Defining and assessing context in healthcare implementation studies: a systematic review

Abstract

Background: The implementation of evidence-based healthcare interventions is challenging, with a 17-year gap identified between the generation of evidence and its implementation in routine practice. Although contextual factors such as culture and leadership are strong influences for successful implementation, context remains poorly understood, with a lack of consensus regarding how it should be defined and captured within research. This study addresses this issue by providing insight into how context is defined and assessed within healthcare implementation science literature and develops a definition to enable effective measurement of context. Methods: Medline, PsychInfo, CINAHL and EMBASE were searched. Articles were included if studies were empirical and evaluated context during the implementation of a healthcare initiative. These English language articles were published in the previous 10 years and included a definition and assessment of context. Results were synthesised using a narrative approach. Results: Three thousand and twenty-one search records were obtained of which 64 met the eligibility criteria and were included in the review. Studies used a variety of definitions in terms of the level of detail and explanation provided. Some listed contextual factors (n = 19) while others documented sub-elements of a framework that included context (n = 19). The remaining studies provide a rich definition of general context (n = 11) or aspects of context (n = 15). The Alberta Context Tool was the most frequently used quantitative measure (n = 4), while qualitative papers used a range of frameworks to evaluate context. Mixed methods studies used diverse approaches; some used frameworks to inform the methods chosen while others used quantitative measures to inform qualitative data collection. Most studies (n = 50) applied the chosen measure to all aspects of study design with a majority analysing context at an individual level (n = 29). Conclusions: This review highlighted inconsistencies in defining and measuring context which emphasised the need to develop an operational definition. By providing this consensus, improvements in implementation processes may result, as a common understanding will help researchers to appropriately account for context in research.

Development of an integrative coding framework for evaluating context within implementation science

“…However, this case study illustrates the benefit of using the context coding framework as it offers a means for exploring the active role of context and the dynamism of the concept "in a world on the move" [51]. It also responds to previous literature [17][18][19][20], by providing a standardised approach to improve the reporting of context during implementation. By accounting for the complexity of context, Tsoukas [52] suggests that the researcher is providing a meaningful contribution as organisational realities are exposed rather than constrained by a simplified, limited representation.…”

“…Using a constructivist approach, the context coding framework builds on the CFIR to enable a deeper exploration of contextual factors independent of the intervention. By grounding the analysis across system levels, the framework accounts for team-level contextual factors previously overlooked in implementation science [20] and rearranges CFIR determinants to the appropriate system level (individual, team, organisational, and system levels) ( Table 2). Figure 1 outlines the process which identified contextual factors relating to these system levels.…”

“…Next, the extant literature was appraised to collate contextual determinants not explicitly referenced by the CFIR [20,[29][30][31][32][33][34][35][36]. Rogers et al's [20] recent review provided a diverse range of novel constructs including team-level contextual determinants (Table 2). Finally, the contextual determinants of CFIR were integrated with the findings from the literature to refine the established CFIR constructs.…”

“…This poor understanding has been attributed to the variability in how context is defined and assessed across studies [6,[15][16][17][18][19][20]. However, Rogers et al's [20] systematic review confirms that inconsistencies exist in the extant literature when defining context. This review also identifies commonalities across the papers included to develop a broad operational definition for the construct.…”

“…To address this gap, a method to adequately account for context within implementation research is necessary This is echoed by Murdoch [22] and Fernandez et al [23] who argue that understanding the dynamics of context requires its conceptualisation to be translated into a practical method of assessment. However, Rogers et al's [20] systematic review observed considerable heterogeneity in the approaches used to assess context. Over 40 methods were identified within the 64 papers included which led to the recommendation that a standardised approach for assessing context is required integrating a qualitative method informed by a comprehensive framework.…”