CONtextual analySis in implEmeNtation Science (CONSENS) - Development of a methodology for studying context and application in transplantation intervention research

As the implementation of interventions—whether that means procedures, programs, products, policies, pills, practices or principles (the 7 P`s)—takes place in real-world settings, its success and sustainability depend heavily on the context in which they will be delivered. Context is “a set of characteristics and circumstances that consist of active and unique factors, within which the implementation is embedded.” Context is multi-level, multi-dimensional and dynamic; it interacts with an intervention and its implementation in the “physical location in which the intervention is put into practice,” i.e., its setting. Thus, context has received significant attention in implementation science, which is “the scientific study of methods to promote the systematic uptake of research findings and other evidence-based practices into routine practice, and, hence to improve the quality and effectiveness of health services and care”. Nevertheless, context remains understudied in implementation science: many researchers fail either to consider or to report contextual information, thereby transforming most of their findings into leaks in the research pipeline, and adding them to what has become an ocean of research waste. Indeed, as few as 14% of evidence is ever implemented into real-world settings. Even then, their sustainability is often limited and the mean lag between the publication of high-level evidence and its implementation has been estimated to be 17 years. Overcoming these challenges, i.e., reducing research waste, bridging the research-to-practice gap and speeding the translational pipeline, requires a rethinking of current approaches in implementation science. Specifically, this includes recognizing context as the key determinant of implementation science projects' success and a major source of information relevant to all of its phases, i.e., intervention development/adaption, selection/adaption of implementation and sustainability strategies, and interpretation of implementation and effectiveness outcomes. In fact, the successful translation of an intervention to a real-world setting always depends on the characteristics of the context, an intervention that fits those characteristics, and which contextually adapted implementation strategies are employed to support intervention's adoption, implementation and sustainability. Therefore, a thorough contextual analysis of the multilevel system in which an intervention will be delivered is a key method of ensuring an implementation's success and sustainability. However, contextual analysis is currently a weak part in implementation science methodology. The concept of context lacks a unifying theoretical and operational definition and terms used to denote context vary across studies and frameworks. These and other conceptual inconsistencies challenge researchers' efforts to find and access implementation science-relevant content and contextual information. Additionally, methodological guidance for conducting a contextual analysis is lacking. And, reflecting a post-positivist understanding, researchers often consider individual aspects of context rather than its complex and dynamic interactions. Therefore, this dissertation's overall aim was to strengthen the theoretical and methodological foundations for contextual analysis in implementation science. It is structured in seven chapters, of which chapters 1 and 2 provide introductory information: Chapter 1 provides a general introduction on the concept of context and contextual analysis as presented by implementation science methodology. It includes current approaches for contextual analysis and their limitations. It also describes The Context and Implementation of Complex Interventions (CICI) framework, which is the basis of this dissertation's conceptualization of context. After briefly introducing key elements of implementation science it outlines this dissertation's guiding rationale and the research gap it attempts to bridge. Chapter 2 describes this dissertation's aims. Chapter 3 reports on a cross-sectional online survey of 56 international implementation science experts. Given a list of journals that publish implementation science-relevant content, and that can be included in IS-relevant search strategies, respondents were asked to specify which of the listed journals they considered relevant to their work. While considerable variability was found regarding most of these journals' relevance, 97.1% of the respondents rated two—Implementation Science and BMC Health Services Research—as relevant. They also proposed additional journals that they considered relevant to specific clinical fields and health science disciplines. Via PubMed and Google searches, we also identified 53 implementation science-focused special issues. Chapter 4 describes contextual analysis as a dedicated phase within an implementation science project—one that serves as the foundation, and that informs all subsequent phases. It also presents the Basel Approach for coNtextual ANAlysis (BANANA), a comprehensive, stepwise approach to guide contextual analysis in implementation science projects. Building on previous work by Stange and Glasgow on patient-centered medical home research, BANANA grew out of brainstorming sessions with implementation science experts and a medical anthropologist. It involves six steps: 1) choice of a theory, model or framework; 2) use of empirical evidence; 3) stakeholder involvement; 4) study design for contextual analysis; 5) determination of relevant contextual factors for implementation strategies/outcomes and intervention co-design; and 6) reporting of the contextual analysis. The first three steps partly run simultaneously, and form the informational basis for the next steps. Each step is described in detail, and a case example demonstrates a successful application of BANANA in an ongoing implementation science research project, the SteM cell transplantatIon faciLitated by eHealth (SMILe) project. Chapter 5 describes our development of an evidence gap map (EGM) that summarizes and graphically depicts the methodological approaches to contextual analysis applied in the mapped implementation intervention studies, as well as highlighting notable gaps in those approaches. As part of this work, a novel approach to literature searches and a framework for summarizing and evaluating methodological approaches for contextual analysis is provided. Our search for publications from the previous six years (2015–2020) yielded 15,286 publications. Utilizing a step-wise approach, we first selected, then screened a random sample of 3017 records–20% of the studies from each year of our search. The screening process left 110 implementation intervention studies for our analysis. Assembling the EGM, we found that only 24 of the 110 included studies (22%) reported on context. Among those reporting on contextual analyses showed we noted high variability both in methods used and in contextual factors assessed. Only one study explicitly reported the use of a theoretical framework for contextual analysis. And while several reported stakeholder involvement, their actual participation was quite limited. Also, those that included contextual analyses gave only sparse descriptions how they used the results. By depicting the publications' data graphically, the EGM literally shows their gaps and heterogeneity. To counter these shortcomings, there is clearly a need to promote standardized approaches to contextual analysis as foundational for the success and quality of implementation science projects. Chapter 6 reflects on limitations of current approaches to contextual analysis—particularly those driven by a post-positivist understanding of context—and describes how the addition of a constructivist perspective would complement these approaches. Five constructivist concepts are introduced that can contribute to a more comprehensive and multilayered understanding of context and can reveal complex dynamics not visible via the post-positivist perspective, i.e., interactions involving individuals and/or contextual factors: 1) social space; 2) social place; 3) agency; 4) sensation; and 5) embodiment. After illustrating how these concepts can be integrated into existing conceptualizations of context, this chapter uses COVID-19 vaccine hesitancy as a case example to demonstrate the value the constructivist perspective adds to contextual analysis. Further, methodological and practical considerations, e.g., regarding data collection and analysis, are discussed. Chapter 7 first discusses and synthesizes the dissertation's major findings in light of current evidence, then outlines implications for research and practice. Strengths and limitations regarding the methods and the applied framework are indicated. Further considerations regarding methodological approaches and designs that help address the complexity of context and to accelerate contextual analysis are described. Finally, a push-pull-capacity model is used to indicate local-, regional- and national-level factors that support capacity-building both for contextual analysis and for implementation science in ways that improve the translational pipeline's performance. This dissertation`s contribution is threefold. First, it presents a case for applying contextual analysis as a separate, foundational implementation science project phase, the results of which inform all subsequent phases. While conducting a contextual analysis requires additional time and financial investments, the results will enhance the quality and speed of intervention translation into real-world settings, thereby enhancing its societal impact. Second, we describe a methodology to guide researchers to conduct a contextual analysis (BANANA). We illustrated its application to an ongoing implementation project (the SMILe project). Third, we suggest ways that contextual analysis can be enhanced–particularly to overcome the shortcomings of the current positivist perspective. Overall, this dissertation both provides a foundation to support enhanced uses of contextual analysis and strengthens implementation science methodology. Ultimately, we are confident that these developments will improve the overall quality and success of implementation science projects