Perspectives of health care professionals on the facilitators and barriers to the implementation of a stroke rehabilitation guidelines cluster randomized controlled trial

The Clinical Practice Guidelines Framework for Improvement [19] and its updates [20‐22] were used to guide the coding framework in the current study. Legaré and colleagues [21] formulated a definition for each type of barrier to promote the standardization in the reporting of barriers and facilitators across different studies. Examples of barriers and facilitators include knowledge (awareness, familiarity), attitudes (e.g., agreement with intervention), and behavior, including environmental factors [20‐22]. The use of the Clinical Practice Guidelines Framework for Improvement [19] was not decided a priori. We decided to use the Clinical Practice Guidelines Framework for Improvement [19] and its updates [20‐22] as it is one of the most recognized frameworks for assessing barriers and facilitators and after analyzing several focus groups and determining that the emerging categories clearly aligned with the barriers outlined in the framework.

This study took a qualitative descriptive approach that consisted of telephone focus groups. A qualitative descriptive approach is well-accepted for researching topics about which little is known and yielding practical answers of relevance to policy makers and health care practitioners [23, 24]. Telephone focus groups were selected because of the geographic dispersion (i.e., national scale) of the study participants.

Participants included staff members from the three different professional groups, nurses, therapists (OTs and PTs), and health care/clinical managers, who had participated in the SCORE-IT and agreed to be contacted at the conclusion of the trial. Participants were contacted by telephone and email about their willingness to participate in the focus groups. Purposive sampling [25] was used to recruit equal numbers of participants across professional groups (nurse, therapist, clinical manager), randomization arms (facilitated KT intervention or passive KT intervention), and geographic locations (Western, Central, Quebec, Eastern). Uniprofessional focus groups were conducted for nurses, therapists, and clinical managers at each of the participating sites. This approach was adopted to mitigate potential power imbalances that may have influenced what participants might be willing to share. Participants were recruited between January 2009 and March 2010. Recruitment ceased when a discussion and review of the responses revealed that saturation had been achieved (i.e., no new responses or themes were emerging) [26].

Each participant took part in a semi-structured telephone focus group lasting approximately 45–60 min. The principal investigators (MB, SWD) and the research coordinators involved in the trial conducted the focus groups. The focus group guide consisted of semi-structured open-ended questions and was informed by the results of our pilot project [11]. The interview guide was pilot tested with a researcher experienced in qualitative methods. Probes or recursive questioning were used during the focus groups to explore issues in greater depth and to verify understanding of the information being collected [25]. The probes were revised and refined as data collection progressed to establish saturation [26]. The complete list of questions is included in Additional file 1. No repeat interviews were conducted. All focus groups were audio recorded. Field notes were made during and/or after the focus groups. The recordings were transcribed verbatim for data analysis. These transcripts were not returned to participants for comment and/or correction.

To facilitate the organization and analysis of the qualitative data, the transcripts were entered into NVivo 10 [27]. Thematic analysis as described by Braun and Clark [28] was used to understand the factors influencing the implementation of the recommended treatments and KT interventions by study arms. The lead author (SM) reviewed the transcripts to develop an initial codebook based on the Clinical Practice Guidelines Framework for Improvement. Following this, two researchers (SM, MB) independently coded a sample of the transcripts (20%), revised the codebook as themes emerged, and met to discuss and reconcile discrepancies until agreement of the coded transcripts was reached. SM is a female post-doctoral fellow and has a PhD in Health Services Research as well as expertise in knowledge translation. She has approximately 10 years of experience conducting qualitative research. MB is a physiatrist (i.e., MD) with expertise in stroke, brain injury, rehabilitation, clinical practice guidelines, prognostic factors, and health services research. He has approximately 10 years of experience conducting qualitative research. Our background in knowledge translation science has influenced the conceptual frameworks that we have been exposed to including our knowledge and selection of the Clinical Practice Guidelines Framework for Improvement [19] for this study. Disagreements/discrepancies around codes, themes, and subthemes were resolved by discussion and reference to the original transcripts. The lead author (SM) analyzed the remaining transcripts. Relevant quotations were identified and selected from the transcripts to illustrate the themes and include the participant’s professional group (nurse, therapist, or clinical manager), and randomization arm (facilitated KT intervention or passive KT intervention). Participants were not provided feedback on the findings.

Research ethics approval was obtained from each site and affiliated university. All participants provided written consent prior to the interview. The trial was registered at ClinicalTrials.​gov (NCT00359593).

Focus groups were conducted with 33 individuals including 11 nurses, 11 therapists, and 11 clinical managers. There were between two and six participants in attendance at each focus group. Participants were from 11 of 20 sites in Western, Central and Eastern Canada as well as Quebec. This sample represented 6 sites from the facilitated KT arm and 5 sites from the passive KT arm.

Overall, five themes were identified. The following themes influencing the implementation of the recommended treatments and KT interventions emerged: facilitation, agreement with the intervention – practical, familiarity with the recommended treatments, and environmental factors (including time pressure, insufficient staff, lack of space and equipment, and organizational constraints). These themes, for the most part, emerged as facilitators and barriers influencing the implementation of the SCORE-IT. Furthermore, the theme of improved team communication and interdisciplinary collaboration emerged as an unintended outcome of the trial across both arms in addition to facilitating the implementation of the treatment recommendations. Representative quotes are given in Tables 2 and 3.

The objective of the current study was to understand the facilitators and barriers influencing the implementation of the recommended treatments and KT interventions from the perspective of nurses, OTs and PTs, and clinical managers following completion of the SCORE-IT. This is one of the first studies to examine the factors influencing the implementation of evidence-informed stroke recommendations and associated KT interventions among allied health care professionals within the context of a trial.

All of the factors influencing the implementation of the recommended treatments and KT interventions including facilitation, agreement with the intervention – practical, familiarity with the recommended treatments, environmental factors, and team communication and interdisciplinary collaboration were identified in both arms of the trial. Team communication and interdisciplinary collaboration also emerged as an unintended outcome of the trial in both arms of the trial. It is particularly noteworthy that facilitation was identified as a facilitator to implementation in the passive KT intervention arm despite the lack of formally instituted facilitators in this arm of the trial. The order of the remainder of this Discussion section is the same as the Results section. As in the Results section, where applicable, we have used the same terms from the Clinical Practice Guidelines Framework for Improvement [19] as headings to organize our Discussion section.

The presence or absence of facilitation in this trial emerged as both a facilitator and barrier to implementation of the recommendations and KT interventions. Indeed, the results of a systematic review on local opinion leaders and their effects on professional practices revealed that opinion leaders alone or in combination with other interventions may successfully promote evidence-based practice [29]. In fact, research on barriers to research use in health care have consistently identified the behaviors of managers and their lack of leadership as major limiting factors to research use by clinicians [30‐34].

One of the main findings of this study was that the theme of facilitation was noted in both arms of the trial, despite the fact that only the facilitated KT arm had formal facilitators. Previous research has raised the question of whether the process by which opinion leaders are selected affects the success of educational initiatives [29]. If self-selected facilitators (i.e., in this case, staff at the passive KT sites) are more beneficial (to implementation outcomes) than facilitators who are selected by external influences, it is possible that other components of the facilitated KT arm of the trial may not have been optimized. Alternatively, the self-initiated facilitation roles taken by staff members at the rehabilitation centers in the passive KT arm may explain why the outcomes at these centers were better than anticipated. Indeed, the results of the trial revealed that while the facilitated KT intervention was associated with a significantly greater improvement in the rate of implementing sit-to-stand training and walking practice, the passive KT intervention was associated with significantly greater improvement in the rate of implementing standing balance training (after adjusting for clustering at patient and provider levels and covariates) (i.e., between group differences) (unpublished work). It should be noted that the original trial was dealt with as a pragmatic trial, which tries to mimic the usual care situation and not impose too many fidelity standards on the basis that they produce a trial result which is not applicable/externally valid for the use of the same intervention under usual care conditions. Furthermore, as Horne [35] noted, staff can be trained to be good managers, but leadership is less susceptible to training and is better obtained by selective recruitment. This phenomenon may explain why facilitation (including a lack of facilitation) was noted across both arms of the trial. At the same time, we are not linking facilitation behaviour to the actual use of the recommendations, rather, we are presenting perceptions of what may or may not have occurred (i.e., in the control arm, it appears that individuals in some sites stepped up to try to mobilize and encourage the uptake of recommendations and the KT interventions themselves). Future research should seek to better understand the specific characteristics/behaviours of facilitators that are associated with successful implementation and clinical outcomes, especially within the context of stroke rehabilitation.

Practicality of and familiarity with the recommended treatments and KT interventions also emerged as significant facilitators and barriers. For example, participants indicated that certain recommendations/tests were not practical to implement because they were time-consuming (e.g., FES unit). Indeed, it is likely that a variety of facilitators and barriers acted together and in combination to influence the implementation of the interventions in the trial (e.g., interaction of time and practicality). Similarly, a lack of familiarity with certain components of the recommended treatments, including equipment such as the FES apparatus and measures such as the CAHAI, limited the implementation efforts. Indeed, results from the SCORE-IT indicated that complex treatments that either involved multiple steps or technology, including the FES, were rarely implemented at baseline and demonstrated either no change or reduced application post-intervention (unpublished work). This finding suggests that the KT interventions did not adequately overcome these barriers. It is possible that these barriers cannot be overcome with KT interventions, especially within the context of a trial (i.e., lack of familiarity with a recommendation(s) and its implementation can only be overcome with a significant amount of time); however, it is possible that facilitation (i.e., mentorship) could be leveraged to overcome barriers associated with practicality and/or familiarity. Furthermore, previous research has reported that insufficient skills and a lack of experience with guideline recommendations are key barriers to implementation of best practices [6, 11, 36‐38]. It could be that a mid-point check of progress and renewed goal setting might be helpful to address these barriers.

Environmental factors, including time pressure, insufficient staff (lack of staff, staff turnover), lack of space and equipment, and organizational constraints (insufficient support from the organizational/senior management) emerged as the most frequently cited barriers to implementation of the KT interventions during the trial as well as the recommendations. In a recent study describing the factors influencing the implementation of stroke clinical practice guidelines among speech pathologists, Hadely and colleagues [36] also reported that factors within the work environment were barriers to implementation. Specifically, the main barriers included lack of time, education, treatment resources, and standardized assessments to carry out guideline implementation [36]. Environmental/work factors as barriers (and facilitators) to guideline implementation have been reported consistently in literature – in the treatment of persons post-stroke as well as other chronic conditions [6, 11, 37, 39, 40]. In the current study, one of the main findings was that environmental factors were seldom noted as facilitators to the implementation of the recommended treatments and/or KT interventions. It should also be highlighted that some of these environmental factors were mitigated by team factors/facilitation. Thus, a main message from our research is that in the absence of more organizational resources (time, money), team factors can be leveraged to overcome such deficits. For example, a high level of staff turnover was noted across the rehabilitation centers; however, if strong leadership/management support was present at the rehabilitation center, this person often ensured that new staff knew the procedures/responsibilities associated with the trial. Horne [35] similarly noted that leaders and even the larger hospital administrative culture could act as key mediators between the environmental factors (time, money, equipment) and the implementation of recommendations.

The presence of team factors, including communication and interdisciplinary collaboration, served as a facilitator to the implementation of the recommended treatments. Donnellan and colleagues [17] also determined that barriers to adherence to generic stroke guidelines related to organization and multidisciplinary team factors [41‐43]. Similarly, team factors played a significant role in influencing the implementation of stroke clinical practice guidelines in the study by Hadely and colleagues [36]. For example, they determined that working in a multi-disciplinary team emerged as a main factor for facilitating the use of guidelines among speech pathologists. These factors have also been reported among physicians, nurses, and OTs [11, 40, 44‐47]. Hadely and colleagues [36] concluded that fostering teamwork can have a significant influence not only in improving guideline implementation but also patient functional gains [48] and length of hospital stay [49].

Findings from the current study also suggest that the Clinical Practice Guidelines Framework for Improvement [19] is relevant in the context of implementing recommended treatments and KT interventions in stroke rehabilitation as agreement with the intervention – practical, familiarity (with the recommended treatments), and three aspects of environmental factors were identified as factors influencing implementation. The other identified factors of facilitation and team communication and interdisciplinary collaboration are not included in the Clinical Practice Guidelines Framework for Improvement [19] and its updates [20‐22] but are included in other implementation frameworks, namely the Consolidated Framework for Implementation Research (i.e., formally appointed internal implementation leaders versus champions, networks and communications) [50] and the Promoting Action on Research Implementation in Health Services (PARIHS) (i.e., facilitation) [51]. Future iterations of the Clinical Practice Guidelines Framework for Improvement [19] could consider these factors, which may improve its ability to address common facilitators and barriers in this context. Lastly, another area of future research would be determining the perceived relative importance of these identified facilitators and barriers (e.g., using a modified Delphi process).

It is noteworthy that many of the identified factors influencing the implementation of the recommended treatments and KT interventions were also identified in our previous multi-site pilot project on the barriers to the implementation of evidence-based recommendation for stroke rehabilitation (i.e., lack of time, inadequate staffing, and equipment). As previously identified, many of these environmental barriers are difficult to overcome and beyond the control of the trial implementation effort. We also previously noted that leaders at the organizational level may be required to overcome these issues; however, in the current study, organizational constraint (insufficient support from the organizational/senior management) was a noted barrier across both arms of the trial. Thus, we may not have adequately addressed these barriers (and the interrelated nature of these barriers) in the current trial. At the same, team functioning and communication was previously noted in the pilot study but was identified as both a facilitator and unintended benefit in the current study.

We acknowledge some limitations. Only one person coded the majority of the data, which may have resulted in bias. The persons conducting the original study also conducted the focus groups, which presents a significant concern about the social desirability of participant responses. At the same time, however, the focus group leaders were careful to avoid biasing the participants towards or away from the intervention and were themselves neutral on its effectiveness (i.e., clinical equipoise). Furthermore, only 11 of the 20 sites participated (but almost equal representation from the facilitated KT and passive KT arms), and thus it is possible that a selection bias operated in that those participants who agreed to take part in this study may have had a greater interest and success with implementing evidence-based recommendations for stroke than those individuals who chose not to participate (i.e., limiting the applicability of the study findings). In discussing factors influencing the implementation of the KT interventions following completion of the trial, participants may have had recall bias; in a focus group setting, participants may also have felt limited in their ability to share their experiences due to social desirability issues. Organizing the focus groups by professional group was an attempt to mitigate this potential barrier. Furthermore, the focus group questions did not specifically ask about all of the factors influencing the 18 recommended treatments of interest (e.g., training for sitting balance, training for standing balance) or did not consistently ask about each of the KT interventions. As such, we are only able to obtain a global sense of the factors influencing implementation of the recommended treatments and their associated KT interventions. This may mask specific issues for specific interventions. The focus group questions were not anchored on the actual performance of the rehabilitation centres; more specific knowledge about the facilitators and barriers to implementation would have been obtained if this approach has been adopted (as discussed above in the Role of Facilitation section). Lastly, the trial and the subsequent focus groups were conducted a number of years ago; it is unknown how a more recent implementation of stroke recommendations and interventions to increase their uptake would affect the current results (e.g., with health system advances such as electronic medical records with reminders).