Organizational contextual features that influence the implementation of evidence-based practices across healthcare settings: a systematic integrative review

While these implementation frameworks include context, no single framework is sufficiently comprehensive about what comprises context. In addition, the authors of the frameworks are often inconsistent in how context is theoretically and operationally defined. Without a shared understanding of context and its characteristics and features, there is little direction to which features of context are most influential to KT efforts [22]. Extending beyond conceptual theories, models, and frameworks; this review aims to synthesize and summarize organizational contextual features commonly reported to influence the implementation of EBPs in actual healthcare settings.

The search strategy (see Additional file 1) was developed on MEDLINE in consultation with two librarians and applied to Cochrane databases, CINAHL, MEDLINE, EMBASE, PsycINFO, and gray literature. The journal Implementation Science (from journal inception to June 2017) was hand searched to uncover additional relevant articles. The search included four categories of search key terms: (a) Implementation, (b) Context, (c) Evidence, and (d) Organization. Inclusion and exclusion criteria (Table 1) were applied during screening.

Following the removal of duplicates, two reviewers (SAL, PEA) independently double-screened the titles and abstracts of a randomly selected sample (generated from an online randomized website) of 20% (n = 1034) of the retrieved citations to ensure interrater reliability. Once suitable agreement (k = 0.85) was achieved, title and abstract screening was undertaken for all citations. Citations missing an abstract during screening were retained for full-text screening to establish eligibility. Citations meeting eligibility criteria were included for full-text screening. Full-text screening followed the same strategy as the title and abstract screening to ensure interrater reliability.

Data extraction was performed by the same reviewers (SAL, PEA) for all included articles, independently and in duplicate. A third reviewer was available to resolve any disagreement between the two reviewers; however, all disagreements were resolved via consensus without involving a third reviewer.

The Mixed Methods Appraisal Tool (MMAT [25]) guided the quality appraisal for all qualitative, quantitative, and mixed methods studies. Quantitative and qualitative studies were each assessed by four criteria with overall scores varying from 0% (no criterion met) to 100% (all four criteria met). For mixed methods studies, three components were appraised: qualitative, quantitative, and mixed methods component, with the overall score determined by the lowest component score. In keeping with integrative review methods [24, 26], all records were retained in the analysis, regardless of score. For each article, two reviewers assessed methodological quality independently and discrepancies were resolved via consensus.

For each study, the steps of data reduction, data display, data comparison, and drawing conclusions and verifications were followed [24]. To ensure trustworthiness and rigor during data abstraction and synthesis, a table was developed to summarize the organizational contextual features. The abstracted information was compared, and patterns of findings were recorded as they emerged, followed by groupings of similar data and the identification of several key themes [24].

Analysis followed a qualitative descriptive approach, given that most of the study designs were qualitative and the results from quantitative studies could not be aggregated due to heterogeneity in study design, outcomes, and type of intervention [27]. Inductive content analysis was used to uncover themes related to organizational context [28, 29]. This analytic approach involved reading and rereading the articles to uncover any salient codes and categories, subsequently collapsing them into themes (organizational contextual features) [27].

Table 2 presents the characteristics of the included studies, including study setting, study aim, sample, guiding framework (if applicable), study design and data collection methods, main findings, and MMAT quality score. The studies were published between 2007 and 2017 and were based in 11 countries. At least 8094 participants were included in this review. Study participants comprised a wide range of stakeholders including physicians, nurses, and allied health professionals working as coordinators, medical staff, and senior managers from many different healthcare settings. Two reported on the number of participating pediatric hospital units (n = 16) and medical centers (n = 12) instead of the number of individual participants. Twenty (56%) of 36 studies used a theory, model, or framework to guide their data collection and/or analysis.

The included studies were of moderate to high methodological quality (Table 1) based on the MMAT [25] appraisal. Of the 36 studies, 22 received a score of 75% (moderately high quality), 8 received 50% (moderate quality), and 5 received 100% (high quality).

Six organizational contextual features included organizational culture; networks and communication; leadership; resources; evaluation, monitoring, and feedback; and champions. A series of sub-features included collaboration, teamwork, communication, financial resources, time, staffing and workload, and education and training. Table 3 illustrates the features and sub-features.

Organizational culture was included as an organizational contextual feature in 22 of 36 (61%) studies. Organizational openness to trialing new innovations and a learning culture were highly associated with implementation success [30‐35]. Conversely, an absence of a learning culture can act as a major hindrance to successful implementation [36]. Organizational cultures comprised of staff who have too much autonomy (i.e., physicians experiencing a high level of autonomy when making decisions about how to treat patients) [37], are resistant to trial new innovations [38], or are unclear about organizational values and beliefs [39] can be barriers to successful implementation. Sites demonstrating high implementation fidelity were marked by a strong culture of innovation, accompanied by positive staff attitudes and behavior toward the new initiative [40]. In one study, unit culture (measured by work creativity, work efficacy, questioning behavior, co-worker support, and emphasis on continuing education) was a significant predictor of nurses’ research use [41]. These results corroborated with other studies investigating organizational context and nurses’ research utilization [32, 42]. Organizational culture significantly moderated the effect of nurses’ instrumental (direct use of research knowledge) and conceptual research use (indirect use of research knowledge) on pediatric pain assessment in hospitals [43].

An association of organizational networks and communication with implementation success was evident in 22 of 36 studies (61%). Three sub-features were commonly associated with implementation outcomes.

Collaborative relationships that occur within and external to the organization were important for carrying out implementation plans. For instance, Barnett et al. highlighted two main purposes of interorganizational collaborations. First, materially based partnerships provided the organization with the resources required for the implementation and diffusion of new programs. Second, symbolically based interorganizational collaborations allowed organizations to gain local consensus to bolster the new programs with legitimacy, which in turn serves as an important social exchange that assisted with communicating the innovation’s impact through gaining a broader consensus. Harvey et al. [39] described how close collaboration with an external implementation improvement team can support staff and leadership development geared toward implementing change.

Teamwork was characterized as good working relations, the ability to communicate clearly and effectively, and the capacity to solve problems together during EBP uptake [44‐47]. Using an ethnographic case study design, McCullough et al. [47] observed that strong teamwork among staff, when combined with strong belief in evidence, led to high adoption of a dosing algorithm in anticoagulation clinics. However, when staff were dismissive of the evidence, strong teamwork served to reinforce resistance to implementation efforts. Teamwork was highly relevant in new programs that required participation from professionals in multidisciplinary teams. In a mental health organization where multidisciplinary staff (peer specialists, practitioners) were required to implement a new person-centered recovery program for their patients, poor multidisciplinary teamwork resulted in poor program uptake [38].

Communication greatly impacted the implementation of EBPs [37, 39, 44, 46, 48‐50]. The establishment of systems and processes to more effectively manage information and communication about the change initiative influenced implementation success [39]. Communication between healthcare professionals in a Dutch intensive care unit (ICU) was an important barrier for a successful start of the implementation phase of a delirium scoring system [48]. Vamos et al. [50] and Stevens et al. [49] articulated various communication channels that facilitated implementation in hospital units, including active (scheduled meetings, debriefings, emails) and passive (flyers, announcements on bulletin boards, auto-generated reminders) communications.

Leadership was reported in 20 of 36 (56%) studies as an important feature for implementation effectiveness. Leaders were often seen as providers of new knowledge and as key influencers in new implementation initiatives [44]. Leaders who created environments with high staff morale allowed staff to perceive themselves as part of the implementation team. Transformational leadership often gave rise to clear roles and effective teamwork structures and cultivated a culture of learning [44]. Leaders ensured changes were sustained, without which staff were reported to “fall back into the old ways of doing things” [37]. Senior leaders were important for ensuring that new processes were integrated as “business as usual” [33]. Senior leaders were also essential for overall hospital staff involvement and buy-in [33, 37, 50]. The initial decision to begin an implementation effort within the hospital and the subsequent ongoing changes during the implementation process required the engagement of leadership at different levels and from multiple stakeholders across hospital departments [50]. The willingness of middle managers to partake in the implementation process was often contingent upon the support expressed by senior leaders [31]. The absence of senior leader support or tension between middle managers and their direct supervisors meant middle managers were significantly more reluctant to participate. Leaders work to optimize implementation success by expressing enthusiasm for the change; being present, supportive, and attentive to the implementation process; and demonstrating willingness to ask for feedback from staff regarding the change. Leadership that is lacking in authority and unsupportive of change, or that neglects to hold staff accountable for the change, presents barriers to implementation [38, 51]. Staff feel unmotivated to change when leaders were too controlling or unresponsive to requests for more training by staff who were required to implement the practice change [40].

This feature is divided into four sub-features that are interrelated and appear to work synergistically to hinder or promote the implementation process.

This feature was important for successful implementation in 14 of 36 studies (39%). Appropriate feedback mechanisms benefited EBP implementation by preserving engagement among staff who implemented the change. Active and engaged leaders who sought feedback about the change and who provided feedback to staff were associated with higher rates of implementation success [33, 34, 53]. Soliciting early feedback from middle managers can help assuage their concerns about the change initiative, and ongoing staff communication and monitoring increased the likelihood of EBP sustainability over time [31]. Three studies reported that evaluation and feedback were important predictors of research use among nurses [32, 54, 60]. Evaluation moderated the effect of nurses’ use of research for pain management [43]. Audit and feedback were effective for improving nursing practice in pain management and assessment for children [49].

Presence of champions was important for implementation success in 11 of 36 studies (31%). Champion was the strongest and most consistent feature related to the use of a new systematic framework for prevention delivery services (including the use of implementation guidelines) [45]. Having a champion to advocate for the “new way of doing things” led to more complete and refined use of these guidelines. Supporting champions can be difficult in contexts that lack engaged leadership or dedicated resources to encourage and monitor adherence. Identified champions who rise to the challenge may succumb to feelings of frustrations when the organization does not support change. Key attributes of successful champions included the following: (a) being an expert on the EBP, (b) being available for troubleshooting and for training other staff “on the floor,” and (c) providing a sense of familiarity among colleagues and belief in the champion’s expertise. In one study, the management team chose staff members who were initially unsupportive of the implementation project and motivated them to take an active role in the project, which prevented them from thwarting the implementation progress [53].

Potential interrelationships between features were identified in 12 of 36 studies. Figure 2 illustrates the direction of influence between each feature. Leadership influenced all other features of this review: (a) the use and selection of champions [33, 37, 40, 50], (b) the allocation of resources (funds and additional staffing) to accommodate for the implementation [36‐38, 52, 53], (c) the facilitation or hindrance to the monitoring and feedback mechanisms during the implementation process [52, 53], and (d) organizational culture [34, 39, 50, 52].

Networks and communication worked synergistically with other features to promote or hinder the implementation process. For instance, an organizational culture of innovation can be cultivated by ongoing and explicit communication of new innovations [44]. Leaders can help champions communicate observable impacts of the new initiative to other staff, as observed in a study investigating the adoption of clinical practice guidelines in long-term care homes [37]. Similarly, communicating with middle managers and senior leadership to gain buy-in for an EBP was a significant contributor to implementation success [31, 39].

Organizational culture interacted with other organizational contextual features. Bergström et al. [44] found that the organizational culture had set the stage for supportive, inter-professional teamwork and was more important than training staff about implementing the EBP. Sommerbakk et al. [53] demonstrated that a culture characterized by trust and open communication was a facilitator for EBP uptake. Chuang et al. [31] reported that a culture of learning builds teamwork and contributes positively to the performance of the hospital unit that is implementing change. On the contrary, lack of support from colleagues was a barrier to constructing a change culture [53]. Strong leadership, coupled with a culture of learning or openness to innovation, was important to successful implementation [31, 32, 34, 35, 37, 40, 41, 43, 52].

Leadership, culture, resources, and networks and communication contributed to implementation success in at least 50% of the 36 selected studies; among these, 12 studies (33%) identified at least one feature or a sub-feature that influenced or worked synergistically to either act as an enabler or a constraint to the implementation process.

This review was limited to published journal articles in English; the results may have limited transferability to non-English-speaking nations that have very different healthcare systems. This review was also limited to studies that investigated organizational contextual features during the implementation, adoption, and uptake phases of EBPs; these studies provided little understanding of how organizational contextual features impact the sustainability of EBPs. The search strategy of this review used the term “context” in the organization to identify empirical studies that investigated organizational context. However, it is likely that other researchers who examined the same organizational contextual features identified in this review may not use the term “context” in their report. As such, these studies could not be retrieved. For example, Williams et al. [73] reported an increase in EBP uptake through improved organizational culture among mental health clinicians in 14 children’s mental health agencies. This study was not captured in this review but proves to be highly relevant to inform implementation researchers about the value of organizational culture change on implementation effectiveness. Guerrero et al. [74] observed that the leader’s openness to and expectations about implementing EBPs were strongly associated with the implementation of a contingency management strategy in substance abuse treatment programs. Therefore, readers should approach the review findings with caution, bearing in mind the limitations of the search strategy in this review.

Several limitations at the level of individual studies warrant discussion. This review identified potential interrelationships between the organizational contextual features but did not explore the nature of these relationships, with one exception [46]. Study findings were reported very briefly in the “Results” sections, which precluded reviewers from drawing further conclusions about these interrelationships. The extent to which these features may be more effective for implementation if considered in concert or individually remains an empirical question that needs further exploration. The organization contextual features identified as consistently influential to implementation efforts were contingent upon the study authors’ decisions as to which features belonged at the organizational level. It is possible that other less frequently explored contextual features can also influence implementation outcomes.

Although the term “implementation success” frequently appeared in the “Results” and “Discussion” sections of the included studies, this term was not defined. Implementation success can be measured or conceptualized differently in different healthcare settings. Implementation studies should describe how “implementation success” is conceptualized or operationalized in the implementation project, or report on any pre-determined targets that represent implementation effectiveness. Most of the included studies used qualitative approaches to identify, describe, or explain the organizational contextual features that emerged from this review; however, it was unclear whether the conceptual or operational definitions for each of these features (e.g., culture, leadership) were consistent across the included studies. Defining each feature being investigated will enhance the clarity and consistency of the feature and facilitate external validity.

Even though 27 out of 36 included studies were rated as moderately high to high quality according to the MMAT, the included studies did not follow any standard methods of reporting, which is consistent with existing literature that articulated the low reporting standards of implementation studies [74, 75]. Implementation researchers should consider using the Standards for Reporting Implementation Studies (StaRI) [76] to ensure transparent and accurate reporting of implementation studies. StaRI requires researchers to provide an extensive description of context, which will help readers assess the external validity of the reported study, and decide how the implementation context in the study compares to their own setting. A rich description of the study’s implementation context is crucial to readers who are considering whether the implementation strategy can be directly adopted or will need modifications [77].