The guideline implementability research and application network (GIRAnet): an international collaborative to support knowledge exchange: study protocol

Guidelines are syntheses of best available evidence that, along with professional judgment and patient preferences, support decision making by clinicians, managers, and policy makers about the organisation and delivery of healthcare. However, they continue to be underused [1‐7]. Research has shown that guideline format and content influence perceptions about and use of guidelines. Specifically, these intrinsic guideline qualities have been shown to promote greater understanding of how users are to apply the recommendations, stimulating confidence in users' ability to practice the recommended behavior, leading to greater intent to use guidelines and actual use [8‐13]. Thus, use of guidelines might be optimised by improving their format and content. The concept of implementability was first defined by Shiffman as characteristics of guidelines that may enhance their implementation by users, and he issued consensus recommendations for generating guidelines with actionable wording [14]. To further investigate the concept of implementability, we reviewed the medical literature to identify features desired by different users or associated with guideline use [15]. The guideline implementability framework included 22 elements organised within eight domains: adaptability, usability, relevance, validity, applicability, communicability, resource implications, implementation, and evaluation (Table 1). Our analysis of guidelines on various clinical indications judged by experts as high quality found that most did not contain implementability elements, highlighting numerous opportunities to potentially improve guideline development and use by integrating one or more of these elements.

Further research is needed to operationalise the implementability concept by developing and evaluating implementability tools. Validation and definitive testing of tools and interventions based on the implementability framework will require considerable collaboration with guideline developers, implementers, and users to more strategically generate knowledge, build a consolidated research base, and accelerate its application into practice. To accomplish this, a more formalised network is needed to leverage and sustain existing relationships and resources and to create capacity for collaboration on research and application of guideline implementability. With funding from the Canadian Institutes of Health Research, we are launching the Guideline Implementability Research and Application Network (GIRAnet). Partners include guideline agencies in Canada, United States, Australia, New Zealand, The Netherlands, Italy, Scotland, England, and Finland and the Guidelines International Network (G-I-N). G-I-N is a nonprofit association of guideline developers, implementers, and users, including 94 organisational and 76 individual members from 46 countries http://​www.​g-i-n.​net. G-I-N represents a natural gateway for research collaboration with stakeholders. The purpose of the network is to

The GIRAnet structure includes the administrative site, steering committee, research group, and various levels of membership. The administrative site will lead and coordinate all network activities; lead the development of implementability tools; and support distributed development, implementation, and evaluation of implementability tools. Prioritisation and direction for network activities will be provided by a steering committee that will advise on these matters. It is comprised of representatives from guideline-development agencies in nine partner countries plus G-I-N. Some examples of network activities include compiling a directory of existing implementability tools; developing and sharing guidance on the development and use of implementability tools, including toolkits and training sessions; conducting a needs assessment of guideline developers and implementers to better understand the resources needed to develop implementability tools; planning mechanisms by which to sustain a research network focused on guideline implementation; and generating a user-informed research agenda. A research group will provide input on research activities, including methods for the development and evaluation of tools and interpretation of evaluation findings. It is comprised of the manuscript co-authors, who are collaborating on operationalising the implementability concept by focusing their efforts on differing domains, and others who undertake academic research related to the development, implementation, and use of actionable guidelines. Members refers to anyone interested in guideline development, implementation, or related research who may wish to remain informed about network activities and products or take a more active role and participate in the evaluation of implementability tools as described in the next section.

The specific mandate of GIRAnet is to enable research that will develop, implement, and evaluate tools and interventions based on the implementability framework. The overall research plan is based on the MRC Framework for Developing and Evaluating Complex Interventions [33]. This approach includes five steps: development, pilot testing, evaluation, reporting, and implementation. However, evaluation must be informed by, and tailored to, the results of development and pilot testing; therefore, the proposed work will take place over three years and address development and pilot testing and planning for subsequent evaluation (Table 2). We will focus on identifying and/or developing implementability tools for guidelines on arthritis, cancer (breast, prostate, colorectal, lung), chronic obstructive pulmonary disease, depression, diabetes, ischemic heart disease, and stroke. These are relevant to national funding priorities in Canada, affect both men and women, and are the major cause of death and disability worldwide.

We will focus on three specific implementability domains/elements not under investigation by others, for which there is little or no research and, according to our research, are seldom included in guidelines (Table 3). Resource Implications refers to equipment or technology needed; industrial standards; policies governing their use; type and number of health professionals needed to deliver services; education, training, or competencies needed by staff to deliver services; anticipated changes in workflow or processes during or after adoption; and costs. Health professionals we interviewed said that this type of information would help them prepare for the impact on practice of adopting new guidelines [34]. Implementation refers to identifying barriers associated with adoption and selecting and tailoring implementation strategies that address those barriers. Our research found that professionals who fund, manage, and deliver health services lack knowledge about how to implement guidelines. Interviews with policy makers, managers, health professionals, and researchers revealed confusion about responsibility and approaches for implementation [32]. Evaluation refers to tools based on performance measures to assess baseline and postintervention compliance with guidelines. Our research found that self-assessment tools based on guidelines are lacking [35]; physicians said they lacked tools to monitor their own performance [36], and providing physicians with self-assessment tools and instructions resulted in identification of learning needs in 66.7% of patient cases they reviewed and modifications in intended care plans for 34.2% of those cases [37].

Cognitive science theory suggests that guidelines may be difficult to use because they present complex information that prescribes action that may not match clinical circumstances or user preferences, and individual knowledge and experience shape the way that information in a guideline is processed [38]. Guidelines featuring implementability tools may overcome these limitations to facilitate guideline interpretation and use. However, we need greater insight on the mechanism(s) by which this occurs to better understand how guideline format and content can be optimised. Through literature review, we compiled a conceptual framework that describes how implementability elements may influence type and process of decision making and guideline use and outcomes. Implementability elements may support various types of decisions: evidence informed (effectiveness data), experiential (professional expertise), shared (negotiation with patients/caregivers), and policy decision making (resource allocation) [15]. Implementability elements may support different types of decision-making processes: intuitive–trigger or reconcile with previous experience–and analytic–create or simulate new mental model [16‐20]. This framework will guide data collection and analysis. Study findings will validate and extend the framework and help us to refine prototype implementability tools (Figure 1).

The content of international guidelines on indications of interest will be analysed to identify exemplar implementability tools reflecting domains of interest (Resource Implications, Implementation, Evaluation). Content analysis describes phenomena in written, verbal, or visual communication to generate or validate a framework or model [39]. We will use a directed approach [40]. This means that explicit content in guidelines will be coded using elements from the implementability framework. We used these methods when we first examined guidelines for implementability [15]. To supplement exemplars, or if none are identified, literature searches will be conducted to identify information reflecting implementability elements with which to populate tools and templates. For example, if guidelines for arthritis management contained little or no resource-implications information, we will search the medical literature to identify relevant technical, regulatory, human resource, and workflow issues. Similarly, if no diabetes guidelines contained performance-assessment instructions, we will search the medical literature for examples of assessment measures with which to create program or evaluation tools. The search strategies themselves will ultimately be used in toolkits for developers, along with indication-specific and generic implementability tools that are relevant across disease indications so that they can find similar information for other guidelines.

Cognitive interviewing will be used to understand how developers can best integrate implementability tools in guidelines and how health professionals can use them for interpreting and applying guidelines. Findings will be used to refine indication-specific prototypes and generic templates in the instructional manual. This approach is based on cognitive theory to understand human information processing (attention span, word recognition, language processing, action, problem solving, reasoning) and has been used widely to understand how respondents interpret survey questions and for usability testing of information technology [41]. It therefore can be applied to study how developers and users perceive the content, format, use, and impact of implementability tools. The fundamental procedure is the semistructured interview. Interviewing can take place concurrently or subsequent to the respondent reviewing the prototype implementability tools in question, and we will use both to minimise recall bias. The process involves probing, where respondents paraphrase information, define meanings, and identify information that is difficult to understand or use. For example, questions may address the content and format of the tool, feasibility of using the tool and anticipated acceptability of the tool by colleagues, and perceptions of how the tool will be used. It also involves think-aloud protocols, where respondents are asked to verbalise their thought processes as they read and interpret and consider information in the tool, its meaning, and how it will be used. The interviewer classifies problems and determines how the tool should be refined, including aspects that should be either removed or modified and how, and the reason for the change. Types of problems include lexical (understanding of the meaning of words and phrases), inclusion/exclusion (scope of the information), temporal (time taken to read, interpret, and consider information), logical (relevance of the information in general, and as presented), and computational (any problems that do not fall in above categories). Qualitative methods will be used for sampling, data collection, and data analysis [42].

Through analysis of exemplar guidelines and supplementary literature search, phase-one development will assemble and create indication-specific and generic implementability tools and templates. Through interviews with guideline developers and users, phase-two pilot testing will assess the feasibility of, and resources needed to develop, implementable guidelines featuring those tools and the impact of those implementability tools on attitude, confidence, outcome expectancy, and intent to use guidelines. An international, multisite quantitative and qualitative study can then be conducted to evaluate the impact of newly developed or modified guidelines featuring implementability tools on actual behaviour or use of guidelines. While before-after observational design is not the strongest test of impact, this is methodologically the most appropriate next step in the evaluation of implementability tools [43]. It will enable comparisons of impact across type of implementability tool and clinical indication and inform the planning of a future, more definitive, time series or pragmatic randomised study to evaluate the cost-effectiveness of implementable guidelines versus usual guidelines or other type of intervention. Furthermore, evaluation in multiple international sites enabled through G-I-N and GIRAnet means that fewer participating health professionals are needed from each site, recruitment and data collection can take place in a shorter period of time, and results are more broadly generalisable and more rapidly translated to practice. However, the conduct of such an international multisite study will be logistically challenging. Moreover, conduct will be dependent on the findings of phase-one development and phase-two pilot testing. Therefore, an intermediate step is needed to prepare for phase-three evaluation. This will take the form of a small-scale pilot test of the planned evaluation, including quantitative analysis of chart-based behavioural outcomes and qualitative analysis of interviews with participants to gather feedback that will inform and refine the subsequent full-scale evaluation.