n = 253 ii) Post: 180 (71) iii) Post: 154 (60) iv) Post: 221 (87) | i) Post: 32 (27), n = 247 ii) Post: 188 (79) iii) Post: 183 (77) iv) Post: 229 (96) | +0.115 11.5% higher adherence in the intervention group i) 0.14% ii) 0.08% iii) 0.17% iv) 0.09% |
| Targeted behaviour: appropriate thyroid function testing Guideline tool used: | Global Guideline Conformity Rate | 1412 physicians (6 general hospitals) | Pre: 62.0% (95% CI 47.7 to 76.4) | Dual intervention group: Post: 77.9% (95% CI 68.9 to 87.0) Note: only results for the dual intervention presented here | +0.159% |
| Targeted behaviour: appropriate duration of intravenous antibiotic therapy for treatment of pneumonia GL tool used: detail sheet/ treatment recommendations+ telephone reminder | No adherence outcomes reported, only proxies | 545 physicians (7 not‐for‐profit hospitals) | ‐ | ‐ | ‐ |
| Targeted behaviour: improved cardiovascular risk factor management in people with diabetes Guideline tool used: printed educational material | No adherence outcomes reported, only proxies | 2 separate studies: Administrative data study: n = 4007 practices: Intervention: 2008; Control: 1999 Clinical data study. n = 80 practices (1592 patients); Intervention: 40 practices (8795 patients); Control: 40 practices (8797 patients) | ‐ | ‐ | ‐ |
Two of the four included studies (Bekkering 2005; Daucourt 2003) reported one or more measures of healthcare professionals' adherence to guidelines. The overall median absolute risk difference (range) (five comparisons) was (range: 0.115 to 0.159), i.e. a median difference in adherence of 13.5%, with the effects ranging from 11.5% to 15.9% increase in adherence.
Fine 2003, in which physicians received an educational mailing, a daily assessment of (pneumonia) patient stability and an additional sheet to the medical notes with a follow‐up recommendation for converting from intravenous to oral antibiotic and hospital discharge, compared with education mailing alone, reported that those in the intervention group had a more rapid rate of conversion to oral antibiotics (hazard ratio (HR) 1.23, 95% confidence interval (CI) 1.00 to 1.52, P = 0.06). Shah 2014 did not report health professional outcomes.
Healthcare resource use and costs
Fine 2003 reported similar percentages in each group of patients being readmitted (intervention group 14% versus 11% in the control group), and a similar length of initial hospital stay (median of five days in each group) at 30 days after index visit.
Shah 2014 reported similar or slightly lower (= undesired effect) use of different types of coronary artery disease (CAD) assessment tools in practices that received the guideline tool compared to those who received the updated guideline only (administrative data study), as was the case for the medication initiation outcomes (both were secondary outcomes).
Bekkering 2005 reported mean annual direct medical costs for the intervention group of EUR 374 versus EUR 449 in the control group. Direct costs included costs of the dissemination of the guideline and healthcare resource use by the patient. Daucourt 2003 reported prescribing cost data in a paper in French (Saillour‐Glénisson 2001) which awaits translation.
Patient outcomes
SeeTable 4 for details
Bekkering 2005 reported similar quality‐of‐life scores for patients with non‐specific low back pain at 12 months.
Fine 2003 reported similar scores on the SF‐12 physical component score (intervention group 45 (standard deviation (SD) 7) versus control group 45 (SD 7)) and the mental component score (intervention group 45 (SD 6) versus control group 45 (SD 7)) at 30 days after index stay, and little or no difference for mortality (intervention group 8% versus control group 9%), and return to work (HR 0.99, 95% CI 0.63 to 1.58). The same authors reported fewer hospital complications in the intervention group compared with control (157 (55%) and 206 (63%) respectively, P = 0.04).
Shah 2014 reported little or no difference between groups (Intervention 2.5%; Control 2.5%; odds ratio (OR) 1.00, 95% CI 0.96 to 1.03, P = 0.77) for death and non‐fatal myocardial infarction (primary outcome in the administrative data study), and also little or no difference for any of the other (secondary) clinical events reported (see Table 4 for details).
ii) Interventions targeting the organisation of care
No studies reported results for this comparison.
iii) Interventions targeting the patient
No studies reported results for this comparison.
iv) Interventions targeting the healthcare professional, the organisation of care and/or the patient
No study reported results for this comparison.
Effectiveness of different approaches of guideline dissemination
We include four studies in this review, of which one evaluated the effectiveness of two short tailored educational workshops, and the other three studied the effects of using paper‐based tools, including order forms or reminders, or both. As the types of multifaceted interventions, the clinical condition and behaviour targeted varied across studies it was not possible to determine which of the different approaches used to improve implementation of guidelines was most effective.
Discussion
Summary of main results
We identified four eligible cluster‐RCTs for inclusion in this review, evaluating the effects of tools developed by existing guideline producers to improve implementation of their guidelines.
All included studies evaluated tools that targeted the healthcare professional. However, meta‐analysis was not feasible, since the targeted clinical conditions and behaviour, as well as the guideline tools used, all varied between studies. The variation in the duration of interventions and follow‐up also made comparisons difficult. Tools developed by guideline producers, and given to healthcare professionals as an aid to improve compliance, probably lead to greater adherence to guidelines (median absolute risk difference (ARD) 13.5%) at an average four weeks follow‐up (moderate‐quality evidence). The effect ranged from 11.5% in one study (two tailored short educational workshops to improve management of non‐specific low back pain) to 15.9% in the other (a pocket memorandum card and test‐request form to improve thyroid‐test ordering). Neither study reported baseline adherence, and it appeared that no guideline for the specific targeted behaviours and conditions was previously in place. There was low certainty of evidence from one trial for little or no difference in costs between groups.
Due to the few eligible studies identified, and the variety of interventions implemented, we could not determine which approaches are most effective, which was the secondary objective of this review. Two of the included studies reported on cost data, and one of these awaits translation. While it is not possible to directly address the investment made by guideline producers in developing implementation tools, the cost is not likely to differ substantially from other organisations that develop tools to support the implementation of guidelines. It should be noted that even small to moderate intervention effects may be highly cost‐effective if the targeted clinical condition is highly prevalent and the implementation tools used are inexpensive to develop and to disseminate.
There is no evidence available for the effectiveness of interventions targeting the organisation of care or the patient.
Overall completeness and applicability of evidence
In all included studies the interventions targeted the healthcare professional. None of the included studies used tools that targeted health system managers or policy makers, the patient (e.g. versions of the guideline developed for the patient), or the organisation of care (e.g. benchmarking tools, costing templates or programme evaluation, audit tools, performance measures and quality indicators to evaluate the implementation of the CPG), and no study evaluated the effects of mass media interventions. The implementation tools used were delivered alongside the CPG, and none was imbedded within the CPG (e.g. tailoring of the CPG for a specific audience). In addition, only guideline tools to promote the use of CPGs for a few clinical conditions and behaviours have been evaluated.
Quality of the evidence
The evidence was from cluster‐RCTs that had all taken clustering into account in the analysis. We downgraded all included studies from high to moderate certainty of evidence for the main outcome (adherence to guidelines), due to high risk of bias. As only a single study provided evidence for the effectiveness of a certain implementation on costs, our confidence in the evidence was further downgraded to low due to imprecision.
Potential biases in the review process
We searched a large number of databases using a strategy that was designed by a senior information specialist, and then adapted for different databases. We also searched a large number of websites of relevant guideline producers. Four review authors sifted a number of references identified by the electronic searches, excluding papers that were irrelevant and clearly not eligible, and producing a long list for a second review author to go through. Two review authors independently assessed all potentially eligible titles and abstracts against the eligibility criteria to ensure no important references were missed. We also performed data extraction and assessment of risk of bias in duplicate.
Agreements and disagreements with other studies or reviews
We are not aware of any other reviews that have evaluated the effectiveness of tools developed by recognised guideline producers to improve implementation of their own CPGs. However, our results of a median 13.5% greater adherence to guidelines in the intervention group (two studies: one evaluating a paper‐based intervention, and one an intervention consisting of two short educational workshops) are greater than the reported median absolute improvement in performance for point‐of‐care computer reminders of around 4% (Shojania 2009), 2% for printed educational materials (Giguère 2012) and 6% for educational meetings (Forsetlund 2009). These reviews, however, included a much larger number of studies and participants, which may explain the differences in effect.
Risk of bias summary: review authors' judgements about each risk of bias item for each included study. White space indicate studies not reporting non‐objective outcomes and for which risk of bias could not be assessed.
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies. White spcace indicate studies not reporting non‐objective outcomes and for which risk of bias could not be assessed.
| Clinical practice guideline (CPG) + implementation tool compared with CPG only for adherence to guidelines | ||||
| Patient or population: Healthcare professionals (physiotherapists, hospital physicians) providing care for people with different clinical conditions (patients with non specific low back pain, patients with symptoms indicating a need for a thyroid function test) Setting: Private physiotherapy clinics in the Netherlands, general hospitals in France Intervention: CPG + guideline implementation tool (e.g. training workshops, paper‐based materials and order forms, reminders, web‐based tools) | ||||
| Outcomes | Median ARD (Absolute risk difference) | No of Participants | Certainty of the evidence | Comments |
| Adherence to guidelines | Guideline tools provided to healthcare professionals as an aid to improve the use of a CPG probably lead to improved adherence with the CPG, as compared to guidelines only. Median ARD: 0.135 (0.115 to 0.15.9) at mean 4 weeks follow‐up | 68 physio‐ therapy practices; and 6 hospitals (2 C‐RCTs) | ⊕⊕⊕⊝ | 2 of the 4 included studies reported a proxy measure for adherence to guidelines, and results from these studies could therefore not be included in the ARD calculation |
| Costs | Guideline tools aimed at improving the use of guidelines may lead to little or no difference in healthcare costs | 68 physio‐ therapy clinics (1 C‐RCT) | ⊕⊕⊝⊝ low2 | 1 trial reported no difference in mean direct annual cost* per patient between intervention and control groups. 1 French paper belonging to 1 of the included trials (6 hospitals) and reporting on costs awaits translation |
| * Direct costs included costs of the dissemination of the guideline and healthcare resource use by the patient GRADE Working Group grades of evidence | ||||
| 1We downgraded the certainty of the evidence one step due to high risk of bias. 2We downgraded the certainty of the evidence two steps due to imprecision. | ||||
| Author Year Targeted behaviour | Guideline developers | Literature review | Critical appraisal | Consensus processes | Key stakeholder involvement | Barriers/facilitator assessment |
| Targeted behaviour: management of non‐specific low back pain Number of recommendations: 4 main recommendations | The Royal Dutch Physiotherapy Association. | CPGs2 were constructed on the basis of the phases of the physiotherapy process, using the Dutch method of developing physiotherapy guidelines, and evidence from systematic reviews were identified through searching electronic databases | Not mentioned but probably included in the Dutch method of developing CPGs | Based on scientific evidence. If no evidence was available, consensus between experts was obtained | The CPGs were pilot‐tested among 100 physiotherapists and reviewed by an external multidisciplinary panel | Barriers to change were assessed through a survey as part of the CPG development process |
| Targeted behaviour: appropriate thyroid function testing Number of recommendations: not reported | The Committee for Co‐ordinating Clinical Evaluation and Quality in | The CPG developers conducted a comprehensive review of the literature | ‐ | CPG3 development involved an expert consensus process. | ‐ | ‐ |
| Targeted behaviour: appropriate duration of intravenous antibiotic therapy for treatment of pneumonia Number of recommendations: a 2‐step recommendation | Researchers who were part of the Pneumonia Patient Outcomes Research Team (PORT) | The CPG was based on a review of the medical literature, and empiric evidence on time to reach clinical stability | ‐ | The CPG development process involved the consensus of an 8‐member national guideline panel | The guideline was reviewed by clinical opinion leaders at each study site, and was approved for local use by the relevant utilisation management department. | ‐ |
| Targeted behaviour: management of cardiovascular risk Number of recommendations: no information | Canadian Diabetes Association (and Expert Committee members) | Expert Committee members evaluated the relevant literature, and guidelines were developed and initially reviewed | After formulating new recommendations or modifying existing | Based on scientific evidence/review of the literature | A draft document was circulated nationally and | ‐ |
| 1Khunti 1998. Development of evidence‐based review criteria for the management of patients with depression in general practice. No published version of the guideline found. | ||||||
| Author Year | Delivery of the intervention | Theoretical models/ frameworks used | Evidence base | Targeted to barriers | Key stakeholder involvement |
| Type of intervention: simple, but multicomponent; active only Intervention target: the healthcare professional Targeted behaviour: management of nonspecific low back pain | Mode: face‐to‐face Provider: The primary investigator and 1 of 2 additional trainers with adequate clinical experience in the management of low back pain supervised the training sessions | ‐ | "The sessions were based on interventions that have all been shown to be effective, such as interactive education and discussion, feedback, and reminders".1,2,3,4,5 | The content of the strategy was determined on the basis of information about the expected barriers for implementation gathered during the development of the CPGs | 2 experts gave advice on the content of the strategy |
| Type of intervention: multicomponent; passive only Intervention target: the healthcare professional Targeted behaviour: appropriate thyroid function testing | Mode: Paper‐based materials Provider: none | ‐ | "Among the clinical guideline diffusion strategies, the most effective are feedback, reminders, academic detailing and financial incentives1,2,4,5 Administrative procedures such as the implementation of test request forms have also proved effective." | ‐ | ‐ |
| Type of intervention: single; active + passive Intervention target: the healthcare professional Targeted behaviour: appropriate duration of intravenous antibiotic therapy for treatment of pneumonia | Mode: Paper‐based material (detail sheets/treatment recommendations in patient records) and telephone reminder Provider: nurse delivered telephone reminder | ‐ | "The multifaceted guideline dissemination strategy consisted of interventions of proven benefit, including real‐time physician reminders" 1,6,7,8 | ‐ | ‐ |
| Type of intervention: passive Intervention target: family physicians (and diabetes patients at high risk of cardiovascular disease) Targeted behaviour: management of cardiovascular risk factors and outcomes of cardiovascular disease in people with diabetes | Mode: printed educational materials Targeting the family physician: The cardiovascular disease toolkit was a collection of printed educational materials, packaged in a brightly‐coloured box with CDA branding, sent to Canadian family physicians. The contents included an introductory letter from the Chair of the practice Provider: NA13 | The implicit theory behind its | "The literature has demonstrated that the benefits of printed educational interventions are, at best, modest. A systematic review of methods to improve practice guideline adherence demonstrated an absolute improvement of 8% for educational materials. A more recent Cochrane review found that printed educational materials led to a median absolute improvement in performance of only 2% (25). Studies of printed materials specifically tied to clinical practice guidelines also showed modest benefits. A small English study randomised 42 family physicians to receive an algorithm for monitoring and treatment of hypertension of diabetic patients based on practice guidelines, but found no difference in blood pressure control between the intervention and control groups. However, some processes of care were slightly improved: patients in the intervention group were prescribed higher doses of antihypertensive medications, and had more physician visits to monitor blood pressure. In a larger Canadian study, family physicians were randomised to receive a 1‐page summary of a 3‐year‐old practice guideline on anti‐anginal therapy from the local medical governing body. No differences were noted in prescription of β‐blockers, antiplatelet agents, or lipid‐lowering drugs between groups in the 7000 patients reviewed" 9,10,11,12 | ‐ | The toolkit was |
| 1Bero 1998 Closing the gap between research and practice: an overview of systematic reviews of interventions to promote the implementation of research findings. | |||||
| Author Year | Tailoring | Feedback | Educational outreach/ Academic detailing/ Small group discussions | Reminders (paper, electronic, telephone) | Decision support tools | Other (test order forms, supportive materials etc.) |
| The content of the strategy was determined on the basis of information about the expected barriers for implementation gathered during the development of the clinical guidelines. | ‐ | 2 interactive training sessions, each lasting 2½ hours, for groups of 8 – 12 physiotherapists (including feedback on current management and reminders). For each session a preparation time of 2 hours was recommended. | ‐ | ‐ | ‐ | |
| ‐ | ‐ | ‐ | Pocket memorandum card. | ‐ | Test request form. | |
| ‐ | ‐ | ‐ | Paper‐based detail sheet/ treatment recommendations put into the patient's record + real‐time nurse telephone reminder | ‐ | ‐ | |
| ‐ | ‐ | ‐ | ‐ | ‐ | Printed educational materials |
| Author Year | Clinical outcomes; Medical complications | Quality of life Satisfaction with care | Mortality | Healthcare resource use (including medications prescribed) | Costs |
| ‐ | Quality of Life (assessed with the EQ‐5D1), mean (SD): BL: Intervention: 0.6730 (0.2042); Control: 0.6134 (0.2661), P = 0.006. At 6 weeks: Intervention: 0.7778 (0.1978); Control: 0.7497 (0.2316) At 12 weeks: Intervention: 0.8141 (0.1988); Control: 0.7873 (0.2210) Note: results for 26 and 52 weeks reported graphically | ‐ | ‐ | Mean annual cost (Euros) per patient (SD): Direct cost2: Intervention: 374 (437). Control: 449 (572). The costs (Euro) of releasing a new guideline for low back pain to 18,000 physiotherapists: Intervention (active strategy): 87,416 Control (passive strategy): 63,101 | |
| ‐ | ‐ | ‐ | ‐ | Cost paper awaits translation | |
| In‐hospital medical complications, number (%): Intervention:157 (55); Control: 206 (63), P = 0.04 Functional status3 SF‐12 physical health composite score: Intervention:45 ± 7, n = 181; SF‐12 mental health composite score: Intervention: 45 ± 6; Control: 45 ± 7, P = 0.71 | Patient satisfaction with care4 , number (%): Not satisfied with overall care: Intervention: 12 (5.3), n = 228; Control: 11 (4.0), n = 273, P = 0.67 Believed length of stay was too short: Intervention: 59 (26.1); Control: 54 (20.2), P = 0.16 Return to usual activities5,Hazard ratio (95% CI): Nonworkers: 1.09 (0.83 to 1.43), P = 0.55 Workers: 0.85 (0.54 to 1.35); P = 0.49 Return to work (workers) 0.99 (0.63 to 1.58), P = 0.98 | Mortality6 all‐cause, number (%): Intervention: 22 (8), n = 283; Control: 29 (9), n = 325, P = 0.70 Pneumonia‐related mortality, number (%): Intervention: 15 (5); Control: 23 (7), P = 0.44 | Length of index hospital (days) stay, median (IQR): Intervention: 5.0 (3.0 to 7.0); Control: 5.0 (3.0 to 8.0); Hazard ratio (95% CI): 1.16 (0.97 to 1.38), P = 0.11 Rehospitalisation7 number (%): Intervention:37 (14); Control:33 (11), P = 0.42 Duration (days) of intravenous antibiotic therapy, median (IQR): Intervention: 3.0 (2.0 to 5.0),n = 283; Control: 4.0 (2.6 to 6.0), n = 325; Hazard ratio (95% CI): 1.23 (1.00 to 1.52), P = 0.06 | ‐ | |
| Clinical data study: Intervention: n = 40 practices/795 patients; Control: n = 40 practices/797 patients Cardiovascular risk reduction (secondary outcomes): Proportion of participants reaching glycaemic control targets (HbA1c < 7.0%): Intervention: 58.5%; Control: 58.8%; OR 0.93 (0.71 to 1.21), P = 0.58 Proportion of participants reaching blood pressure control targets (< 130/80): Intervention: 52.8%; Control: 63.5%, OR 0.72 (0.53 to 0.98), P = 0.04 Proportion of participants reaching LDL‐cholesterol control targets (< 2.0 mmol/L): Intervention: 59.2%; Control: 61.7% , OR 0.90 (0.68 to 1.18), P = 0.43 Proportion of participants reaching Total to HDL‐cholesterol ratio (< 4.0): Intervention: 74.2%; Control: 76.8%, OR 0.85 (0.63 to1.14), P = 0.27 Clinical (secondary outcomes): When HbA1c > 8.0%: Intervention: 11.8%; Control: 13.0%, OR 0.98 (0.48 to 1.98), P = 0.95 When blood pressure > 140/90: Intervention: 5.6%; Control: 7.2%, OR 0.67 (0.25 to 1.82), P = 0.43 When LDL cholesterol > 3.0 mmol/L: Intervention: 43.5%; Control: 45.2%, OR 0.94 (0.53 to 1.67), P = 0.83 Administrative data study : Intervention: 2008 practices (467,713 participants); Control: 1999 practices (466, 076 participants) Secondary outcomes: | ‐ | Administrative data study : Intervention: 2008 practices (467,713 participants); Control: 1999 practices (466,076 participants) Primary outcome: Death or non‐fatal myocardial infarction: Intervention: 2.5%; Control: 2.5%; OR 1.00 (0.96 to 1.03), P = 0.77 Secondary outcomes: Medication initiation (ACEI/ARB > 1 antihypertensive class, or > 2, or > 3, statin, glucose‐lowering drugs, insulin, nitrate): OR, range: from 0.96 to 1.02, P values from 0.03 to 0.94 | Clinical data study: Primary outcome: Proportion of participants prescribed statins (new or renewed prescription): Intervention: 700 (88.1%); Control: 725 (90.1%); OR 0.73, 95% CI 0.42 to 1.26, P = 0.26 Proportion of participants prescribed an ACEI/ARB: Intervention: Control: Secondary outcome. Administrative data study : Secondary outcomes: CAD assessment (electrocardiogram, cardiac stress test, nuclear imaging, coronary angiography, coronary revascularisation, cardiology or internal medicine visit): OR, range: from 0.96 to 1.00, P values from 0.02 to 0.83 | ‐ | |
| 1EQ‐5D: a standardised instrument for use as a measure of health outcome.The EQ‐5D has five dimensions: mobility, self‐care, usual activity,pain/discomfort and anxiety/depression. Each dimension has three levels, no problems, some problems and serious problems. Hence, EQ‐5D has 243 possible health states. Utility values of the general public for these health states as measured with the time tradeoff technique on a random sample of the adult population of the United Kingdom, the MVH‐A1 tariff, were applied in this study. The scores range from −0.594 (worst situation) to 1.0 (perfect health). | |||||
| Author Year | Adherence Outcomes | Participants (Settings) | Control Adherence | Intervention Adherence | Median ARD |
| (Hoijenboos 2005) Targeted behaviour: management of non specific low back pain GL tool used: interactive training workshop X2 | Adherence to 4 guideline recommendations: i) Limit number of sessions in normal course back pain ii) Set functional treatment goals iii) Use mainly active interventions iv) Give adequate information Note: an increase was desirable for all outcomes | 113 physiotherapists (68 private physiotherapy practices) | i) Post: 14 (13), n = 253 ii) Post: 180 (71) iii) Post: 154 (60) iv) Post: 221 (87) | i) Post: 32 (27), n = 247 ii) Post: 188 (79) iii) Post: 183 (77) iv) Post: 229 (96) | +0.115 11.5% higher adherence in the intervention group i) 0.14% ii) 0.08% iii) 0.17% iv) 0.09% |
| Targeted behaviour: appropriate thyroid function testing Guideline tool used: | Global Guideline Conformity Rate | 1412 physicians (6 general hospitals) | Pre: 62.0% (95% CI 47.7 to 76.4) | Dual intervention group: Post: 77.9% (95% CI 68.9 to 87.0) Note: only results for the dual intervention presented here | +0.159% |
| Targeted behaviour: appropriate duration of intravenous antibiotic therapy for treatment of pneumonia GL tool used: detail sheet/ treatment recommendations+ telephone reminder | No adherence outcomes reported, only proxies | 545 physicians (7 not‐for‐profit hospitals) | ‐ | ‐ | ‐ |
| Targeted behaviour: improved cardiovascular risk factor management in people with diabetes Guideline tool used: printed educational material | No adherence outcomes reported, only proxies | 2 separate studies: Administrative data study: n = 4007 practices: Intervention: 2008; Control: 1999 Clinical data study. n = 80 practices (1592 patients); Intervention: 40 practices (8795 patients); Control: 40 practices (8797 patients) | ‐ | ‐ | ‐ |
