Background

Knowledge synthesis has the potential to inform the management of health problems [1] and is integral to the health of the Canadian population [2]. A knowledge synthesis summarizes all pertinent studies on a specific question, can improve the understanding of inconsistencies in diverse evidence, and can define future research agendas [1, 3]. Knowledge synthesis is also an important part of the knowledge translation (KT) process (and ideally should form the ‘base unit’ of KT strategies for providers and policy makers), and be used to provide the evidence base for KT products including clinical practice guidelines, policy briefs and decision aids [4]. As such, knowledge synthesis can be used to interpret results of individual studies within the context of the totality of evidence. This is an important consideration, given that basing practice and policy decisions on a single study or expert opinion can be misleading [5].

Knowledge synthesis activities in healthcare have often focused on the methodologically rigorous Cochrane reviews, most commonly of interventions. The definition of a systematic review according to the Cochrane Collaboration is “A review of clearly formulated questions that uses systematic and explicit methods to identify, select, and critically appraise relevant research, and to collect and analyse data from the studies that are included in the review. Statistical methods (meta-analysis) may or may not be used to analyse and summarise the results of the included studies” [6]. However, Cochrane-like review methods may not always be applicable for answering all knowledge synthesis questions, particularly those investigating complex and multidisciplinary topics [7, 8]. For example, members of our team recently attempted to conduct a systematic review to better understand the relationship between the perceived characteristics of clinical practice guidelines and their uptake by clinicians, and found that a flexible approach that borrowed relevant components of less traditional knowledge synthesis methods (i.e., including realist reviews and meta-ethnography) was more relevant to determine the mechanisms and circumstances underpinning guideline implementation [9]. This example highlights the need for less traditional methods for completing a review. By matching the appropriate design to fit the question, synthesis outputs are more likely to be relevant and be useful for end users.

Furthermore, a traditional review such as a Cochrane review cannot always explain why particular interventions work in some settings but not in others [10]. For example, a Cochrane review found that school feeding programs significantly improved the growth and cognitive performance of disadvantaged children [11], but failed to provide direction for policy-makers to decide which intervention should be implemented and under what circumstances. By conducting a realist review alongside the Cochrane review (which can be used to understand ‘what works for whom and under what circumstances’ [10]), the authors were able to provide concrete recommendations that could be implemented in practice and policy making [7]. To address these types of questions and adequately incorporate the needs, preferences and experiences of patients into healthcare delivery, there is an increasing need to consider less traditional review methods of complex evidence (i.e., heterogeneous, methodologically diverse, difficult to classify, and contradictory) [12, 13]. Another approach is to consider conducting a systematic review as a “first step” to better understand complex evidence (or to conduct them in parallel with novel reviews), particularly for evidence generated from philosophy and the social sciences. The increasing number of synthesis methods that have recently emerged within the healthcare literature supports this need [14–17].

Table 1 summarizes a selection of knowledge synthesis methods that currently exist in the literature across multiple disciplines (identified through consultation with knowledge synthesis experts and qualitative researchers). Although many of these approaches can be applied to healthcare situations, the methods for conducting them have not been as clearly operationalized as traditional reviews of interventions. Consultation with researchers and end users of reviews that we conducted in preparation for this research indicate a lack of clarity around how to match the appropriate review method to the research question, the methods used to conduct these reviews, and how to analyze and present the results from the review to inform decision making. These issues are challenging for researchers interested in tackling reviews of complex questions and for decision makers trying to interpret and apply this evidence. Other identified challenges involve locating the numerous synthesis methods (particularly those for synthesizing qualitative research), which can be problematic and resource-intensive since they are scattered widely within the literature and across many different disciplines and databases. The terms used to describe the different synthesis methods are often similar (e.g., ‘meta-synthesis’, ‘meta-ethnography’, ‘meta-narrative’, ‘meta-study’, ‘meta-interpretation’) and their definitions can overlap [12]. This area of research is further complicated because some of these methods are referred to as a ‘complete’ synthesis method (i.e., providing guidance on the search strategy, study selection, appraisal, and analysis), while others provide guidance only on specific parts of the process, such as data analysis [12].

Table 1 Characteristics of a preliminary list of existing knowledge synthesis methods
Full size table

Some researchers have attempted to outline methods for the synthesis of qualitative [47] and mixed-methods research [36, 45] and to build a typology of such reviews [41], while others have highlighted methods for knowledge synthesis reviews to inform specific end-user targets such as for management and policy-making in the health field [45]. A recent overview by Gough and colleagues attempted to outline the differences between review designs and methods by describing the important conceptual and practical differences amongst them [8]. However, a comprehensive manual for all of the different synthesis methods (quantitative/qualitative or mixed), outlining how they are related and how to decide which methodology is the most appropriate for a particular research question does not currently exist. To our knowledge, the current study will be the first to describe an overall taxonomy of all existing types of knowledge synthesis methods, to characterise the differences between them, and to develop a strategy for knowledge users to be able to select the most appropriate method to answer their research questions.

The specific objectives of the current study are to: (1) to conduct a systematic search for knowledge synthesis methods across multi-disciplinary fields, such as health and philosophy; (2) compare and contrast the different knowledge synthesis methods; and, (3) map out the specific steps to conducting the knowledge synthesis methods, which will be used to inform the development of a knowledge synthesis methods manual/tool.

Methods/Design

Search strategy

We will use the methodologically rigorous scoping review approach proposed by Arksey and O’Malley [48] to conduct a systematic search across the disciplines of health and philosophy. We will search the following electronic databases from inception onwards: MEDLINE, CINAHL, EMBASE, PsycInfo, the Cochrane Methodology Register, Cochrane Database of Systematic Reviews, Social Sciences Abstracts, LISA, Philosopher’s Index, and ERIC. We will also perform targeted searches for grey literature (i.e., difficult to locate or unpublished material) by searching 1) Google, 2) relevant discipline-based listservs (e.g., CANMEDLIB, MEDLIB), and 3) the websites of agencies that fund or conduct knowledge synthesis (e.g., CIHR, Canadian Agency for Drugs and Technologies in Health, Agency for Healthcare Research and Quality, Cochrane and Campbell Collaborations, Joanna Briggs Institute, Centre for Reviews and Dissemination).

The draft literature search for MEDLINE can be found in Additional file 1, which uses a combination of medical sub-headings (MeSH) and free text terms. It will be modified as necessary for the other databases. The search strategy will not be limited by study design, year or language of dissemination and will be peer reviewed by another information specialist using the Peer Review of Electronic Search Strategies (PRESS) checklist [49]. The literature search will be supplemented by scanning the reference lists of included studies, searching authors’ personal files, and contacting methodological experts in each field.

Study selection: inclusion criteria

Study design: All study designs will be considered including qualitative and quantitative methods such as methodology reports; knowledge syntheses (including a description of the synthesis method); short reports describing the development, use, or comparison of methods for knowledge synthesis. Type of knowledge synthesis: We will focus on synthesis methods above and beyond traditional systematic reviews and exclude methods on economic analysis or clinical practice guidelines. Disciplines: Health: “A state of complete physical, mental and social well-being and not merely the absence of disease or infirmity” [50] (and thus includes the disciplines of psychology, education and sociology) and philosophy. These were selected because many of the knowledge synthesis methods originated from these disciplines (e.g., systematic review methods rooted in education and psychology; realist reviews based on philosophy).

Study selection: screening

Prior to commencing the screening process, a calibration exercise will be conducted to ensure reliability in correctly selecting articles for inclusion. It will entail independently screening a random sample of 5% of the included citations by two reviewers. Eligibility criteria will be modified if low agreement is observed between the reviewers (e.g., a kappa statistic less than 50%). The reviewers will then independently screen the remainder of the search results using a pre-defined relevance criteria form for all levels of screening (e.g., title and abstract, full-text review). Discrepancies will be resolved by discussion with a third reviewer.

Data abstraction

A data abstraction form will be tested independently by two reviewers on a random sample of 10 articles and revised iteratively, as needed. It is anticipated that the data items will include study characteristics (e.g., first author, year of publication) and characteristics related to the method (e.g., general description of the review method, discipline) (Additional file 1). Two investigators will independently read each article and extract the relevant data. Differences in abstraction will be resolved by discussion or the involvement of a third reviewer. We will not formally appraise methodological quality because the aim of a scoping review is to identify gaps in the evidence base and to target topic areas for future reviews.

Data analysis

We will analyze the data according to a three-stage process aimed at addressing the three research objectives: to characterize the synthesis methodologies (Synthesis objective 1); to identify the similarities and differences amongst these methods (Synthesis objective 2); and to map out a process for conducting different synthesis methods and to provide an approach for matching the research question to the appropriate methods (Synthesis objective 3). Table 2 shows the analysis plan and anticipated outputs for each of these objectives. Data analysis will involve quantitative (e.g., frequency analysis) and qualitative (e.g., thematic analysis) methods. We anticipate that this multi-layer synthesis process will also identify existing gaps in the literature, and reveal potential topics for conducting other systematic or novel reviews in the future.

Table 2 Analysis plan and anticipated outputs for each of the 3 synthesis objectives
Full size table

Engagement of knowledge users and KT plan

We have adopted an integrated KT approach to this project through the inclusion of knowledge users (i.e., systematic review methodologists, journal editors, review funders, policy makers, students and educators who teach knowledge synthesis methodology), who have been and will continue to be involved in every step of the process through to the reporting format and the methods for disseminating and implementing findings, drawing on Graham’s Knowledge-to-Action (KTA) framework [51]. We plan to develop an active KT plan by: 1) identifying the key messages arising from this research project; 2) determining the principal target audiences for each of these messages; 3) seeking out the most credible messenger for these messages and engaging their interest in becoming involved in the communication of these messages; and 4) launching a KT strategy grounded in the best available research evidence. We will use a diverse range of approaches to disseminate the results of this review to the different stakeholder groups (including an interactive workshop that will bring together the key target audiences for our research). These strategies will ensure that the research continues to reflect the relevant needs of the end users of this information, and to facilitate appropriate dissemination of outputs.

Anticipated challenges

We foresee some potential challenges related to this scoping review. First, the yield of the literature search might be more extensive than anticipated—the team will work closely with the information specialist to ensure that the scope is manageable. Second, it might be challenging to categorize the knowledge synthesis methods accurately (e.g., distinguishing between quantitative/qualitative or mixed / hybrid approaches or those not formally categorized) and to appropriately match a research question with a synthesis method. However, we have a strong team with diverse experience in different research methods, and are planning to hold stakeholder meetings to iteratively receive in-depth feedback from our end users.

Discussion

The proposed scoping review has the potential to impact practice and policy and will make several contributions to the KT and health services research literature. First, the work will advance the science of knowledge synthesis by providing a systematic process for key knowledge users to make informed decisions about which synthesis method is the most appropriate to answer their research questions. This may also augment the quality of the research evidence produced. In particular, the work will highlight the potential for novel knowledge synthesis methods to clarify complex, multi-component, and multi-disciplinary healthcare interventions [13], and to contribute to the advancement of evidence-based practice and evidence-based decision-making. Second, there is currently no comprehensive manual for all available synthesis methods (quantitative/qualitative or mixed). To develop this manual, a taxonomy and comparison of all available synthesis methods are needed. Our work aims to develop the taxonomy of synthesis methods across multiple disciplines such as health and philosophy. Third, the scoping review will help map the literature, identify gaps where primary methods evidence is lacking and needed, and where systematic reviews are required; we anticipate that this work will lead to multiple subsequent systematic reviews. For example, one future systematic review may focus on knowledge synthesis methods for health services research and another may focus on knowledge synthesis of qualitative data. Fourth, the work has the potential to directly influence knowledge synthesis funders such as the Canadian Institutes of Health Research (CIHR) in developing resources (e.g., modules) that can be used to increase awareness of novel synthesis methods and their relevance for addressing complex evidence. This information is especially imperative for those conducting peer review of knowledge synthesis grants. Fifth, the scoping review can be used by publishers and editors to assist with the peer review of manuscripts describing these types of knowledge syntheses. Sixth, our findings have the potential to influence health research methods curricula within clinical epidemiology programs, by expanding the current understanding of synthesis methods. The development and evaluation of complex interventions has emerged as an important component of KT, so expertise in conducting non-traditional review methods will become increasingly important for researchers, teachers, and students. Lastly, the work will be targeted across a broad scope of health disciplines, which will provide the opportunity to elicit more generalizable findings that can directly inform practice and policy decisions within these disciplines. Results from this work will be the starting point of a comprehensive manual and decision algorithm on how to conduct the different synthesis methods and the proposed KT strategy will serve to engage the relevant stakeholders in clarifying and fulfilling the research agenda proposed in the scoping review (Table 3 summarizes the anticipated products that will be generated).

Table 3 Anticipated products generated by the scoping review
Full size table

Conducting a scoping review of available knowledge synthesis methods across multi-disciplinary fields will help funders, publishers, policy-makers, researchers, teachers, and students make informed decisions about the most appropriate synthesis method to answer research questions about complex evidence, and provide the opportunity to elicit findings directly informing practice and policy decisions.

Funding

The study was funded by a Canadian Institutes of Health Research (CIHR) Knowledge Synthesis grant. MK holds a CIHR Banting Postdoctoral Fellowship, ACT a CIHR/DSEN new investigator award, and SES a Tier 1 Canada Research Chair in Knowledge Translation.