Evaluating the Implementation of the SWITCH® School Wellness Intervention Through Mixed Methods

Background: School wellness programming is important for promoting academic achievement and healthy lifestyles in youth; however, research is needed on methods that can help schools implement and sustain programs on their own. The purpose of this study was to investigate factors within and outside the school environment that impacted school capacity for implementation and potential sustainability of wellness programming. Methods: As part of the School Wellness Integration Targeting Child Health (SWITCH®) intervention, elementary school wellness teams (N=30) were guided through a capacity-building process focused on promoting the adoption of healthy lifestyle behaviors in students. Data on implementation were collected through three standardized surveys and interviews (pre-mid-post) and a post-implementation interview. Indicators of organizational capacity were assessed using the School Wellness Readiness Assessment (SWRA). Paired t-tests were run to assess changes in implementation (classroom, physical education, lunchroom settings), capacity, and stakeholder engagement over time. One-way analysis of variance (ANOVA) tests were run to examine how implementation of best practices (low, moderate, high) explained differences in capacity gains. Qualitative data were analyzed through inductive and deductive analysis, following the Consolidated Framework for Implementation Research (CFIR). Results: Paired t-tests showed non-signicant increases in school and setting-specic capacity and implementation of SWITCH best practices over time, in addition to a consistent level of engagement from key stakeholders. ANOVA analyses revealed positive, non-signicant variances between implementation group and gains in school capacity (F[2,24]=1.63; p=.21), class capacity (F[2,24]=0.20 p=.82), lunchroom capacity (F[2,24]=0.29; p=.78), and physical education (F[2,24]=1.45; p=.25). Qualitative data demonstrated factors within the outer setting (i.e., engaging community partners) that facilitated programming. Inner-setting factors (i.e., relationships with administration and staff) signicantly affected implementation. Implementation process construct themes (e.g., planning, adaptation of resources to meet school capacity/needs, and engaging students as leaders) were cited as key facilitators. Schools discussed factors impacting sustainability, such school

To advance research on school wellness programming, it is important to test approaches aimed at building capacity for more effective school wellness programming.
The School Wellness Integration Targeting Child Health (SWITCH®) intervention adopts a capacitybuilding process whereby schools are provided with training and professional development on school wellness programming and evaluation practices (7,8). Through such a process, school wellness teams (SWT) are provided with autonomy to implement program quality elements (i.e., implementation strategies) and setting-speci c best practices (using implementation materials) to the degree that works best for their school and context.
Foundational research on the original Switch intervention (9,10) demonstrated the potential of multiecological-level programming for improving children's intake of fruits and vegetables, increasing physical activity, and decreasing non-educational screen time. Subsequent work based on an established implementation framework (11,12) developed and re ned the SWITCH implementation process to enable schools to carry out programming more independently (7). A recent evaluation has documented the speci c value of the online platform for building self-monitoring skills in youth (blinded paper-2019); however, there is considerable variability in the use of this platform by the schools as well as their adherence to the setting-speci c best practices (blinded paper-in review). Therefore, further research is needed to understand the contextual factors that in uence implementation and subsequent capacity change using mixed methods. Such information is critical to improving training and implementation materials and providing ongoing support to schools as a means to promote comprehensive school wellness environments fostering student health behavior change.
The dissemination and implementation (D&I) literature offers many approaches and tools for understanding implementation with the Consolidated Framework for Implementation Research (CFIR) being one of the most comprehensive and widely used frameworks (13,14). The CFIR has predominantly been used in clinical research (15), but recent applications to school-based health interventions have shown value for evaluating implementation of programming (3,16). Although CFIR has been used to frame studies testing the implementation and potential success of evidence-based programs in schools (15-17), limited research has been conducted to understand the factors that affect school wellness implementation. Further, scant research has been conducted to understand stakeholders' perceptions of capacity and the ability to sustain evidence-based programming over time. Sustainability has been cited as a key issue in the D&I literature (6). To advance research on sustainability, the authors speci cally advocated for the use of 'methodologically stronger primary research, informed by theory', (p. 26) (6). The use of mixed methods offers advantages for triangulating implementation data, and for identifying areas of dissonance between more objective indicators of implementation (18).
The speci c purpose of this study was to examine relations between organizational capacity and implementation of recommended SWITCH best practices. Consistent with guidelines from CFIR, analyses captured the factors both within and outside the school environment that affected capacity for program implementation and potential sustainability.

Study Design
The study design encompasses an evaluation of implementation and the factors within and outside school settings that in uence implementation of SWITCH quality elements and best practices (19). We sought to test implementation of SWITCH as a "real world trial," one step prior to full dissemination in the Scale-Up Continuum of evidence-based interventions (14). Consistent with this continuum, the purpose of this implementation evaluation was to "determine the extent to which the intervention is delivered to the target population as planned" (p. 3) (14). The adoption of CFIR and the use of mixed methods approaches provides advantages for studying the complex relationships in uencing school wellness programming.

Participants And Training Procedures
The 2019 SWITCH program involved 30 elementary schools from 22 counties (of 99) across the state (blinded). Summary data on school characteristics is provided in Table 1. Schools were recruited/enrolled in the summer/fall of 2018 through communication/solicitation and were instructed to form a SWT of three (or more) school staff members and assign a leader. Training was offered from September to December of 2018 and consisted of a series of preparatory webinars and a required one -day conference at the University campus. The composition of SWT representation for the sample and information on SWT leaders is illustrated in Table 2. The purpose of the webinars was to orient SWT to SWITCH quality elements and best practices and their importance/rationale as a means to enhance student health behavior (shown in Table 3). This provided SWT with an opportunity to learn the history of SWITCH, the key facets of programming, and all the resources available for implementation. The purpose of the in-person conference was to provide SWT with rich professional development on module implementation, assessing their school wellness environment for strengths and areas for improvement, and strategic planning.   Moreover, in the 2018-2019 iteration, a key focus for implementation was to establish reciprocal and collaborative relationships with 4-H Extension staff that facilitate county-level 4-H programming. These staff were encouraged to attend the conference and meet with SWT from schools in their county to collectively learn about SWITCH and plan programming. The 4-H Extension staff facilitated goal setting through provision of ideas and support, and worked with SWT to develop an action plan. A separate evaluation of 4-H Extension engagement and program satisfaction was conducted and is outside the scope of this paper (blinded article, 2020).

School Capacity Measurement
The SWITCH evaluation has incorporated indicators to capture elements of organizational capacity or readiness as well as environmental characteristics (20). A speci c tool called the School Wellness Readiness Assessment (SWRA) (20) was developed based on previous conceptions of organizational readiness (21). It consists of questions to assess readiness in the overall school setting, as well as classroom, lunchroom, and physical education settings. SWT were asked to respond on a 5-point Likert scale ranging from "Strongly disagree" to "Strongly agree". The SWRA was administered prior to implementation (week 0) and again at the end of implementation (week 12+).

Implementation Measures Checkpoint Surveys
To assess the degree to which SWT were able to implement the SWITCH program, checkpoint surveys were administered halfway (week 6) and toward the end (week 12) of implementation. Wellness teams were asked to report the degree to which they implemented each of the quality elements and best practices (see Table 3) in the rst and second half of implementation, by selecting between three levels of implementation (Not Yet/Partially/Fully) for each quality element. Setting-speci c implementation was also gathered by asking schools to report best practice utilization in the classroom, lunchroom, and physical education settings. These data provided valuable information on Implementation Process (Planning, Executing) factors most prevalent to schools in the implementation phase. Additional questions asked about the perceived engagement and awareness of school stakeholders (i.e., classroom teachers, administration, food service staff -1 = low; 2 = moderate; 3 = high) to gain a sense of how the SWT are networked within their school community and the degree of collaboration within and outside the building.

Checkpoint Calls
Checkpoint calls were utilized before and throughout the implementation phase to connect with schools and facilitate goal setting/monitoring. The rst checkpoint call was scheduled before implementation (week 0), half way (week 5/6), and toward the end (11/12) of the active implementation phase. Each call was grounded in Motivational Interviewing (MI) principles (22,23), whereby researchers and trained 4-H Extension staff posed questions to SWT with the goal of guiding them to a strategy/solution that works best for their team capacity and school environment. In doing so, support to schools was individualized and speci c to their needs as an implementation site. Schools were asked to allot ~ 30 minutes before/during/after the school day for these conversations. Information regarding training provided to 4-H staff, their experiences with supporting SWITCH implementation, and perceptions of program dissemination can be found in two other articles (blinded author articles, in review). Questions asked in these calls pertained to the Innovation Characteristics (Adaptability, Complexity, Cost), Implementation Process (Planning, Engaging, Executing), Inner Setting (all constructs), and Outer Setting (Cosmopolitanism, Peer Pressure) factors which attributed to SWT's ability to implement SWITCH and develop comprehensive programming (13). All calls were audio recorded and transcribed.

Semi-formal Interviews
After implementation concluded (post-week 12), a semi-formal interview was conducted with SWT by an independent contractor, highly-skilled in qualitative methods, to gather re ective input and feedback on the program and the ways in which schools were able to perceive their ability to sustain wellness programming after formal implementation ended. The interview guide was designed to address how training and support from the SWITCH research team/Extension, as well as factors within/outside the school environment, affected SWT ability to implement the program. An example question was: "How would you describe the level of support you received from your county youth Extension o cer(s)? To what degree did they facilitate SWITCH implementation?". Interview data provided valuable insights that could not be gleaned from checkpoint calls, as SWT re ected on all 12 weeks and provided objective feedback on the program (24). Interviews lasted between 10 and 27 minutes, were audio recorded, and transcribed verbatim. Figure 1 illustrates the number of interviews/checkpoint calls conducted at each time point, showing a fairly even distribution between new and returning schools.

Data Analysis
All quantitative data were entered into Statistical Analysis Software (SAS; Version 9.6, Raleigh NC), processed, and tested for normality of distribution. Checkpoint survey and SWRA data were merged with school demographic data and sorted by time, providing two time points (pre-and post-SWITCH) for each school. To establish a composite measure of implementation, an average quality element score (5 elements; scored 1-3) was added to average scores for each of the setting best practices (lunch, physical education, classroom; scored 1-3), yielding a score between 2-6. Subsequent tertile splits were computed to separate high versus moderate and low implementing schools; such method has strong support and utility in modeling research especially when a linear relationship cannot be assumed among variables (25). Scores were calculated for overall SWRA scores, and changes in capacity in each school setting (post-pre) were computed to determine capacity change over time.
Pearson bivariate correlations were conducted to examine relations between changes in implementation over time and changes in school-level wellness capacity. In addition, paired t-tests were conducted to examine changes in setting-speci c capacity over time. Finally, a series of one-way ANOVAs were run to understand the differences between implementation (low, moderate, high) and subsequent gains in capacity (setting speci c and overall). Statistical signi cance was assumed at α < 0.05.
Qualitative data collection yielded 90 transcripts across the four time points, requiring a rigorous and systematic procedure; thus, analysis followed constant comparison techniques (26). Based on this approach, data were analyzed rst through open coding, followed by axial coding while maintaining an inductive approach, before deductive analysis was conducted through the lens of CFIR domains. This provided a clearer understanding of how factors related to the intervention and those within and outside the school environment affected implementation and SWT capacity to achieve systems change in their school community. The purpose of axial coding was to test the un-coded data against the initial codebook for agreement and to make re nements to themes (27,28). A separate subsample of transcripts from each time point were reviewed and coded into the dataset. Where new ideas/themes emerged, the coders developed new nodes in NVivo to accommodate emerging concepts. Multiple rst-order themes were generated with several second order themes encapsulated within these concepts. The nal product of this phase was a more comprehensive codebook with many main (> 10) themes and subthemes that represent qualitative data.
Deductive analysis comprised a systematic approach to conceptualize how the data aligned with CFIR. Use of the comparative method and negative case analysis were highly integrated, as it was important to look for cases that do not align with the overall implementation framework structure (27,28). When initial themes were developed, the lead author and another co-author reviewed the themes and began to merge the inductive codebook with the NVivo CFIR codebook template (29). This process allowed the researchers to carefully consider each domain and construct, and to determine appropriate alignment to the model, where relevant. Furthermore, researchers sought to code data related to the sustainability of SWITCH and the degree to which SWT believed they could continue programming as part of school culture.
Guided by experts in qualitative analysis (30,31), a series of strategies were employed to ensure the trustworthiness of the data and analyses. First, a comprehensive audit trail was kept throughout the analytical process between the two primary coders to document key issues, trends, and modi cations that arose, facilitating transparency in collaborative research (32). Second, peer debrie ng took place whereby two other researchers reviewed codes and provided feedback on the interpretation of data, identifying gaps or areas of potential oversight (31). Finally, thorough negative case analyses were integrated to ensure that themes accurately represented the data and identi ed areas of divergence from themes (30,31).

SWITCH Implementation
Overall, SWT reported implementing most of the program quality elements (see Fig. 2). Almost all schools reported regular SWT meetings (n = 27, 93.1%), followed by getting students online and adopting selfmonitoring skills through the website (n = 25, 86.2%). Comprehensive integration of using SWITCH resources across the school setting (n = 23, 79.3%) and working to involve parents (n = 23, 79.3%) were implemented fairly well. Only 3 schools (10.3%) reported involving youth in a leadership capacity to deliver wellness programming (see Fig. 2). For setting-speci c best practices, at week 6, implementation in the classroom was lowest compared to the lunchroom and physical education. In week 12, the opposite occurred, whereby classroom implementation was the highest, followed by physical education and lunchroom. The lunchroom implementation scores all decreased aside from reinforcement of themes (see Fig. 3).
The SWT rated the relative engagement of key stakeholders in SWITCH at several timepoints (Fig. 4). The reported engagement at week 6 and 12 was highest among physical education (2.9 ± 0.4 and 2.8 ± 0.5, respectively) and students (2.8 ± 0.4 and 2.7 ± 0.5, respectively) and lowest for parent engagement (2.0 ± 0.6 and 1.5 ± 0.5, respectively). All scores decreased slightly, except for student engagement, which remained the same. The SWT reported increases in capacity over time, as captured by the SWRA tool, in each setting (see Fig. 5). The biggest increase in capacity was in the lunchroom, from 2.3 ± 0.5 to 4.1 ± 0.5, equaling a 78% change in score, followed by physical education (from 2.9 ± 0.4 to 4.3 ± 0.5; 48% increase). School capacity increased 6% (from 3.2 ± 0.9 to 3.4 ± 0.5) and classroom capacity increased 2.5% (from 3.9 ± 0.6 to 4.0 ± 0.5), with classroom having the highest capacity at baseline.
Pearson bivariate correlations revealed small-moderate (non-signi cant) relationships between implementation composite scores and percent changes in capacity for the school (r = .16), classroom (r = .35), physical education (r = .02), and lunchroom indicators (r = .06). Paired t-tests also showed nonsigni cant changes in setting-speci c capacity over time (all p > .05). ANOVA results also showed nonsigni cant ndings. As Fig. 6 shows, small differences were found between the levels of implementation

Innovation Characteristics
Findings from interview and checkpoint call data highlight the ways in which SWT adapted SWITCH best practices and resources to t the needs of their school system. One of the examples commonly cited was the adaptation of module/curricular content intended for one setting into another setting, as illustrated in Table 4, where the classroom module was adopted into science curricula to synergize with concepts taught in this content area. Further, schools discussed how they adapted activities, such as conducting taste tests, on a limited budget, demonstrating the ability of the intervention model to be adapted to school needs and capacity.
Many SWT discussed how SWITCH implementation forced them to think about how to best implement all its parts, which required extensive planning and converging at the in-person training conference and the weeks thereafter. Several SWT saw the need to revisit how classroom instruction was conducted and took it upon themselves to provide professional development and resources to their peers as a means to enhance the implementation.

Outer Setting
In 2019 the partnership with state 4-H Extension grew and the collaborating staff were encouraged to support school wellness initiatives by meeting with schools and helping with implementation. This support ranged in its intensity; some SWT felt that Extension support was vital to their success, stating that their local 4-H staff were a "part of the team" and took responsibility for various tasks. Other SWT talked about how Extension staff kept them accountable for the goals they had set in the rst checkpoint call and were responsive to requests for help with various programming aspects.
Schools also collaborated with local agencies, predominantly for help with supply of food and equipment for programming and for hosting certain school-wide events for students and families. Finally, some schools reported hearing about the ways in which other returning schools had been successful in the 2017-18 SWITCH cycle at the in-person conference, and about what other schools in their district were doing, using this as motivation for planning their wellness activities/programs in the 2019 cycle.

Inner Setting
Most SWT discussed relationships with other teachers outside the SWT as a factor in their perceived ability to execute SWITCH best practices; most reported a positive reaction to programing and that their teachers were getting on board, whereas others mentioned a lack of staff buy-in which hindered their efforts. There was a sense of staff members feeling overwhelmed with other responsibilities, which may have attributed to the lack of buy-in from the whole school.
Most SWT reported good support from administration, and in relation to the implementation climate, some SWT reported that they were able to schedule additional opportunities and make room for wellness in their schedule, such as adding another block of recess. Unfortunately, some SWT lamented the lack of time provided to plan as a team and lack of facilities to conduct programming. Finally, some SWT discussed the challenges faced when trying to change unhealthy "norms" in school culture, such as using recess as punishment, food as a reward, or more veteran staff not wishing to innovate in certain areas of the school setting. These factors acted as barriers to implementation of programing and relate to other ndings (e.g., lack of buy-in, limited time/resources, etc.).

Individual Characteristics
Throughout the formal programming phase, it became clear that SWT motivation was varied, as some teams were more motivated/excited than others. Table 4 re ects some of the key extracts from SWT, highlighting the divergent perspectives and situations. Most of the low motivation was linked to SWT feeling overwhelmed, and was reported predominantly in new schools, suggesting that lack of motivation/excitement may be attributed to inner setting factors. Further, for the most part, SWT felt a strong sense of collaboration and engagement from others in their school, but some discussed sociopolitical tension in their school setting, which impeded their perceived ability to implement programming across the school setting.

Implementation Process
The in-person training conference was cited as a key socialization factor and motivator for school teams, particularly new schools, as they were able to learn from other SWT about their implementation practices and launch goal setting/planning for the semester ahead of time. Throughout the implementation phase, many schools discussed ways they were able to meet and accomplish planning responsibilities while balancing different schedules, such as biweekly meetings and sending emails.
SWT expressed their success with changing perceptions of students; many reported how strategies such as taste tests, physical activity breaks, and monitoring screen time at home fostered behavior change in students and enhanced awareness of the importance of health behavior. The integration of students as leaders of wellness programming was described by many as a way to bolster their efforts to deliver SWITCH school-wide, as students also served as implementation champions, expanding SWT reach across the school setting. This aligns with ndings from the checkpoint survey, showing that students were one of the most engaged stakeholders (see Fig. 4). Reports of parent involvement were somewhat mixed. Although some schools were successful in generating awareness, some SWT reported an overall lack of parental involvement with SWITCH programming. SWT were able to engage local organizations, such as grocery stores and representatives from non-pro t/federal groups (i.e., AmeriCorps, Food Corps) to facilitate with programming and even visit the school to help implement lessons and assist in schoolwide events. Further, SWT alluded to speci c strategies and practices adopted by their local Extension o cer that greatly facilitated programming across the school.
With regard to executing the intervention according to plan (i.e., program best practices), several speci c approaches were mentioned throughout the implementation phase as particularly successful. Many SWT found that the lessons in modules (i.e., classroom and physical education) were too long to implement (~ 20 minutes) or they didn't have enough time to plan, therefore modi ed lessons to t their schedule or split them up over two content periods. Other practices included teaching lessons in other parts of the school curriculum, such as science or social studies, which worked especially when school days were missed due to inclement weather. Interactive posters in the classroom, lunchroom, and physical education setting were commonly used and served to create a culture of health by driving student interest through posting trivia questions and challenges.
Finally, a key emphasis toward the end of SWITCH implementation was on sustainability, and how programs/practices could be sustained once the formal 12-week implementation phase was over. Many SWT felt con dent that the lessons, resources, and programming would be easy to sustain and discussed how they had begun to link programming to their school wellness policy. Many tied this within their triannual assessment phase as per policy mandates, and found it was an important step which allowed them to revisit many practices and policies in existence. On the other hand, some SWT felt that their policy was not a priority, and was merely a document to satisfy external requirements by administration. This was often due to administration who were either not supportive or aware of the need to revise wellness policies as a result of SWITCH implementation.

Discussion
The study provides novel insights about school wellness implementation and the factors within and outside the school environment that affected school capacity and SWITCH program implementation. The adoption of CFIR and the use of mixed methods approaches were critical features for studying the complex relationships in uencing school wellness programming. Overall, SWT were able to adapt the intervention to meet the needs of their school setting and stakeholders, demonstrating the adaptability of SWT to mitigate structural barriers in the school environment. Findings from surveys and checkpoint calls demonstrated moderate-to-strong implementation of the best practices and an overall increase in settingspeci c implementation over the 12-week period. Unlike prior studies that showed organizational readiness/capacity as a key predictor of implementation (33,34), our ndings indicated that changes in capacity followed a sequential pattern in relation to implementation, whereby the moderate and high implementing groups demonstrated greater capacity change in multiple school settings. These results have implications for further study of the relation between implementation and capacity, and the possibility of a bidirectional relation between these constructs.
Schools reported moderate-to-high engagement from students, classroom teachers, and physical education teachers, but lower engagement was reported from administration, food service staff, and parents. These ndings may be, in part, due to the SWT representation which comprised mostly classroom and physical education teachers. A key nding was that school, classroom, physical education, and lunchroom capacity increased over the course of the implementation phase. These several points during implementation. Through these touchpoints, Extension staff helped schools work through implementation barriers by providing advice, information on strategies that had been effective in other schools, and in some cases, they directly assisted with special events or programming.
Findings from interviews document the additive bene t of extension support; many schools felt that 4-H Extension o cer was instrumental to implementation and would not have been able to implement SWITCH as well without this support. Interestingly, this type of support is often lacking in schools trying to implement wellness programming but was regarded as tting within the scope of the 4-H Extension staff scope of responsibilities for their position (blinded paper, 2020). The provision of funding for (or direct sta ng of) a wellness coordinator is potentially effective but not sustainable (36). Establishing effective linkages between schools and community partners that can assist with wellness program implementation in a sustainable manner is crucial for establishing long-term, effective wellness programming.
The SWRA provided a means to understand 'Characteristics of Individuals' and 'Inner Setting' factors which may contextualize the implementation practices adopted by schools. Previous studies have investigated the association between school capacity for wellness and program implementation and outcomes (33,35,37). Findings from these studies document the potential importance of capacity as an in uential factor associated with the implementation of school wellness initiatives. A retrospective study evaluating implementation of the CATCH intervention in 36 Illinois elementary schools identi ed that organizational readiness (i.e., capacity) was a key factor facilitating or hindering school staff implementation as well as classroom teacher implementation in the classroom setting (37). The authors reported that classroom teachers were required to do more work to implement CATCH compared to physical education and cafeteria supervisors/managers and there was no direct or external incentive for their efforts (37).
In relation to 'compatibility' and 'relative priority' from the inner setting, interview data provided valuable contextual information on school capacity to implement programming. One reason why some schools reported lower capacity at baseline may relate to the nding that SWT felt they were overwhelming staff with asking them to implement SWITCH lessons or use posters in each of the school settings, as reported tension over changing the "status quo". Some schools also reported di culty gaining staff buy in; such tension has been reported in prior research (4,38) and highlights complexity when a new innovation is introduced to school settings. To combat this issue, some SWT reported using staff wellness challenges to gain buy-in and enhance the overall school culture, and making SWITCH modules/resources easier to use through creating binders/photocopying lessons for teachers to take (Planning-Striking a balance).
Strategies such as these demonstrate the adaptability of SWITCH materials and messaging to align with the needs and context of different school environments, ultimately enhancing implementation and systems change for wellness promotion.
In the present study, classroom teachers had the lowest level of implementation in the early phase of SWITCH compared to the lunchroom and physical education, but the highest level of implementation in the later phase of SWITCH. This may be due to the teachers having more time to explore the SWITCH module and plan for integrating the content within their curriculum and teaching schedule. In addition, the check-ins with Extension staff at the mid-point of the intervention may have been useful for helping teachers critically consider and plan for better integration of the SWITCH module during the second phase of the intervention. Interview data also highlight the positive impact of concrete 'planning' (implementation process) opportunities, such as the training conference and pre-implementation webinars, on implementation. This gave SWT a chance to learn vicariously from other schools' successes (outer setting -'peer pressure') and meet as a team to plan wellness initiatives for the upcoming implementation phase.
However, other factors within the inner setting (available resources), such as lack of time and facilities/personnel, hindered implementation. This barrier is a highly cited issue in many school-based health behavior interventions (39)(40)(41)(42), and in this study manifested as a lack of time to implement SWITCH quality elements and best practices. SWITCH emphasizes the importance of regular SWT meetings and evaluation procedures including meetings as a component of implementation measurement; checkpoint surveys and interviews did suggest that SWT met regularly as a team to plan and coordinate implementation strategies. Through interviews, schools discussed their adaptation of this best practice by meeting brie y before/after school, within mutual planning time, and/or planning implementation through email. As such, although time is often cited as a barrier, the in-depth data analysis process facilitated understanding of how schools overcome these issues to maintain a high degree of implementation delity. Findings provide practical strategies for developing school-based implementation frameworks based on empirical evidence.
With regard to 'executing' (implementation process), SWT reported ways in which the SWITCH materials were adapted to meet their school needs, such as teaching classroom modules in science classes, integrating the physical education module into warm-up lessons, and other practical strategies which would not have been gleaned through checkpoint surveys alone. Other researchers have found that delity of implementation of HIV-prevention intervention curriculum was strongly associated with student HIV knowledge and self-e cacy outcomes (43). In their study, Wang et al. (43) categorized teachers into low, moderate, and high implementers based on their usage (number of lessons) and delity (carrying out lessons as planned) of curriculum indicating that delity to lesson plans was of great importance. In the present study, due to the comprehensive nature of SWITCH and that its quality elements/best practices target many aspects of the school environment, " delity" is a complex and multifaceted concept. The fact that SWT were able to adapt content to meet the needs of their individual school contexts, we believe, is a positive nding and aligns with the CFIR Intervention Characteristics construct of "adaptability" (13,15).
Furthermore, when working with schools, it is critical to understand the pedagogical relevance of intervention materials in order to meet the needs of a wide range of school stakeholders.
Finally, the ways in which SWT re ected on implementation and the degree to which SWITCH programming could be sustained provided valuable information for future improvement. Most schools were extremely positive in their perceptions of SWITCH as a sustainable program, yet when discussing implications for their school or district wellness policy, several SWT had little to no involvement in this process. Furthermore, some reported that, when asking about their policy, were not given clear information on this document or its whereabouts, highlighting a disconnect between policy and practice (44). Although training was provided at the in-person conference about the Final Rule mandate, it is likely that more training and support is needed to help SWT as a means to enhance the connection between practice and policy.
The ndings add rich understanding to the steps needed to sustain school wellness programming in school settings (6,45), but it is important to acknowledge limitations. First, the SWITCH schools selfselected to enroll in the program, and were not representative of all schools in the state or country, limiting generalizability of these ndings. Second, the checkpoint calls and interviews were conducted at times convenient for SWT members, and sometimes not all group members could not be present, limiting our understanding of all aspects of implementation. However, it must be noted that this is a logistical challenge when working with schools and it is imperative to adopt exible approaches to data collection in such dynamic settings.

Conclusions
The purpose of this study was to examine the factors within and outside the school environment that in uence capacity for program implementation. Findings underscore the need to study implementation using multiple methods, and the organizational capacity that might predict the degree to which schools implement the program. Multiple constraints were found to arise from the inner setting, such as lack of time/resources, staff buy-in, and administrator support. However, the provision of 4-H Extension support (outer setting) and the adaptability of the SWITCH curriculum/resources (intervention characteristics, implementation process) to be tailored to meet school context and capacity. This additional support may mitigate some of the barriers and constraints faced by SWT, enhancing the potential for sustainability of SWITCH. Finally, results show the importance of the school wellness conference as a training and socialization process, acting as a catalyst for planning and readiness for implementation. Further improvement to the SWITCH program is needed to support schools in taking their practice to policy, as a means to enhance school wellness environments and promote student health behavior.

Declarations
Ethics approval and consent to participate:   Changes in setting-speci c capacity over time

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. StaRIchecklist2019ImplementationPaper.docx