Various theories, models, and frameworks have described the factors, processes, and strategies influencing the successful implementation of evidence-based interventions (EBIs) (1). Facilitation is an overarching strategy suggested to support the implementation of EBIs and constitutes the “active ingredient” in the implementation process in the integrated-Promoting Action on Implementation in Health Services (i-PARIHS) framework (2). The i-PARIHS framework proposes that facilitators play a key role in the implementation of EBIs by supporting the tailoring of implementation strategies to the characteristics of the innovation (e.g., clarity, relative advantage, and usability), its recipients (e.g., their knowledge and skills, motivation, and resources) and the inner and outer context (e.g., leadership support, culture, previous experiences with implementation, and policy) in which implementation takes place (2). A facilitator is thus a person who enables the change by easing collaboration and participatory problem-solving between stakeholders. The facilitator is uniquely positioned to help leaders and staff customize the implementation plan by choosing strategies appropriate for the organizational context and stakeholder dynamic (3–6). Facilitators can thus intentionally set cognitive and social mechanisms in motion to promote the normalization of EBIs (7). Teams and organizations where facilitation techniques are employed are more likely to make complex practice changes by, for example, implementing more EBIs (e.g., guidelines) (8), increasing adherence to them (9), or by improving the delivery of care (10).

Aside from supporting the normalization of new behaviors, facilitation has been highlighted to build internal implementation capacity in organizations (11), both through the specific roles and activities that facilitators carry out (12), and through their position, which enables them to train organization staff in implementation (6). For implementation to happen, normalization, i.e., the embedding of interventions in practice, is required. The Normalization Process Theory proposes four mechanisms that need to be triggered to achieve normalization (13). Coherence involves the work individuals and groups do when faced with the challenge of operationalizing a set of practices. Cognitive participation entails the relational work people do to build and sustain a community of practice around a new technology or complex intervention. Collective action encompasses the practical implementation of a set of practices. Reflexive monitoring is the appraisal work people do to assess and understand how a new set of practices impacts them and others. Each mechanism comprises four subconstructs (called submechanisms in this article).

The effectiveness of facilitation hinges on facilitators' knowledge, skills, and self-efficacy. Reviews of the healthcare services and implementation science literature have listed and described the type of knowledge, skills, and characteristics that facilitators should possess (14, 15). Apart from hard skills (e.g., searching, retrieving, appraising, and synthesizing evidence), facilitators should also demonstrate a wide range of soft and open skills, commonly falling within the categories of self-awareness, self-management, social and cultural awareness, and relationship management (14, 16).

How to best prepare facilitators for their role is not fully understood (17). Facilitation training can increase the chance of successful implementation and normalization of EBIs, which may lead to positive outcomes for patients in the long run (18). In contrast, training that does not match the facilitators' needs and expectations (19) or is insufficient can create barriers in the implementation process (19, 20). Although there are some interventions aiming to train professionals as facilitators (21–23), very few longitudinal evaluations of such interventions are described in the literature.

This longitudinal study aimed to evaluate the results the Building implementation capacity for facilitation (BIC-F) intervention had on the participants' knowledge, skills, and self-efficacy to facilitate implementation and the normalization of a systematic implementation model into their work routines, as well as its use into their respective organizations.

This longitudinal study used a mixed methods sequential explanatory design, in which the quantitative results are explained using the qualitative findings (24). The collected quantitative data was analyzed and used as a starting point to develop the interview guide. The integration of the quantitative and qualitative results is carried out in the discussion section of the manuscript. The Template for Intervention Description and Replication (TIDieR) (25) was followed, and the filled-out checklist is provided as a supplementary file (Supplementary Material: TIDieR Reporting Standards Checklist).

This study evaluated the BIC-F intervention, which aimed to train individuals who hold a facilitating role in organizations. The results, which we will discuss in more detail in subsequent paragraphs, indicate: (1) Statistically significant increases in participants’ perceptions of their knowledge and skills to facilitate implementation and their self-efficacy as a facilitator post-intervention, compared to pre-intervention; and (2) An early stage in the process of normalization of the systematic implementation model in the participants' work routines, with strong coherence but relatively low activation of other NPT mechanisms, and at the organizational level, where many facilitators were still working towards achieving coherence and cognitive participation among relevant stakeholders.

Earlier research has shown that self-efficacy is a good indicator of behavioral modification, and increasing participants’ self-efficacy regarding one or a set of tasks may improve the odds that they perform the task(s) even after the training (36). The importance of self-efficacy in training transfer is also suggested by previous research, which has studied it as an outcome of capacity-building interventions (44). In our study, the statistically significant increase in participants' self-efficacy post-intervention could result from the strong emphasis placed on practice throughout the intervention. Specifically, the participants practiced creating implementation plans for several cases and took part in role-playing exercises. These activities could have served as sources of self-efficacy in the form of mastery experience (also called performance accomplishments), which seems to be the most effective way of developing self-efficacy (45). Another source of self-efficacy that may have been activated during the training was vicarious experience, which entails learning from other participants during group discussions and presentations and by observing workshop leaders apply the knowledge during workshops (45).

Self-efficacy is strengthened by repeated success, achieved through practice (46, 47). The decrease in self-efficacy scores between the first post-intervention measurement and the second post-intervention measurement could be explained, in part, by limited opportunities to apply the knowledge and skills participants acquired in the training after the intervention, as reported in some interviews. Factors external to the organization (e.g., the Covid-19 pandemic) as well as internal to the organization (e.g., lack of prioritization of implementation work) were mentioned as limiting conditions. Previous research has suggested that facilitators can practice their role better if organizations prioritize training and implementation work (48). It is, thus, likely that the professional development of facilitators who undergo even high-quality training is hindered without regular practice, which can only happen under auspicious organizational conditions. Increased contact between the workshop leaders and the facilitators could help the latter maintain momentum after the intervention. Furthermore, workshop leaders and facilitators could find strategies targeted at management to increase the prioritization of implementation work in organizations.

The results of this study indicate that the participants were in the early phases of normalizing the systematic implementation model in their work routines. Six months after the intervention, participants, on average, were not fully familiar with the model, and for many, it did not feel like a natural part of their work routines. However, it was stated that the model could become part of their work routines. There was a general understanding of the steps and value of the model for their implementation work, which was suggested by low S-NoMAD scores on the differentiation, individual specification, and internalization items (lower scores suggest higher contribution to normalization), and by positive interview accounts referring to these submechanisms of coherence. This aligns with previous research showing that an implementation process starts with achieving coherence (49). In light of qualitative data describing that only a few facilitators used the implementation model systematically, the medium rating on interactional workability suggests that the normalization process was in its early stages for many participants. A too short follow-up period (one year) could partly explain this finding, given that a similar study noted that change in behavior and applied knowledge in their sample typically occurred 18–24 months after the intervention (50). The qualitative data give further insight into how participants applied the systematic implementation model. Not all participants utilized all the steps in the model, which is consistent with findings from a prior study that evaluated an intervention teaching the model (51). In our study, a detailed application of the model was perceived by some participants as making the planning process too difficult. This might clarify why a few facilitators were still grappling with how to apply the model to more complex implementation projects. Using the model flexibly or as inspiration was also suggested in a few interviews. This type of use could be understood according to recent theoretical perspectives, which underscore the significance of considering mental models when describing how individuals approach implementation (36). Mental models can determine how information is understood and what changes individuals will consider, and they are shaped over time by continuous learning, experiences, and exposure to different contexts (52), which participants in our study reported. This lends support to the idea that facilitation involves flexibility in applying and blending different approaches (6) and that the “implementation of implementation science knowledge” (54), like any other type of scientific knowledge, will not be linear and straightforward, but rather the subject of individual interpretation and contextual adaptation (55).

Apart from their role in supporting the implementation of EBIs, facilitators can also contribute to building an organization's implementation capacity (6, 16). Intervention participants were encouraged to share the learnings from the training and engage other staff in implementation planning. Medium scores on S-NoMAD items assessing the coherence and engagement of other staff in using the systematic implementation model (e.g., communal specification, initiation) and perceived organizational support for the use of the model (e.g., contextual integration) suggest that the process of its normalization was not very advanced in most organizations. In interviews, there were examples indicating that participants did not mention the model at all, mentioned it briefly, or just created implementation plans based on the model, which they then shared with the group they supported. The limited dissemination of knowledge and engagement of other staff in applying the systematic implementation model might be, to some extent, clarified by the facilitators' need to first become more comfortable with using the model in their routines. This finding is in line with a previous study where learning over time was a mechanism that led to an increased sense of confidence and an improved ability to enact the facilitation role (20), and could also account for why some participants responded with Not relevant at the moment on S-NoMAD items related to collaborative work with other staff members. Additional factors that could have hindered the use of the model in some organizations, as indicated by pre-intervention data and qualitative results, were implementation climate and collegial support.

A strength of this study was its mixed methods sequential explanatory approach, where qualitative data were used to provide possible explanations of the quantitative results. The longitudinal design involving several data collection timepoints is also a strength, as it allowed us to understand better how participants used the knowledge and skills they acquired in the training, what challenges they encountered even after the intervention ended, and the type of support they could benefit from. Furthermore, the diversity of participants in terms of their roles, responsibilities, and organizations indicates that this training might be relevant for a wide range of professionals activating within healthcare and other public services and strengthens the external validity of the study.

The study results should be interpreted considering a series of limitations. An important limitation is the lack of a control group, which impeded us from estimating the effect size of the intervention. Another limitation was the attrition of study participants over time. This could have been because of self-selection bias, leading to not capturing other outcomes and perspectives that might have provided a more detailed understanding of the participants' experiences of the intervention and its results. Furthermore, the extraordinary circumstances in which the intervention was delivered (i.e., the Covid-19 pandemic) and the one-year follow-up period are also considerable limitations which could have impacted the results. However, given that this was a pilot study, a one-year follow-up period was considered sufficient to observe some meaningful changes and to guide a longer evaluation. Another weakness is that the quantitative data were collected using self-reported measures, which could have led to an under- or overestimation of the measured constructs. Furthermore, responses in interviews could have been affected by social desirability bias, but this risk was mitigated by not involving the interviewer in the delivery of the intervention.

Using NPT to frame the longitudinal evaluation and as an aid in interpreting the results is a strength of this study, as the theory has been previously found to provide a robust explanation of how interventions are embedded in practice. However, the deductive approach we used to analyze the qualitative data highlighted overlaps between NPT constructs, which was also observed by other studies (49). This occasionally made coding meaning units with a single NPT construct challenging. For example, individual specification and internalization necessarily required applying the systematic implementation model (i.e., interactional workability). In some cases, interactional workability was also concomitant to reconfiguring the model. Furthermore, NPT is a relational theory, which was relevant for interpreting the perspectives of the facilitators, who are professionals who often initiate and create support for implementation through relational work. Nonetheless, collecting data only from facilitators limited our understanding of the normalization process of the model at the level of the organization, as we tried to make sense of collective processes such as communal specification and enrolment based only on the participants' points of view.

The intervention seemed to improve participants' knowledge, skills, and self-efficacy, and it started the process of normalizing the systematic implementation model in the participants' work routines. Nevertheless, in most cases, the normalization process was in the early stages and was often not approached in an intentional and systematic way because personal or organizational factors hindered transfer of training. To ensure improved transfer of training and higher normalization of the methods they teach, the BIC-F and other interventions need to provide long-term support to facilitators, and include components designed to tackle contextual obstacles and actively build implementation capacity at the organizational level.