Implementation and systems impact of a systems-based health promotion programme at Danish vocational schools: results after 2-years follow-up Authors: Clara Heinze, Helene Rald Johnsen, Rikke Fredenslund Krølner, Charlotte Demant Klinker Published: 2026 ABSTRACT BackgroundVocational schools are important settings for health promotion among young people from lower socioeconomic backgrounds, who are at an increased risk of engaging in health-compromising behaviours. However, health promotion interventions are rarely sustained in such settings. Systems approaches have been proposed as a promising approach to health promotion but remain unexplored in vocational school settings in Denmark. This study aimed to examine the implementation and systems impact of a systems approach to health promotion in Danish vocational schools and to elucidate the overarching pathways through which such impacts may emerge.MethodsThis longitudinal qualitative study was conducted as part of the Data Health programme, a systems-based initiative implemented across eight Danish vocational schools and four municipalities. Quarterly interviews with school coordinators or department managers (n = 59) were analysed using principles from Ripple Effect Mapping to visualise evolving actions and perceived impacts over a two-year period and identify enablers and barriers to systems impact. A cross-case analysis was used to identify plausible pathways enabling perceived systems impact.ResultsAt two-year follow-up, 42 actions had been implemented, and systems impact (structural, relational, and/or transformative change) was perceived to have occurred at six out of eight schools. Two reinforcing systems impact pathways emerged: (1) internally driven change via committed staff and distributed leadership, and (2) externally supported change through collaboration with community actors. Key enablers included dedicated change agents, financial resources, student participation, external networks, and backbone support. Barriers included staff turnover, time constraints, and students alternating between school and workplace-based training.ConclusionsThese findings provide valuable insights for the future adaptation of the Data Health programme and for informing broader systems-based strategies in vocational school settings.Trial registrationRegistered in ClinicalTrials.gov: NCT05308459, retrospectively registered on 4th April 2022.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1186/s12889-026-28208-0. Background Schools serve as an important setting for supporting and promoting students’ health and wellbeing throughout their time at school and can further support lifelong healthy behaviour [1]. A large proportion of vocational students come from lower socioeconomic backgrounds and are more likely than their peers in academic high schools to engage in unhealthy behaviours such as smoking, poor dietary habits, and physical inactivity [2–5], making health promotion important at such schools. Prior initiatives in vocational schools have targeted various health aspects, such as well-being [6], sleep habits [7], physical activity (PA) [8], and smoking prevalence [9]. However, health promotion initiatives in schools are generally not sustained after the intervention period despite being effective during the trial. This primarily reflects the dynamic and interdependent nature of school systems, where factors such as culture, leadership, resources, and external societal changes continuously interact to shape practices and influence the sustainability of health promotion efforts [10–12]. In health promotion research and practice, there is a growing recognition that intervention approaches targeting multiple parts of a system simultaneously have the potential to create a greater impact compared to interventions only targeting the individual level [13, 14]. Systems approaches have been proposed as a way to address complex public health challenges by transforming the environments that shape individual behaviour. Such approaches refer to an overarching intervention perspective that targets multiple interacting components within a system, including organisational structures, relationships, and underlying norms, often through the involvement of stakeholders across levels [15]. In schools, systems approaches to health promotion have been recognised since the 1990s, for example, in the Coordinated School Health framework [16] or the World Health Organisation’s (WHO) concept of Health Promoting Schools (HPS) [17]. The HPS framework adopts a broad, whole-school strategy that addresses health at multiple levels, such as the individual, organisational, and policy domains. This approach seeks to involve stakeholder groups, such as students, educators, parents, and members of the wider community, in the development, execution, and long-term integration of health-related activities [17]. Evidence supporting the effects of whole-school approaches at the individual level is well documented, including enhancements in health-related behaviours and academic outcomes [17–19]. Moreover, some research has pointed to strengthened organisational capacity and supportive structural shifts in school culture as other beneficial system-level outcomes [19–21]. However, such whole-school approaches have only to a limited extent been implemented and studied in vocational school settings [21] and remain unexplored in a Danish context. Transforming a system requires shifts in beliefs, attitudes, and norms of organisations and individuals within a system [22]. Various theories and models have been proposed to define systems change [23–26]. One such definition, developed by Kania et al. (2018), suggests that systems change in organisations can occur across three interrelated levels: structural (policies, practices, and resource flows); relational (patterns of collaboration and power dynamics); and transformative (shifts in underlying mindsets, beliefs, and assumptions) [26]. In the case of health promotion in vocational schools, systems change refers to how health and wellbeing are valued and integrated into the daily routines and mindsets of individuals within the school. As systems are continuously evolving, evaluating systems change requires methodologies that go beyond linear causality and simple input–output models [27–31]. Recently, a growing body of theoretical and methodological literature has explored how to evaluate the systems impact of such efforts [22, 27–30, 32]. However, actual evaluations remain limited. Most existing literature on the evaluation of systems approaches focuses on identifying the potential for systems changes [33, 34] or evaluating the exposure, dose received, and reach related to implementation [30, 35]. There is a growing recognition of the need to capture complexity, adaptation, and emergent outcomes in dynamic systems, and methodological approaches to evaluate systems impact are rapidly developing in response [27, 30]. Ripple Effects Mapping (REM) is one such method which engages stakeholders in plotting the ‘ripple effects’ of a programme or collaboration into a visual map [29, 36]. REM can provide researchers or evaluators with insights into how implementation and impacts evolve within a complex adaptive system, often in unexpected or non-linear ways [36]. In addition, a recent study demonstrated that REM could be compared across settings in order to explore key enabling mechanisms to achieve systems impact at a higher level of abstraction and generate more generalisable theories [37]. Based on this backdrop, the aim of this study was twofold. First, we wanted to explore the implementation and systems impact of a systems approach at Danish vocational schools. Second, we examined the overarching pathways (chains of mechanisms) through which systems impact might be achieved. The study was exploratory and conceptually informed by systems thinking, (schools were understood as complex systems shaped by interacting actors and processes), and by whole-school approaches to health promotion, which guided our overall understanding of how implementation and change processes unfolded. The current study was part of the evaluation of the Data-Driven and Systems Approach to Health Promotion Among Vocational Students (Data Health) programme, which was initiated across eight Danish vocational school sites in Year 2022 [38]. Methods Study setting Vocational education in Denmark is a practice-oriented educational programme lasting typically three to four years [39]. In 2024, there were 86 vocational schools across 259 locations in Denmark, enrolling approximately 104,000 students [40]. The educational programmes follow a dual model, combining school-based instruction with practical training in workplaces, with approximately two-thirds of the time spent in workplace-based training [39]. The more than 100 available programmes are organised into four main educational tracks: (1) Technology, construction, and transportation (e.g., plumber or automotive technician), (2) Food, agriculture, and hospitality (e.g., baker or chef), (3) Administration, commerce, and business services (e.g., payroll officer or event coordinator), (4) Care, health, and pedagogy (e.g., dental assistant or health care worker) [41]. Student demographics differ across programmes; technical programmes tend to enrol a higher proportion of male and younger students, while care-related fields tend to have a higher proportion of female and older students [42]. The schools vary in size and are usually managed by a principal manager, with departmental managers overseeing specific areas. The school staff includes both teaching and administrative personnel. Teachers provide theoretical and practical instruction to develop job-relevant skills, whereas administrative staff manage daily operations, student services, and external coordination [41]. The Data Health programme The Data Health programme was a systems-based programme designed to promote health and wellbeing among vocational students through co-created, locally developed actions [38]. The programme was created by the research team, drawing on principles from systems thinking [43, 44], the HPS approach [17], and data-driven approaches [45, 46]. Four schools started the programme in January 2022, and four started in August 2022. This was partly due to practical and evaluation reasons and partly due to when the schools anticipated having resources available [38]. The programme comprised a five-step iterative process, as illustrated in Fig. 1. Fig. 1Overview of the five steps of the Data Health programme [38] Overview of the five steps of the Data Health programme [38] The programme began with formal collaboration agreements between schools and local municipalities, with each appointing a dedicated coordinator (Step 1). School coordinators received one month’s salary to support programme implementation. Local survey data on student health behaviours were collected and presented to management, staff, and students to stimulate dialogue and support the selection of a targeted health problem (Step 2). Four schools selected physical inactivity as the main problem to address, while the other four focused on issues related to student wellbeing. Department managers made the final decisions but were encouraged to consider data and feedback from both students and staff [47]. This was followed by a Group Model Building (GMB) process, a participatory method rooted in systems thinking in which stakeholders collaboratively develop shared representations of system structures and relationships [44]. The GMB process consisted of three workshops at each school (Step 3) and engaged school actors (students, staff and managers), municipal representatives, and community members in mapping system drivers and identifying leverage points for change. Action groups were hereafter established across all study sites to take responsibility for the continued development and implementation of actions. Across the eight study sites, 39 action ideas were co-created during the final GMB session (GMB3). Additional results on implementation fidelity and outputs of the participatory co-creation process have been reported in detail elsewhere [47]. Each school was allocated a budget of 10,000 DKK (≈ €1,350) to support the implementation of actions. To motivate and promote the exchange of knowledge and learning across study sites, the research team established a Community of Practice (CoP) network with quarterly meetings. Municipal and school coordinators were invited to participate in the meetings (Steps 4 and 5). Approximately one year into the programme, the research team withdrew direct support for the schools and facilitated a meeting to promote sustainability at each school. School and municipal coordinators were encouraged to discuss and develop plans for continued collaboration and long-term strategies to integrate programme components (e.g., collaboration, the coordinator role, and the use of data) into school routines and structures (Steps 4 and 5). Study design and boundaries This was a longitudinal qualitative study, involving collection of qualitative data over time to examine processes of change and development within specific contexts [48, 49]. This design is particularly well suited for exploring how implementation processes and system changes unfold over time within complex adaptive settings [49]. Qualitative data were obtained quarterly over a two-year period following the GMB sessions. Systems thinking was used as an overarching lens in our evaluation strategy to explore both the implementation process and perceived impact within the broader system. To detect changes within a system, it is essential to establish clear system boundaries to determine what is considered inside or outside the scope of the evaluation [27]. In the current study, the term ‘whole school’ [50] was used to define the systems, indicating that all components of the schools, including processes, organisational structures, and relevant stakeholders and collaborations (such as community or municipal actors), were considered integral parts of the local systems in relation to the issue of interest (PA or wellbeing). The system boundary was defined to include actors and structures directly involved in or influencing school-based actions, while actors not directly engaged in these processes were considered outside the system, although they may still exert indirect influence. Within this boundary, an “action” was defined as any planned or implemented activity initiated during the data collection period and addressing the issue of interest (PA or wellbeing). Data collection Participants We conducted semi-structured interviews with the person responsible for overseeing and coordinating the majority of actions at their respective schools. Initially, this role was assigned to school coordinators; however, over time, this responsibility shifted in some instances due to staff turnover, reorganisation of responsibilities, or other organisational changes. The interview guide The interview guide was informed by previous work and studies on implementation processes and systems change, which guided the development of key themes and questions [51, 52]. The interview guide focused on three key themes while remaining open to other topics of importance to the participants: (1) Actions: characteristics of the actions delivered, purpose to drive change, and where in the process the actions were (planning, implementing, maintaining, or stalled) and why; (2) Perceived impact: what/who has been affected, and have the action(s) led to perceived intended or unintended consequences (both positive and negative) or inspired new/subsequent actions; (3) Perceived factors leading to impact: financial resources, the people involved in implementation, and other influencing factors of importance according to the participants. In the final interview (24 months follow-up), we incorporated five additional questions inspired by Appreciative Inquiry [36] to explore participants’ experiences further, such as successes, achievements, or challenges that were triggered by or resulted from the programme. The interviewer paid particular attention to identifying links between actions and perceived impacts by asking follow-up questions to participants and clarifying the connections. Additionally, the purpose of the final interviews was to encourage further development of actions and sustainability beyond the evaluation period. Please see Appendix I for the full interview guide, including the questions for the final interview session. Procedure For early-start schools, data collection took place from August 2022 to September 2024, and for late-start schools, from January 2023 to March 2025. Two authors (CH and HRJ) were responsible for conducting the interviews every three months and ensuring that data were collected as planned. These two authors had a broad and in-depth understanding of what was implemented across all sites. The interviews were scheduled at all sites through email contact with the school coordinators or designated responsible persons. If there were staff changes, we contacted the department manager to clarify who would be the appropriate person to interview. All interviews were conducted and recorded via phone or video calls, each lasting approximately 30–45 min. The final interview lasted approximately 60 min to allow time for additional questions. Data analysis All analyses were conducted by two authors (CH and HRJ), and findings were discussed with co-authors to enhance credibility of the findings [53]. As a starting point, the main content of the audio files was transcribed and summarised across all study sites in an Excel-based action registry, including actions, implementation processes, responsible actors, and perceived impacts. Additionally, we documented participants’ perceptions of how the actions had created an impact within the system during the two-year period. After all interviews were conducted, information from the action registry was used to visualise the relationships between actions and perceived impacts over two years, based on principles from REM [29, 36]. We generated separate REM timelines for each study site in Microsoft PowerPoint, following a four-step process: (1) all identified actions were mapped over time; (2) all perceived impacts were incorporated into the maps; (3) connections between actions and impacts, according to the participants, were drawn to visualise the ‘ripple effect over time; and (4) perceived factors affecting impacts (both enablers and barriers) were identified in the data, discussed, and added to the map, with discrepancies resolved through consensus. The action registry and REM timelines provided a structured record of the analytical process, ensuring transparency from raw data to final interpretations. This approach allowed us to illustrate how different actions were connected at each study site and the pathways through which they might generate impact. For example, a data segment describing the implementation of phone-free teaching was coded as an action. Its integration into school policies, action plans, and student handbooks was coded as impacts and linked in the REM timeline, while factors such as management support or student resistance were coded as enablers or barriers. We analysed the REM timelines separately, first counting how many actions were implemented and active at each point in time (e.g., a one-day event appeared at only one time point). Next, we deductively categorised perceived impacts, derived from the REM timelines, based on the definitions of the Six Conditions of Systems Change [26], as shown in Table 1. We then identified all factors influencing the REM timelines and categorised them as either enablers or barriers affecting the identified impacts. Table 1Applied definitions of the ‘Six Conditions of Systems Change’, adapted from [26]ConditionsDefinitionsStructural change PoliciesFormal rules, regulations, and priorities set within school institutions or relevant organisations that guide changes. PracticesStated activities carried out by organisations, networks, or partnerships to support changes. This also includes internal procedures, informal routines, and day-to-day work. Resource flowsThe allocation and distribution of key resources, such as funding, staff, knowledge, and information.Relational change Relationships & connectionsThe quality of relationships and communication between actors in the system, especially between those with different expertise. Power dynamicsHow authority, decision-making rights, and both formal and informal influence are shared or concentrated among individuals and organisations.Transformative change Mental modelsDeeply held beliefs, assumptions, and habitual ways of thinking influence how people perceive the world, make decisions, and interact with others. Table 1Applied definitions of the ‘Six Conditions of Systems Change’, adapted from [26]ConditionsDefinitionsStructural change PoliciesFormal rules, regulations, and priorities set within school institutions or relevant organisations that guide changes. PracticesStated activities carried out by organisations, networks, or partnerships to support changes. This also includes internal procedures, informal routines, and day-to-day work. Resource flowsThe allocation and distribution of key resources, such as funding, staff, knowledge, and information.Relational change Relationships & connectionsThe quality of relationships and communication between actors in the system, especially between those with different expertise. Power dynamicsHow authority, decision-making rights, and both formal and informal influence are shared or concentrated among individuals and organisations.Transformative change Mental modelsDeeply held beliefs, assumptions, and habitual ways of thinking influence how people perceive the world, make decisions, and interact with others. Applied definitions of the ‘Six Conditions of Systems Change’, adapted from [26] Finally, a cross-case analysis [54] was conducted based on the eight REM timelines to identify plausible cross-cutting pathways through which systems impact might be achieved. This process involved an iterative exploration of how and why perceived impact emerged across the study sites, allowing us to trace the underlying mechanisms as well as the contextual and programme-related factors that contributed to change in some settings. It also enabled us to explore why similar pathways were not associated with perceived impact in other cases. To synthesise and visualise these findings, we drew on principles from system dynamics, particularly the focus on identifying interdependencies, feedback loops, and dynamic relationships within complex systems [55]. Based on this, we developed a Causal Loop Diagram (CLD) [23] to illustrate the plausible pathways, key contextual and programme-related influences, and highlight the presence and influence of feedback loops linked to perceived systems impact. The mechanisms and connections between them were included in the CLD if they were applicable to two or more study sites. Results The characteristics of the study sites, along with information on participants involved in this study, are presented in Table 2. A total of 59 interviews were conducted across the eight study sites, out of 64 initially planned. This was due to the withdrawal of School A from the programme nine months after the GMB process, following changes in management staff and shifting priorities at both the school and municipal levels. Therefore, the analysis of School A only reflects the initial nine months. Table 2Characteristics of study sitesEducational trackSchool sizeGeographical location (region)GMB was conductedTargeted health behaviourPrimary responsible person and number of interviews conductedSchool ASocial healthcareMediumZealandMay/June 2022Well-beingDepartment manager, N = 3School BAgri-culturalSmallCapital regionMay/June 2022Well-beingSchool coordinator, N = 8School CCommercialSmallZealandMay/June 2022Physical activitySchool coordinator, N = 8School DTechnicalLargeCapital regionNov/Dec 2022Physical activitySchool coordinator, N = 9School ESocial healthcareLargeCapital regionNov/Dec 2022Physical activitySchool coordinator, N = 4, Department manager (due to staff turnover March 2024), N = 4School FAgri-culturalMediumZealandNov/Dec 2022Well-beingDepartment manager, N = 8School GCommercialMediumCapital regionNov/Dec 2022Well-beingSchool coordinator N = 5, Department manager (due to staff turnover May 2024), N = 3School HTechnicalLargeZealandNov/Dec 2022Physical activitySchool coordinator, N = 4, Department manager (due to changes in priorities January 2024), N = 4School size: Small = < 800 students, Medium = 800–2000 students, Large = > 2000 student Table 2Characteristics of study sitesEducational trackSchool sizeGeographical location (region)GMB was conductedTargeted health behaviourPrimary responsible person and number of interviews conductedSchool ASocial healthcareMediumZealandMay/June 2022Well-beingDepartment manager, N = 3School BAgri-culturalSmallCapital regionMay/June 2022Well-beingSchool coordinator, N = 8School CCommercialSmallZealandMay/June 2022Physical activitySchool coordinator, N = 8School DTechnicalLargeCapital regionNov/Dec 2022Physical activitySchool coordinator, N = 9School ESocial healthcareLargeCapital regionNov/Dec 2022Physical activitySchool coordinator, N = 4, Department manager (due to staff turnover March 2024), N = 4School FAgri-culturalMediumZealandNov/Dec 2022Well-beingDepartment manager, N = 8School GCommercialMediumCapital regionNov/Dec 2022Well-beingSchool coordinator N = 5, Department manager (due to staff turnover May 2024), N = 3School HTechnicalLargeZealandNov/Dec 2022Physical activitySchool coordinator, N = 4, Department manager (due to changes in priorities January 2024), N = 4School size: Small = < 800 students, Medium = 800–2000 students, Large = > 2000 student Characteristics of study sites School size: Small = < 800 students, Medium = 800–2000 students, Large = > 2000 student Actions related to the programme Table 3 shows the number of implemented and active actions during the two-years follow-up period. The number of actions increased over the two years, from nine actions at the three-month follow-up to 42 at 24-month. Numbers varied across sites, ranging from just one active action at School D to 12 at School F by 24-month follow-up. Examples of actions included the creation of a cosy student area (School A), the establishment of a student council (School H), and staff training on integrating PA into teaching, carried out in collaboration with a local sports association (School C). Further examples can be found in the two REM timeline illustrations in Appendix II (Schools G and H). The growth in the number of actions at most schools over time suggests that implementation continued beyond the initial programme phase and may indicate increasing local engagement and integration into existing school structures. Table 3Implemented and active actions during the two years of follow-up3month6month9month12month15month18month21month24monthSchool A155MissingMissingMissingMissingMissingSchool B13346678School C23443335School D00011111School E11222444School F358810111112School G13456856School H02577776Total922313135403842Missing= School A dropped out of the programme after nine months Table 3Implemented and active actions during the two years of follow-up3month6month9month12month15month18month21month24monthSchool A155MissingMissingMissingMissingMissingSchool B13346678School C23443335School D00011111School E11222444School F358810111112School G13456856School H02577776Total922313135403842Missing= School A dropped out of the programme after nine months Implemented and active actions during the two years of follow-up Missing= School A dropped out of the programme after nine months Indications of systems impact The perceived systems impact identified through analyses of the eight REM timelines are summarised in Table 4. The analyses revealed that no systems impact was experienced by informants at two study sites (Schools A (that withdraw after 9 months) and D). However, indications of systems impact were identified at the six other study sites (Schools B, C, E, F, G, and H), involving all or some components of structural, relational, and transformative change (defined in Table 1). Table 4Summary of identified systems impact and key factors of influence across schoolsSystems impactWhich types of systems impact (Structural, relational or transformative) were identified as a result of the programme?Influencing factorsWhich factors facilitated or hindered the systems impact of the programme?Enablers (E)Barriers (B)School A (withdrew after nine months) Missing/No impact identified during the first nine months B3: High staff turnoverB4: Low staff motivationB5: High manager turnoverB6: Low management supportB7: Limited financial resources within the municipalityB8: Limited financial resources within the schoolSchool BPolicies: No impact identifiedPractices: No impact identifiedResource flows: Granted external funding for multi-sport facility; Employment of more mental health councillorsRelationships & connections: Increased collaboration with community actorsPower dynamics: Increased management support; Increased involvement of students who contribute with inputs for actionsMental models: Staff changed mindsets about promoting students’ health at the schoolE1: Presence of structures so students can provide inputsE2: The research team as backbone supportE4: The Community of Practice networkE5: High prioritisation and allocation of resources among community actorsE6: School coordinator who was truly dedicated and passionate about creating changeE7: Financial resources provided by the research teamE9: High staff motivationE10: Perceived success of initial implementation and disseminationB1: Lack of time for coordinating actionsB6: Low management supportSchool CPolicies: PA became an integrated part of school schedulePractices: Capacity building among staff was implemented focusing on changing daily routinesResource flows: No impact identifiedRelationships & connections: Increased collaboration with community actorsPower dynamics: Increased involvement of students who contribute with inputs for actionsMental models: Increased focus on PA among staff and managerE1: Presence of structures so students can provide inputsE2: The research team as backbone supportE3: High management supportE6: School coordinator who was truly dedicated and passionate about creating changeE7: Financial resources provided by the research teamE8: Meeting community actorsE11: Knowledge and awareness raising among school staffE13: Perceived success of initial experiences with students participating in actionsB1: Lack of time for coordinating actionsB2: Changing and diverse student groups and short duration of coursesB5: High manager turnoverSchool D No impact identified B1: Lack of time for coordinating actionsB2: Changing and diverse student groups and short duration of coursesB3: High staff turnoverB4: Low staff motivationB9: Competing priorities: Renovation of the schoolB10: The culture at the school and at the workplaces where students do their internshipsSchool EPolicies: No impact identifiedPractices: No impact identifiedResource flows: The school made it a priority to allocate resources for actions; PA became a regular topic on the agenda at staff meetings; Continuous informal exchange of knowledge among the school staffRelationships & connections: No impact identifiedPower dynamics: Increased inputs from students; Increased staff involvement in decision makingMental models: PA became an overarching focus at the schoolE1: Presence of structures so students can provide inputsE3: High management supportE4: The Community of Practice networkE9: High staff motivationB1: Lack of time for coordinating actionsB2: Changing and diverse student groups and short duration of coursesB3: High staff turnoverB11: Curricular pressuresSchool FPolicies: Mental health was integrated into both action plans and policy documentsPractices: Capacity building among staff was implemented focusing on changing daily routines; One staff member became ambassador for mental healthResource flows: Contributed to expansion to other school departments by knowledge sharing; One staff member was allocated with dedicated time to work on mental health promotion initiatives as a consistent area of focusRelationships & connections: Increased collaboration between councillors, psychologists etc.; A collaboration with the municipality was establishedPower dynamics: Increased staff involvement in decision making; Continuous involvement of students who contributed with inputs for actionsMental models: Mental health became an overarching focus at the schoolE2: The research team as backbone supportE3: High management supportE5: High prioritisation and allocation of resources among community actorsE6: School coordinator who was truly dedicated and passionate about creating changeE7: Financial resources provided by the research teamE8: Meeting community actorsE10: Perceived success of initial implementation and disseminationE14: Financial resources provided by the municipalityB1: Lack of time for coordinating actionsB2: Changing and diverse student groups and short duration of coursesSchool GPolicies: Mental health was integrated into both action plans and policy documents; phone free teaching was written into school policies, action plans, and student handbookPractices: Capacity building among staff was implemented focusing on changing daily routines; Changes in assignment deadlinesResource flows: The school made it a priority to allocate resources for actions; Contributed to expansion to other school departments by knowledge sharingRelationships & connections: An established partnership between the school and a community actorPower dynamics: Increased staff involvement in decision-making and implementation; Increased involvement of students who contributed with inputs for actionsMental models: Promoting students wellbeing became a key focus among school staff and the managerE1: Presence of structures so students can provide inputsE2: The research team as backbone supportE3: High management supportE4: The Community of Practice networkE5: High prioritisation and allocation of resources among community actorsE8: Meeting community actorsE12: Department manager who was truly dedicated and passionate about creating changeB1: Lack of time for coordinating actionsB2: Changing and diverse student groups and short duration of coursesB3: High staff turnoverB12: Lack of student support for actionsB13: Sick leave among managementB14: A group of students who were challenging to manage and required significant resourcesSchool HPolicies: No impact identifiedPractices: Capacity building among staff was implemented focusing on changing daily routines; Expanding PA initiatives to staff membersResource flows: Contributed to the expansion to other school departments; Continuously allocating more resources to PARelationships & connections: Increased collaboration with community actorsPower dynamics: Increased staff involvement in decision-making and implementation; Established involvement of students who contributed with inputs and implementation of actions; Increased knowledge sharing among staffMental models: PA became a natural part of students’ and staffs’ daily life at the schoolE1: Presence of structures so students can provide inputsE2: The research team as backbone supportE3: High management supportE4: The Community of Practice networkE5: High prioritisation and allocation of resources among community actorsE8: Meeting community actorsE9: High staff motivationE11: Knowledge and awareness raising among school staffE12: Department manager who was truly dedicated and passionate about creating changeB1: Lack of time for coordinating actionsB2: Changing and diverse student groups and short duration of coursesCodes for enablers (E) and barriers (B) were assigned in order of frequency, for example E1 represents the most frequently identified enabler across schools Table 4Summary of identified systems impact and key factors of influence across schoolsSystems impactWhich types of systems impact (Structural, relational or transformative) were identified as a result of the programme?Influencing factorsWhich factors facilitated or hindered the systems impact of the programme?Enablers (E)Barriers (B)School A (withdrew after nine months) Missing/No impact identified during the first nine months B3: High staff turnoverB4: Low staff motivationB5: High manager turnoverB6: Low management supportB7: Limited financial resources within the municipalityB8: Limited financial resources within the schoolSchool BPolicies: No impact identifiedPractices: No impact identifiedResource flows: Granted external funding for multi-sport facility; Employment of more mental health councillorsRelationships & connections: Increased collaboration with community actorsPower dynamics: Increased management support; Increased involvement of students who contribute with inputs for actionsMental models: Staff changed mindsets about promoting students’ health at the schoolE1: Presence of structures so students can provide inputsE2: The research team as backbone supportE4: The Community of Practice networkE5: High prioritisation and allocation of resources among community actorsE6: School coordinator who was truly dedicated and passionate about creating changeE7: Financial resources provided by the research teamE9: High staff motivationE10: Perceived success of initial implementation and disseminationB1: Lack of time for coordinating actionsB6: Low management supportSchool CPolicies: PA became an integrated part of school schedulePractices: Capacity building among staff was implemented focusing on changing daily routinesResource flows: No impact identifiedRelationships & connections: Increased collaboration with community actorsPower dynamics: Increased involvement of students who contribute with inputs for actionsMental models: Increased focus on PA among staff and managerE1: Presence of structures so students can provide inputsE2: The research team as backbone supportE3: High management supportE6: School coordinator who was truly dedicated and passionate about creating changeE7: Financial resources provided by the research teamE8: Meeting community actorsE11: Knowledge and awareness raising among school staffE13: Perceived success of initial experiences with students participating in actionsB1: Lack of time for coordinating actionsB2: Changing and diverse student groups and short duration of coursesB5: High manager turnoverSchool D No impact identified B1: Lack of time for coordinating actionsB2: Changing and diverse student groups and short duration of coursesB3: High staff turnoverB4: Low staff motivationB9: Competing priorities: Renovation of the schoolB10: The culture at the school and at the workplaces where students do their internshipsSchool EPolicies: No impact identifiedPractices: No impact identifiedResource flows: The school made it a priority to allocate resources for actions; PA became a regular topic on the agenda at staff meetings; Continuous informal exchange of knowledge among the school staffRelationships & connections: No impact identifiedPower dynamics: Increased inputs from students; Increased staff involvement in decision makingMental models: PA became an overarching focus at the schoolE1: Presence of structures so students can provide inputsE3: High management supportE4: The Community of Practice networkE9: High staff motivationB1: Lack of time for coordinating actionsB2: Changing and diverse student groups and short duration of coursesB3: High staff turnoverB11: Curricular pressuresSchool FPolicies: Mental health was integrated into both action plans and policy documentsPractices: Capacity building among staff was implemented focusing on changing daily routines; One staff member became ambassador for mental healthResource flows: Contributed to expansion to other school departments by knowledge sharing; One staff member was allocated with dedicated time to work on mental health promotion initiatives as a consistent area of focusRelationships & connections: Increased collaboration between councillors, psychologists etc.; A collaboration with the municipality was establishedPower dynamics: Increased staff involvement in decision making; Continuous involvement of students who contributed with inputs for actionsMental models: Mental health became an overarching focus at the schoolE2: The research team as backbone supportE3: High management supportE5: High prioritisation and allocation of resources among community actorsE6: School coordinator who was truly dedicated and passionate about creating changeE7: Financial resources provided by the research teamE8: Meeting community actorsE10: Perceived success of initial implementation and disseminationE14: Financial resources provided by the municipalityB1: Lack of time for coordinating actionsB2: Changing and diverse student groups and short duration of coursesSchool GPolicies: Mental health was integrated into both action plans and policy documents; phone free teaching was written into school policies, action plans, and student handbookPractices: Capacity building among staff was implemented focusing on changing daily routines; Changes in assignment deadlinesResource flows: The school made it a priority to allocate resources for actions; Contributed to expansion to other school departments by knowledge sharingRelationships & connections: An established partnership between the school and a community actorPower dynamics: Increased staff involvement in decision-making and implementation; Increased involvement of students who contributed with inputs for actionsMental models: Promoting students wellbeing became a key focus among school staff and the managerE1: Presence of structures so students can provide inputsE2: The research team as backbone supportE3: High management supportE4: The Community of Practice networkE5: High prioritisation and allocation of resources among community actorsE8: Meeting community actorsE12: Department manager who was truly dedicated and passionate about creating changeB1: Lack of time for coordinating actionsB2: Changing and diverse student groups and short duration of coursesB3: High staff turnoverB12: Lack of student support for actionsB13: Sick leave among managementB14: A group of students who were challenging to manage and required significant resourcesSchool HPolicies: No impact identifiedPractices: Capacity building among staff was implemented focusing on changing daily routines; Expanding PA initiatives to staff membersResource flows: Contributed to the expansion to other school departments; Continuously allocating more resources to PARelationships & connections: Increased collaboration with community actorsPower dynamics: Increased staff involvement in decision-making and implementation; Established involvement of students who contributed with inputs and implementation of actions; Increased knowledge sharing among staffMental models: PA became a natural part of students’ and staffs’ daily life at the schoolE1: Presence of structures so students can provide inputsE2: The research team as backbone supportE3: High management supportE4: The Community of Practice networkE5: High prioritisation and allocation of resources among community actorsE8: Meeting community actorsE9: High staff motivationE11: Knowledge and awareness raising among school staffE12: Department manager who was truly dedicated and passionate about creating changeB1: Lack of time for coordinating actionsB2: Changing and diverse student groups and short duration of coursesCodes for enablers (E) and barriers (B) were assigned in order of frequency, for example E1 represents the most frequently identified enabler across schools Summary of identified systems impact and key factors of influence across schools B3: High staff turnover B4: Low staff motivation B5: High manager turnover B6: Low management support B7: Limited financial resources within the municipality B8: Limited financial resources within the school Policies: No impact identified Practices: No impact identified Resource flows: Granted external funding for multi-sport facility; Employment of more mental health councillors Relationships & connections: Increased collaboration with community actors Power dynamics: Increased management support; Increased involvement of students who contribute with inputs for actions Mental models: Staff changed mindsets about promoting students’ health at the school E1: Presence of structures so students can provide inputs E2: The research team as backbone support E4: The Community of Practice network E5: High prioritisation and allocation of resources among community actors E6: School coordinator who was truly dedicated and passionate about creating change E7: Financial resources provided by the research team E9: High staff motivation E10: Perceived success of initial implementation and dissemination B1: Lack of time for coordinating actions B6: Low management support Policies: PA became an integrated part of school schedule Practices: Capacity building among staff was implemented focusing on changing daily routines Resource flows: No impact identified Relationships & connections: Increased collaboration with community actors Power dynamics: Increased involvement of students who contribute with inputs for actions Mental models: Increased focus on PA among staff and manager E1: Presence of structures so students can provide inputs E2: The research team as backbone support E3: High management support E6: School coordinator who was truly dedicated and passionate about creating change E7: Financial resources provided by the research team E8: Meeting community actors E11: Knowledge and awareness raising among school staff E13: Perceived success of initial experiences with students participating in actions B1: Lack of time for coordinating actions B2: Changing and diverse student groups and short duration of courses B5: High manager turnover B1: Lack of time for coordinating actions B2: Changing and diverse student groups and short duration of courses B3: High staff turnover B4: Low staff motivation B9: Competing priorities: Renovation of the school B10: The culture at the school and at the workplaces where students do their internships Policies: No impact identified Practices: No impact identified Resource flows: The school made it a priority to allocate resources for actions; PA became a regular topic on the agenda at staff meetings; Continuous informal exchange of knowledge among the school staff Relationships & connections: No impact identified Power dynamics: Increased inputs from students; Increased staff involvement in decision making Mental models: PA became an overarching focus at the school E1: Presence of structures so students can provide inputs E3: High management support E4: The Community of Practice network E9: High staff motivation B1: Lack of time for coordinating actions B2: Changing and diverse student groups and short duration of courses B3: High staff turnover B11: Curricular pressures Policies: Mental health was integrated into both action plans and policy documents Practices: Capacity building among staff was implemented focusing on changing daily routines; One staff member became ambassador for mental health Resource flows: Contributed to expansion to other school departments by knowledge sharing; One staff member was allocated with dedicated time to work on mental health promotion initiatives as a consistent area of focus Relationships & connections: Increased collaboration between councillors, psychologists etc.; A collaboration with the municipality was established Power dynamics: Increased staff involvement in decision making; Continuous involvement of students who contributed with inputs for actions Mental models: Mental health became an overarching focus at the school E2: The research team as backbone support E3: High management support E5: High prioritisation and allocation of resources among community actors E6: School coordinator who was truly dedicated and passionate about creating change E7: Financial resources provided by the research team E8: Meeting community actors E10: Perceived success of initial implementation and dissemination E14: Financial resources provided by the municipality B1: Lack of time for coordinating actions B2: Changing and diverse student groups and short duration of courses Policies: Mental health was integrated into both action plans and policy documents; phone free teaching was written into school policies, action plans, and student handbook Practices: Capacity building among staff was implemented focusing on changing daily routines; Changes in assignment deadlines Resource flows: The school made it a priority to allocate resources for actions; Contributed to expansion to other school departments by knowledge sharing Relationships & connections: An established partnership between the school and a community actor Power dynamics: Increased staff involvement in decision-making and implementation; Increased involvement of students who contributed with inputs for actions Mental models: Promoting students wellbeing became a key focus among school staff and the manager E1: Presence of structures so students can provide inputs E2: The research team as backbone support E3: High management support E4: The Community of Practice network E5: High prioritisation and allocation of resources among community actors E8: Meeting community actors E12: Department manager who was truly dedicated and passionate about creating change B1: Lack of time for coordinating actions B2: Changing and diverse student groups and short duration of courses B3: High staff turnover B12: Lack of student support for actions B13: Sick leave among management B14: A group of students who were challenging to manage and required significant resources Policies: No impact identified Practices: Capacity building among staff was implemented focusing on changing daily routines; Expanding PA initiatives to staff members Resource flows: Contributed to the expansion to other school departments; Continuously allocating more resources to PA Relationships & connections: Increased collaboration with community actors Power dynamics: Increased staff involvement in decision-making and implementation; Established involvement of students who contributed with inputs and implementation of actions; Increased knowledge sharing among staff Mental models: PA became a natural part of students’ and staffs’ daily life at the school E1: Presence of structures so students can provide inputs E2: The research team as backbone support E3: High management support E4: The Community of Practice network E5: High prioritisation and allocation of resources among community actors E8: Meeting community actors E9: High staff motivation E11: Knowledge and awareness raising among school staff E12: Department manager who was truly dedicated and passionate about creating change B1: Lack of time for coordinating actions B2: Changing and diverse student groups and short duration of courses Codes for enablers (E) and barriers (B) were assigned in order of frequency, for example E1 represents the most frequently identified enabler across schools Structural impacts included formal changes such as the integration of PA into school schedules (School C), incorporation of mental health into policies and action plans (Schools F and G), and prioritisation of PA in staff meetings (School E). In four study sites (Schools C, F, G, and H), capacity building among staff supported changes in daily practices, such as integrating PA or wellbeing into routines and teaching. Relational impacts were identified at the five study sites, where collaboration with especially community actors was experienced to increase during the programme (Schools B, C, F, G, and H). For example, School G formalised a partnership with a community organisation, while School F strengthened its collaboration with local mental health professionals. In addition to strengthened external collaboration, changes in internal power dynamics were identified at six study sites. These changes were characterised by increased involvement of both staff and students in decision-making processes, resulting in increased distributed leadership. Transformative impacts were also identified at the six study sites, including shifts in how the staff perceived their roles in health promotion efforts. For example, at School B, the staff began to view students’ health and wellbeing as one of their core responsibilities. Meanwhile, at School H, PA was described as an integrated part of daily life for both students and staff. Factors contributing to systems impact Table 4 summarises the perceived factors influencing the experienced impact, categorised as either enablers or barriers. We identified several enablers which were experienced to support change across the study sites. Enabling factors included the presence of a dedicated and passionate school coordinator or department manager (Schools B, C, F, G, and H), knowledge and awareness among school staff (Schools C and H), motivated staff (Schools B and H), and consistent management support (Schools C, E, F, G, and H). Structures that provided students with opportunities to share their views and contribute with ideas, such as student councils, were also important across several study sites (Schools B, C, E, G, and H). Financial resources seemed to play an important role in enabling actions to be implemented and in supporting their potential to contribute to impact, either through the programme or via external funding from municipalities or community organisations (Schools B, C, F, G, and H). The opportunity to meet and engage with community actors (Schools C, F, G, and H) and participate in the CoP (Schools B, E, G, and H) were also found to be key enablers. Some participants highlighted the ongoing communication with the research team as an important factor in supporting the implementation of actions, serving as an enabling factor (Schools B, C, F, G, and H). Lastly, experiencing continuous success with the implementation and dissemination of actions was found to be an enabling factor to support change (Schools B and F). In contrast, several factors were identified that may have constrained the potential for systems impact. A significant barrier identified across multiple study sites was the resignation of key school actors, particularly school coordinators or department managers, which disrupted continuity in implementation and ownership (Schools A, C, E, and G). In addition, lack of time to coordinate actions, along with competing priorities, such as school renovations or teaching obligations, were experienced as barriers among all study sites and limited the ability to prioritise implementation. Lastly, the transient nature of vocational students, where students frequently enter and leave due to high dropout rates and workplace-based training, seemed to limit the implementation of actions developed by students during the GMB process and their potential to contribute to systems impact (Schools C, E, and G). Many of these barriers appeared to be particularly prominent at the study sites where no systems impact was identified (School D), as well as at the school that withdrew (School A). Pathways contributing to perceived systems impact Figure 2 illustrates the identified plausible pathways which might contribute to systems impact, as well as contextual and programme-specific influences. We identified two plausible impact pathways, characterised as two reinforcing feedback loops: one was school-driven changes to distribute leadership (R1), and the other was community-supported changes to change daily practices (R2). Although the identified mechanisms have influenced multiple aspects of systems impact, we have only highlighted the two most prominent ones. We acknowledge that all impact components are interrelated, as illustrated by the arrows between the systems impact conditions. Additionally, while the two pathways could function independently, as observed in one study site (School E), where change was solely school-driven, they also influenced and reinforced each other, thereby maybe contribute to systems impact, as observed in more study sites (Schools B, C, F, G, and H). Fig. 2Plausible pathways contributing to systems impact for health promotion in vocational schools Plausible pathways contributing to systems impact for health promotion in vocational schools School-driven change to distribute leadership (R1) This pathway was driven by committed school staff or managers who demonstrated a strong passion for and took active ownership of the health promotion agenda, resulting in distributed leadership. Their commitment and leadership encouraged motivation among school staff and implementation of actions, which led to early successful experiences with implementation. These successes reinforced staff engagement and, in some cases, resulted in staff being recognised as health promotion champions. Student involvement (supported by structures such as a student council) also played a central role in motivating staff in this process, as it contributed to the relevance and legitimacy of school-based actions. Several contextual factors were found to be key in driving this feedback loop, particularly strong management support, dedicated time allocation, and school resource availability. The allocation of time enabled staff to plan, coordinate and implement health-promoting actions alongside their teaching duties. Similarly, access to financial resources supported the initiation and maintenance of these actions over time. However, high staff turnover in the vocational school setting and the changing and diverse nature of the student populations (due to workplace-based training) challenged continuity and momentum. We also identified key contributions of the Data Health programme that supported this pathway to impact. These included the facilitation of the CoP network, which enabled school and municipal actors to share experiences and knowledge, the research team’s role as backbone support, and the provision of financial resources to enable local action. Community-supported change to change daily practices (R2) This pathway centred on the involvement of community actors in driving change in daily practices among staff and students. The Data Health programme enabled this by creating opportunities for direct engagement with these actors, mainly during GMB3. Through collaboration, community actors often contributed to capacity building among school staff, enhancing their awareness, knowledge, and motivation. Positive experiences with externally supported actions reinforced the value of collaboration and, in some study sites, led to the development of equal partnerships between school and community actors. These external actors provided resources, expertise, and legitimacy, which enabled schools to access wider support and change daily practices among staff and students. Discussion This study aimed to explore the implementation, perceived systems impact, and plausible impact pathways of a health promotion programme based on systems thinking in Danish vocational schools. Over the two-year follow-up period, six of the eight study sites implemented actions that might have contributed to systems impact, exhibiting signs of structural, relational, and/or transformative change. Across the study sites, a pattern emerged whereby schools implementing a greater number of actions were more likely to experience wider systems impact. At the same time, we recognise that not all actions contribute equally to systems change, rendering the number of actions less relevant. This point was also emphasised by Nobles et al. (2022), who categorised systems impact into event-based actions, structures, goals, and those that aim to influence deeper underlying beliefs of a system. Therefore, a high number of actions at the event level may not generate the same depth of sustained impact as fewer, more strategic actions targeting the deeper levels of a system [22]. The findings indicate that the programme might have triggered systems changes across the six study sites that were sustained beyond the formal programme period. The continued increase in the number of actions over time suggests that implementation did not cease when external support was withdrawn but continued within the schools, reflecting strong local ownership and integration into existing school structures. We identified two plausible impact pathways that may explain this induced change. These aimed to (1) strengthen distributed leadership and (2) transform everyday school practices through collaboration with external actors. These pathways consisted of interconnected enabling mechanisms linked to the implementation of actions for change, shaped by both programme-related and contextual factors. Many of the identified mechanisms align with those reported in implementation science and educational research, such as leadership support, staff engagement, and access to resources [56, 57]. The contribution of this study lies in demonstrating how these mechanisms function within a whole-school approach and support the emergence of systems impact, addressing a gap as such mechanisms have been underexplored in vocational school settings. Enabling distributed leadership for health promotion The first pathway emerged through mechanisms situated internally at the schools, which affected the power dynamics to enhance distributed leadership. As noted by Rowling and Samdal (2011), distributed leadership structures are vital to fostering collective responsibility and embedding health promotion into the daily practices and culture of the school [58]. The overall aim of the Data Health programme was to strengthen the capacity, knowledge, and engagement of school actors (including students, staff, and managers), as well as municipal and community actors, so they could take active roles in implementing health promotion actions [38, 47]. The GMB process in the Data Health programme was designed to create a space for distributed leadership to be explored and determined. The original intention was for the working groups established during the GMB process to take collective responsibility for coordinating and implementing the co-developed actions [38, 47]. At none of the study sites the working groups took responsibility for coordinating or carrying out the planned actions. Notably, none of the actions led solely by student groups were carried out. Frequent student turnover due to workplace-based training made this difficult. Similarly, other health promotion interventions within Danish vocational schools have identified school structure and frequent student transitions as key barriers to implementation [59, 60]. Instead, we found that one or two committed individuals (often the school coordinator and/or department manager) took the initiative and responsibility for coordinating and implementing actions during the period after the GMB process. These individuals acted as important change agents, serving not only as internal connectors across the school but also as important links between the school and external community actors. This finding is consistent with recent studies which highlight the importance of local implementation champions leading and supporting school-based health promotion efforts [61, 62]. However, relying on only a few key individuals makes the process vulnerable, particularly in a setting with high staff and management turnover [10]. High staff turnover is a pre-existing condition that must be acknowledged when working with health promotion within the setting of vocational schools. The results of this study showed that committed staff or managers within vocational school settings often served as catalysts for broader staff engagement over time. In addition, although students did not play an active role in implementing actions, their perspectives and involvement in decision-making (supported by a student council) had a positive influence on staff motivation to participate actively. Similar findings have been reported in the implementation of HPS principles in primary school settings, where student participation has been shown to strengthen relevance and engagement. For example, Keshavarz et al. (2010) showed that formal structures for student participation, such as student councils, helped to implement HPS initiatives. However, maintaining a functioning student council in vocational schools can be challenging due to the transient nature of the student group. Consequently, some vocational schools do not have such structures in place [63]. We also identified some aspects related to the programme which contributed positively to transforming the role of distributed leadership. First, access to financial resources, both from the programme budget and from internal school funding, played an important role in enabling the implementation and prioritisation of health-promoting actions. Previous research on HPS approaches has also identified financial resources as a crucial factor for successful implementation (e.g [64]), . The WHO emphasises the importance of flexible funding models that can be adapted to meet the specific needs of individual schools [65]. Second, the CoP, which was established and facilitated by the research team, was found to support commitment among school coordinators and/or department managers. Other studies have similarly highlighted the value of decentralised, practice-oriented networks supporting implementation related to health promotion [66, 67]. Finally, the research team, acting as backbone support, served as a continuous reminder of the importance of prioritising health promotion. Findings from other systems-based studies support this finding by highlighting backbone support as a critical factor in maintaining focus, sustaining momentum [68, 69], and achieving systems impact [70]. The importance of establishing supportive structures to maintain focus and strengthen implementation in complex and dynamic school settings has been well-documented in previous studies [61, 67, 69, 71, 72]. The role of community actors in supporting systems change for health promotion The second pathway emerged through collaboration between schools and community actors, who played an important role in supporting capacity building among the school staff in changing daily practices related to either wellbeing or PA. Many of the community actors who emerged as active change agents within the programme were professionals affiliated with local or regional non-governmental organisations, such as the Danish Gymnastics and Sports Association or the Just Human Foundation [47]. Their expertise and existing responsibilities in areas such as youth engagement and health promotion might have enabled them to contribute to the planning and implementation of health promotion actions in the setting of vocational schools. This finding supports the HPS approach’s argument that external change agents may be better positioned than internal school actors to drive system-level changes in school culture and norms by leveraging external legitimacy [50, 73]. Only one school (School F) managed to build a sustained collaboration with the municipality at the two-year follow-up, even though this had been an intended part of our programme [38]. This school also implemented the highest number of actions and experienced a broad spectrum of systems impact. While this may be coincidental, this relationship could also indicate that active collaboration with the municipality fosters implementation and contributes to broader systems changes. Results from the process evaluation of the GMB process also suggested that municipal actors played an important role as connectors between schools and the broader community [47]. Bartelink et al. (2022) explored the role of external health promotion professionals in relation to HPS implementation. In alignment with our results, the authors found that such a person can play an important role in encouraging schools to prioritise health, guiding the implementation process, offering support, and acting as a link between the school and the wider community [72]. Currently, only few larger municipalities in Denmark have adopted such structures and are actively involved in promoting health in vocational schools [74, 75]. We encourage future research and practice to explore further how diverse municipalities in Denmark can take on a more defined and strategic role in supporting health promotion in vocational schools, drawing on international experiences [21, 72, 76, 77]. A tendency emerged in the data indicating that study sites which increased collaboration with external community actors during the study period also implemented a higher number of actions and generally experienced broader systems impact. This might suggest that collaboration with external actors supported both continued implementation and capacity building among staff, which in turn enabled the conditions for distributed leadership to appear within the schools. The two interconnected feedback loops identified reflect the core aim of systems approaches: creating conditions that encourage further change over time [23]. Implications for research and practice The identified plausible impact pathways may serve as leverage points for fostering sustainable and embedded systems changes for health promotion in Danish vocational schools or similar settings. However, the withdrawal of one school from the programme and the evaluation raises the question of whether systems-based health promotion programmes are appropriate for all Danish vocational schools, particularly those with limited capacity or resources. Notably, the withdrawn school was a social and healthcare school. This type of institution has previously been found to have staff who score significantly higher on measures of health promotion capacity compared to staff in other educational tracks [63]. Similar relations between this type of school and strong implementation capacity have also been found in other studies [59, 78]. Therefore, we assume that the withdrawal was not primarily due to a lack of capacity but rather the result of a combination of contextual factors and structural changes. During the study period, the department manager at the school and the municipal coordinator resigned, and the school coordinator was assigned a different position. These shifts emphasise the importance of understanding vocational schools as complex systems that constantly evolve over time and underscore the need for distributed leadership. Based on the findings of this study, we identified several implications for future systems-based practice in vocational schools, relevant to both researchers and practitioners committed to fostering sustainable change within vocational school settings: Acknowledge vocational schools as complex adaptive systems: High staff turnover, fluctuating student populations, and curricular pressures challenge implementation and systems impact. Approaches to health promotion in vocational schools must be flexible and tailored to the contextual realities, and systems approaches may be a viable approach.Strengthen structures for student participation: Despite high student turnover in vocational schools, it is possible to create supportive structures for student involvement. Involving students in decision-making processes related to health promotion can motivate staff and may reinforce broader systems change.Invest in distributed staff capacity building: Systems change depended on committed staff and managers acting as local change agents across the organisation. Professional development, shared responsibility, and collaboration with external actors may strengthen the capacity to sustain actions beyond the initial programme phase.Support cross-sector collaboration and municipal coordination: Municipal actors can play an important coordinating role by connecting vocational schools with community organisations and aligning health promotion efforts across sectors. Strengthening these relationships may support sustained implantation and systems change.Establish long-term supportive infrastructure: Providing ongoing support to schools is crucial for maintaining focus and driving progress over time. Relevant structures might include ongoing backbone support and facilitating CoP networks, roles that municipalities might be well placed to fulfil in vocational school settings. Acknowledge vocational schools as complex adaptive systems: High staff turnover, fluctuating student populations, and curricular pressures challenge implementation and systems impact. Approaches to health promotion in vocational schools must be flexible and tailored to the contextual realities, and systems approaches may be a viable approach. Acknowledge vocational schools as complex adaptive systems: High staff turnover, fluctuating student populations, and curricular pressures challenge implementation and systems impact. Approaches to health promotion in vocational schools must be flexible and tailored to the contextual realities, and systems approaches may be a viable approach. Strengthen structures for student participation: Despite high student turnover in vocational schools, it is possible to create supportive structures for student involvement. Involving students in decision-making processes related to health promotion can motivate staff and may reinforce broader systems change. Strengthen structures for student participation: Despite high student turnover in vocational schools, it is possible to create supportive structures for student involvement. Involving students in decision-making processes related to health promotion can motivate staff and may reinforce broader systems change. Invest in distributed staff capacity building: Systems change depended on committed staff and managers acting as local change agents across the organisation. Professional development, shared responsibility, and collaboration with external actors may strengthen the capacity to sustain actions beyond the initial programme phase. Invest in distributed staff capacity building: Systems change depended on committed staff and managers acting as local change agents across the organisation. Professional development, shared responsibility, and collaboration with external actors may strengthen the capacity to sustain actions beyond the initial programme phase. Support cross-sector collaboration and municipal coordination: Municipal actors can play an important coordinating role by connecting vocational schools with community organisations and aligning health promotion efforts across sectors. Strengthening these relationships may support sustained implantation and systems change. Support cross-sector collaboration and municipal coordination: Municipal actors can play an important coordinating role by connecting vocational schools with community organisations and aligning health promotion efforts across sectors. Strengthening these relationships may support sustained implantation and systems change. Establish long-term supportive infrastructure: Providing ongoing support to schools is crucial for maintaining focus and driving progress over time. Relevant structures might include ongoing backbone support and facilitating CoP networks, roles that municipalities might be well placed to fulfil in vocational school settings. Establish long-term supportive infrastructure: Providing ongoing support to schools is crucial for maintaining focus and driving progress over time. Relevant structures might include ongoing backbone support and facilitating CoP networks, roles that municipalities might be well placed to fulfil in vocational school settings. Strengths, limitations, and future research This study contributes with knowledge on the implementation and perceived impact of a systems approach in vocational schools, an overlooked setting in health promotion research and practice. A strength of this study is the longitudinal design, with data collected quarterly over a two-year period. This enabled a dynamic view of how actions evolved over time and how impact gradually unfolded across the different levels represented within the Six Conditions of Systems Change framework. Further, the longitudinal design allowed us to connect implementation processes with perceived impact as they unfolded in real time, offering insights that are not easily captured through more traditional qualitative designs. In doing so, the current study identified plausible mechanisms that enable impact across more sites, extending beyond the conventional concepts of implementation facilitators and barriers [77, 79]. This allowed us to move beyond isolated case descriptions and provide early empirical contributions towards understanding mechanisms of change in systems-based health promotion in vocational schools. However, it is important to consider the following limitations. First, because complex adaptive systems constantly evolve, our findings capture a snapshot of the changes that occurred during the study period rather than providing a comprehensive account of all actions or impacts. The data in this study were generated through interviews with participants and shaped by researchers as part of the evaluation. However, broader external conditions may have influenced both implementation and outcomes. At some study sites, especially those targeting wellbeing, the programme fitted well with existing national and regional health priorities, for example, the newly established Commission for Children and Youth Wellbeing [80]. This overlap may suggest that broader societal momentum may have affected and supported the identified impact. However, as other researchers have noted, the goal of a systems-based evaluation is not to capture every possible influence or effect related to a specific initiative but rather to illustrate a range of actions and impacts within the scope of the evaluation [27, 29]. In line with this, we also found that it was not always possible to draw clear distinctions between actions and impacts. For instance, staff capacity building acted both as a concrete action and as an impact, reflected in shifts in everyday practice. Secondly, we originally intended to collect data for this study through REM workshops with action groups across both school, municipal, and community actors [38]. However, as these groups did not materialise in practice, we chose to collect data from individual interviews with the person responsible for implementation at each school instead. Therefore, the data only reflected one perspective per study site. This may have introduced social desirability bias, as the participants were often closely involved in coordinating the actions. While the study draws on principles from REM, the mapping process was researcher-driven rather than fully participatory, as originally intended within the REM approach [36]. Consequently, the visualisations were not co-constructed with participants in real time, which may have limited opportunities for participants to validate interpretations and contribute additional perspectives. This may have implications for interpretive validity and the depth of insights. However, we sought to mitigate this by grounding the analysis closely in longitudinal interview data and through iterative cross-case comparisons. Finally, the researchers were closely involved in the programme and evaluation process, which may have introduced a risk of confirmation bias and influenced the interpretation of findings. At the same time, this close involvement provided in-depth contextual understanding and enabled continuous insights into how implementation and perceived impact unfolded across study sites over time. We therefore consider this dual role to represent both a limitation and a strength when evaluating systems-oriented programmes in complex adaptive settings. To our knowledge, the current study is the first to provide insights into potential pathways to systems change in vocational schools and in school settings more broadly. Future studies could benefit from exploring enabling mechanisms more broadly across the system by applying participatory methods, such as REM [29] or Realist REM [81], to capture diverse perspectives from a wider range of school and community actors. Future studies could replicate these pathways by providing deeper qualitative insights into the key mechanisms identified. In addition, future research may benefit from examining systems impact in relation to student-level outcomes (such as wellbeing or PA behaviour) to better understand the links between individual- and system-level change. Conclusion This study explored the implementation and systems impact of a health promotion programme grounded in systems thinking across Danish vocational schools, with the aim of identifying plausible pathways to systems change. Over the two-year follow-up period, systems change occurred in six out of eight study sites, exhibiting signs of structural, relational, and/or transformative change that appeared to be sustainably embedded within these schools. These changes were enabled by the implementation of local actions shaped by both contextual and programme-related factors. Although the specific actions varied across schools, the mechanisms activated showed considerable overlap. Two central and mutually reinforcing plausible pathways to impact emerged: one centred on internally driven change through the development of distributed leadership, and the other on externally supported capacity building that influenced staff practices in relation to health promotion. Our findings underscore the importance of recognising the dynamic context of vocational schools, which are characterised by high staff and student turnover and competing educational priorities. These findings offer valuable insights for the future adaptation of the Data Health programme and for informing broader systems-based strategies in vocational school settings. Supplementary Information Supplementary Material 1. Supplementary Material 1. Supplementary Material 2. Supplementary Material 2.