Understanding the dynamics of implementing the Norwegian School Fruit Programme: integrating implementation science and system dynamics Authors: Mahshid Zolfaghari, Nanna Lien, Caroline Løvik Brandvik, Knut-Inge Klepp, Birgit Kopainsky, Hanne C. Lie, Biljana Meshkovska Published: 2025 ABSTRACT BackgroundMany implementation determinants of school nutritional interventions have been identified, but a deeper understanding of their interplay within a system is needed. The present study aimed to investigate how the Norwegian School Fruit Programme, particularly the No Parental Payment Programme (NPPP), is being implemented in schools and how it interacts with the school system structure and dynamics as perceived by school personnel and pupils. Moreover, it aimed to understand the dynamic interplay between the determinants of the Consolidated Framework for Implementation Research (CFIR) in the NPPP’s context.MethodThe overarching methodology was qualitative system dynamics modelling, with data collected through a series of group model building (GMB) workshops at four schools offering the NPPP. Using GMB scripts, facilitated discussions between school personnel and pupils were held for a better understanding of the NPPP’s implementation dynamics and to create diagrams illustrating how implementation determinants interact with each other and each school’s unique context. The individual school diagrams were compiled into a single diagram. The CFIR was subsequently mapped on to this diagram.ResultsLack of communication with pupils, the food stores outside schools, and the NPPP’s design weakened the reinforcing mechanism that fostered pupils’ eating of fruit. Not eating enough of the fruit increased implementation costs because schools had to deal with leftover fruit and littering issues. Implementation costs caused schools to adopt implementation processes, like reflecting, engaging, and adapting to reduce the costs, although some of these did not work or counterintuitively decreased pupils’ eating of fruit. The success of these processes was contested by the design of the NPPP and mechanisms that were shaped by the goals and priorities of the schools.ConclusionsThis research employed qualitative system dynamics to investigate NPPP implementation dynamics. Recommendations for improving implementation include aligning NPPP’s design with pupils’ preferences, monitoring eating of fruit, and involving internal stakeholders (pupils and school personnel) for gathering feedback, and external stakeholders (municipalities) for monitoring.Supplementary InformationThe online version contains supplementary material available at 10.1186/s43058-025-00834-6. The present study underlines the value of systems sciences methods in studying the structure and dynamics of implementation interventions, leveraging the synergy between these fields to generate new insights.It identifies underlying mechanisms and cultivates causal thinking by eliciting the perspectives of pupils and school personnel on the Norwegian School Fruit Programme implementation dynamics, thereby contributing to theory building.Using the school fruit programme as a case, the present study situates implementation determinants within a systems narrative, for example, it demonstrated that implementation processes are feedback loops intended to lower the costs associated with limited innovation reach. The present study underlines the value of systems sciences methods in studying the structure and dynamics of implementation interventions, leveraging the synergy between these fields to generate new insights. The present study underlines the value of systems sciences methods in studying the structure and dynamics of implementation interventions, leveraging the synergy between these fields to generate new insights. It identifies underlying mechanisms and cultivates causal thinking by eliciting the perspectives of pupils and school personnel on the Norwegian School Fruit Programme implementation dynamics, thereby contributing to theory building. It identifies underlying mechanisms and cultivates causal thinking by eliciting the perspectives of pupils and school personnel on the Norwegian School Fruit Programme implementation dynamics, thereby contributing to theory building. Using the school fruit programme as a case, the present study situates implementation determinants within a systems narrative, for example, it demonstrated that implementation processes are feedback loops intended to lower the costs associated with limited innovation reach. Using the school fruit programme as a case, the present study situates implementation determinants within a systems narrative, for example, it demonstrated that implementation processes are feedback loops intended to lower the costs associated with limited innovation reach. Background The promotion of healthy dietary habits is considered to be a key public health strategy to prevent non-communicable diseases [1]. Schools play an important role in implementing evidence-based primary prevention interventions that promote such habits [2]. Although the effectiveness of these interventions has been established through randomized controlled trials [3], their implementation is often low to moderate [4, 5]. Therefore, understanding the implementation of evidence-based primary prevention interventions in schools has become a priority to improve their effectiveness. Implementation involves introducing evidence-based interventions into complex systems, such as schools [6]. As complex systems, schools are made up of various components, including different agents, policies and physical environments, and are interconnected and influenced by other systems, such as the state education system and families [7]. The interconnected components constitute feedback loops, which produce systems’ dynamics [8, 9]. Implementation of an evidence-based intervention within a complex system influences and changes its structure and dynamics [9]. Structure pertains to the interrelationship of system components, whereas dynamics or behaviour refers to how system components change over time [10]. An understanding of complex system structure and dynamics is necessary for improving implementation interventions [9, 11]. Research on implementing primary prevention interventions at schools has primarily focused on understanding system components, rather than system structure and dynamics. For example, the synthesis of evidence on school-based nutritional programmes, such as fruit and vegetable distribution [12, 13], has identified important facilitators: sufficient and stable funding, personnel recognizing programme advantages and communication between agents at various levels, including policy, school and family [12, 13]. However, lack of priority and alignment with schools’ objectives have hindered the success of implementation [12, 13]. In response to the need to translate evidence into practice, implementation science has emerged [14], offering multiple conceptual frameworks, such as the Consolidated Framework for Implementation Research (CFIR), to guide empirical research [15, 16]. However, CFIR, including its recent addendum [17], lacks guidance on tracking the complex processes connecting implementation determinants to each other and implementation outcomes, necessitating a different set of methods and tools to track the structure and dynamics of systems within which implementation interventions take place. Integration of systems science into implementation science has been suggested for tracking implementation dynamics [11, 18–20]. However, guidelines are lacking, resulting in limitations in previous research, especially in the context of school-based programmes and policies. Existing studies often lack a clear definition of complex systems and focus more on multi-level interventions rather than considering the setting as a complex system [21, 22]. Although system framing and terminology are commonly used, established systems science methodologies are underutilized [19, 21–24]. These methodologies offer researchers a toolkit for analysing complexity such as causal loop diagrams (CLD), which illustrate interdependencies and feedback loops for dynamic reasoning [10]. Dynamic reasoning refers to an analytical process to investigate the role of feedback loops and interconnections in driving changes within a system [25]. There are two types of feedback loops [25]: a reinforcing feedback loop amplifies or reinforces changes in a system, producing growth, or accelerating decline [25]. Conversely, balancing feedback loops constrain changes within a system and tend to bring the system to a desired state [25]. Moreover, previous research mostly focuses on university or local community initiatives in the USA and the UK [22, 24, 26, 27], necessitating further research on different programmes and locations. One recommended systems science method for studying implementation issues is system dynamics [28]. Qualitative system dynamics modelling [29], a subset of this methodology, enables us to investigate the dynamic complexity of public health and implementation issues [30]. This methodology is suitable for studying implementation interventions because it assumes that the underlying system’s structure can explain system’s dynamic behaviour [8, 11]. System’s dynamic behaviour is one of the key concepts in the field of system dynamics, referring to pattern of change in key components of complex systems over time [25]. Structure is commonly represented using tools like CLDs, whereas system’s dynamic behaviour is depicted using graphs illustrating changes in system components over time [10]. Various data sources, including previous literature, secondary qualitative and quantitative data and stakeholders’ input, are used to elicit a system’s structure and dynamics. Group model building (GMB) is a commonly used method to engage stakeholders directly in the system dynamics modelling process [31]. In a GMB workshop, a facilitator invites participants to interact and share their perceptions about the structure and dynamics through standardized procedures known as scripts [32]. Interest in applying GMB to public health research has increased in recent years [30]. In implementation research, its application is recommended as a novel approach for investigating implementation mechanisms [33]. In addition, GMB has proved valuable in guiding the design of implementation plans and strategies in its few use cases in the field [34, 35]. Evidence from other research domains suggests that GMB is more effective than traditional group facilitation techniques in fostering communication and consensus-building among stakeholders [36]. This finding is also supported by systematic reviews of its applications in public health [30, 37]. The Norwegian School Fruit Programme The Norwegian School Fruit Programme (the Programme) is a national programme designed to promote eating of fruit and vegetables among pupils [38]. Its core component is distributing fruit and vegetables to improve their availability and accessibility in primary and lower secondary schools (grades 1–10). The Programme was implemented nationally in 2002. Its current financing model combines government subsidies with contributions from parents (Parent-paid Program). Financial contributions from municipalities, schools and parent associations could replace the parental contribution [38]. The latter financing model is now referred to as the No Parental Payment Programme (NPPP). Since 2019, the capital city of Oslo has been one of the key municipalities subsidizing the remaining subscription costs and offering NPPP for all pupils in lower secondary schools (grades 8–10) [39]. There is evidence of the Programme’s effectiveness in improving pupils’ eating of fruit and vegetables, especially when it is offered as the NPPP [40–43]. However, there is limited evidence about dynamics of its implementation, which may influence its effectiveness, warranting further investigation. Aim The present study aimed to investigate how NPPP is being implemented in schools and how it interacts with the school system structure and dynamics as perceived by school personnel and pupils. Moreover, it aims to understand the dynamic interplay between CFIR determinants within the NPPP’s context. Methods Study design The present study is part of the FRESKO project, which seeks to advance the understanding of adoption, implementation and impact of school-based primary prevention interventions by integrating system dynamics with implementation science. To achieve the aims of both the FRESKO project and the present study, we employed qualitative system dynamic modelling [29], as the overarching methodology, and GMB [31] for data collection. The first step in the system dynamics modelling process is to identify the system’s behaviour [25]. Accordingly, we considered changes in fruit orders over time, calculated as a percentage of the total pupils, as the systems’ behaviour that best indicated implementation dynamics. The second step in the modelling process is to develop a dynamic hypothesis by illustrating the system’s structure, which can explain the system’s behaviour [25]. Using insights from GMB workshops, we developed a narrative in retrospect about how changes in fruit orders occurred over time due to feedback processes and interconnections. Ultimately, the CFIR was mapped on to the final CLD to study the dynamic interaction of its determinants. This research was completed in accordance with the Consolidated Criteria for Reporting Qualitative research (COREQ) checklist (see Additional file 1) [44]. Participants We purposely selected public lower secondary schools (grades 8–10) and combined primary and lower secondary schools (grades 1–10) that offer the NPPP. To facilitate in-person workshops, schools were chosen from Oslo, both east and west boroughs that represent different socioeconomic neighbourhoods, and a neighbouring county in vicinity of Oslo. Our goal was to include schools with different degrees of implementation based on their fruit order information for autumn 2022, available on the Programme organizer’s webpage [45]. Schools were invited by phone and email. Once they agreed to participate, information letters for workshop participants (school personnel and pupils) were sent to the schools. We asked schools to invite student council members from diverse backgrounds in grades 8 to 10. The invited school personnel should be directly or indirectly involved in implementing the NPPP. Before the workshops, all participants signed a consent form, with parents or guardians providing consent and pupils providing assent. Workshops We scheduled two 90-min GMB workshops at approximately 2-week intervals in each school between January and April 2023. The activities for the first workshop were pilot tested with a group of clinical nutrition students beforehand. The workshops were facilitated in Norwegian by C.L.B. or N.L., with M.Z. drawing the CLD and a project team member taking notes. After each workshop, the facilitation team held a debriefing session to reflect on the participants’ discussion and summarize the lessons learned. The workshops were recorded, summarized in Norwegian (C.L.B.) and subsequently translated into English (M.Z.). M.Z. worked closely with C.L.B. to translate and analyse the first sessions. First workshop During the first workshops, the ‘causal mapping with seed structure’ script served as the primary tool for working with participants in co-creating a CLD [46]. The purpose of the seed structure was to familiarize participants with system dynamics concepts and encourage discussion [47]. The seed structure for this workshop (see Additional file 2) was derived from a CLD, which the lead author developed and discussed with co-authors (N.L., C.L.B. and B.K.), based on analyses of focus groups and interviews conducted for the FRESKO project [48]. Participants were instructed to expand the seed structure in small groups and then shared their ideas with the whole group. Pupils expanded on the factors that result in their eating fruit in the NPPP, whereas personnel expanded on factors shaping their perception of the need for it. The facilitator (C.L.B. or N.L.) and modeller (M.Z.) collaborated to help the group translate their ideas into words and arrows, using Stella Software. After the first workshops, the developed CLDs were compared with the summaries and notes (M.Z. and C.L.B.) to ensure that they accurately captured the discussions. Second workshop The purpose of the second workshops was to evaluate the CLDs from the first workshops and ask clarifying questions. The CLDs were presented in three or four overarching, colour-coded themes with a narrative for each, and participants were given A3 colour prints of the CLD and the narratives, following the ‘model review’ script [49]. They were then asked to verify that the CLD incorporated the key points discussed during the previous workshop and encouraged them to propose any necessary modifications. After working in small groups, participants shared their suggested changes with the whole group. The next activity involved asking for follow-up, clarifying questions on topics discussed in the first session and investigating topics that, according to previous literature and systems perspective, were important but not yet discussed. These questions covered themes such as participants’ understanding of the NPPP’s implementation aim, pupils’ interactions within the NPPP, mechanisms influencing personnel motivation, schools’ infrastructure and interaction with suppliers. A summary of the first and second workshop activities is provided in Additional file 3 and a detailed protocol can be made available from the lead author on request. After the second workshops, the lead author modified the CLDs based on the analysis of the notes and summaries. These modifications to the CLDs were made completely after the workshops, rather than during them. The analysis procedure to modify the CLDs adhered closely to a well-established method for developing CLDs [50]. Merging CLDs The lead author compared the final CLD of each school and identified both similar and unique system structures. All similar structures were merged and each unique structure was added to the merged CLD, no structure was discarded. Iterative reviews ensured the merged CLD remained consistent with the workshop summaries and the source CLDs throughout the process. This approach was informed by the examples from previous studies [51, 52]. All the individual CLDs after the first and second sessions can be provided by the lead author on reasonable request. Mapping CFIR to CLD To understand the dynamic interplay between CFIR’s determinants in the context of the NPPP, the lead author systematically reviewed the explanation of each feedback loop and mapped relevant determinants to the explanation in an Excel spreadsheet. The identified determinants were then visually mapped to either a factor or an entire feedback loop in the merged CLD based on their compatibility. To enhance trustworthiness, the lead author consulted with B.M. and N.L. throughout this step. Results The GMB workshops were held in four schools with the following characteristics: two lower secondary schools (grades 8–10) and two combined schools (grades 1–10). Three schools were in Oslo: two in high socioeconomic neighbourhoods, one in a low socioeconomic neighbourhood, and one school was in a rural area in a neighbouring county; sizes ranging from 250 to 700 pupils (Table 1). In terms of the dynamic behaviour of the NPPP implementation over time, each school’s fruit order is plotted between spring 2019 and spring 2023 in Fig. 1. Table 1Key characteristics of the schools and their corresponding participants who attended the group model building workshopsSchoolDistrict in OsloSchool typeSession 1Session2School size (number of pupils)Fruit location in schoolPupilsTeacherAdmin personnelPupilsTeacherAdmin personnel1East (Urban-Low socioeconomic area)8–104–24–2300–500Mainly canteen only during lunch break, andclassrooms was also possible if teachers allowed2West (Urban-High socioeconomic area)8–1092–72–> 500Mainly classrooms, andcanteen only during lunch break3West (Urban-High socioeconomic area)1–10623522> 500Entrance and recreation room4County in Oslo vicinity (Rural area)1–10512411300–500ClassroomsSchool type 8–10 denotes lower secondary schools, whereas type 1–10 refers to combined primary and lower secondary schoolsFig. 1Dynamic behaviour of the No Parental Payment Programme implementation according to the schools’ fruit orders from spring 2019 to spring 2023. Note. The fruit order between spring 2020 and autumn 2021 was also influenced by COVID-19 Table 1Key characteristics of the schools and their corresponding participants who attended the group model building workshopsSchoolDistrict in OsloSchool typeSession 1Session2School size (number of pupils)Fruit location in schoolPupilsTeacherAdmin personnelPupilsTeacherAdmin personnel1East (Urban-Low socioeconomic area)8–104–24–2300–500Mainly canteen only during lunch break, andclassrooms was also possible if teachers allowed2West (Urban-High socioeconomic area)8–1092–72–> 500Mainly classrooms, andcanteen only during lunch break3West (Urban-High socioeconomic area)1–10623522> 500Entrance and recreation room4County in Oslo vicinity (Rural area)1–10512411300–500ClassroomsSchool type 8–10 denotes lower secondary schools, whereas type 1–10 refers to combined primary and lower secondary schools Key characteristics of the schools and their corresponding participants who attended the group model building workshops Mainly canteen only during lunch break, and classrooms was also possible if teachers allowed Mainly classrooms, and canteen only during lunch break School type 8–10 denotes lower secondary schools, whereas type 1–10 refers to combined primary and lower secondary schools Dynamic behaviour of the No Parental Payment Programme implementation according to the schools’ fruit orders from spring 2019 to spring 2023. Note. The fruit order between spring 2020 and autumn 2021 was also influenced by COVID-19 Participants included 24 pupils and 12 personnel, of whom 4 and 2, respectively, did not participate in the second workshops (Table 1). There was, overall, an even distribution across grade levels and between teachers and administrative personnel. Figure 2 depicts the system structure underlying the dynamics of the NPPP implementation. Table 2 lists the feedback loops forming this structure, along with representative quotes from participants. In the next section, the dynamics of the NPPP’s implementation, resulting from the interaction between various components of the structure, are explained in four primary themes. Furthermore, parts of the structure relevant to each theme are extracted from Fig. 2 and presented throughout the text. Within these figures, the CFIR’s determinants that were matched with the CLD are shown in blue boxes. Finally, a summary of the dynamic interactions among the different determinants of the CFIR framework, based on the findings, is provided.Fig. 2The merged causal loop diagram synthesized from group model building workshops in four Norwegian schools investigating the No Parental Payment Programme (NPPP) implementation structure and dynamicsNote. Positive signs indicate that variables change in the same direction, whereas negative signs show that they change in the opposite direction, assuming that other factors are constant. Causal loop diagrams feature two feedback loops: balancing loops (B) that resist changes and reinforcing loops (R) that amplify changes. Each colour in the diagram represents one theme: dark blue (schools see the benefits of the NPPP, which facilitates more fruit orders), red (pupils not eating enough fruit is a barrier to implementation), green (schools made changes to the implementation to deal with waste and cost and the degree to which the design of the NPPP could be changed) and light blue (schools’ goals clashed with and took precedence over changes to implementation that could have addressed pupils’ needs)Table 2Main themes, corresponding feedback loops and examples of quotations from the group model building workshopsThemeLoopQuotationSchools see the benefits of the No Parental Payment Programme (NPPP), which facilitates more fruit ordersR1School personnel, school 1: Fruit contributes to a bit of leveling … evening out social inequalities. … Yes, right? We need to offer something for everyone regardless.B1School personnel, school 3: We see that not all pupils have a healthy packed lunch. Some do not eat breakfast. They do not bring a packed lunch. And we can see throughout the day that their energy decreases. Their mood and blood sugar levels and such. So, there is simply a need for some children to have access to something else. We do have a bit of a breakfast arrangement as such, but, still, it is not enough.Pupils eating insufficient fruit is a barrier to implementationR2Pupil, school 2: … the fact that you feel that you’re eating healthier, the fact that you get more energy …. And things like that. That you gain energy, you have enough stamina. You get in a better mood. … It is much better for the pupils to relate to each other when we are in a good mood.R3Pupil, school 4: If someone else is eating and says the apples were really good today, you might feel like having one. But what if they say, ‘Stay away from the clementines, they weren’t so good’.R4Pupil, school 2: And then, in our class specifically, a large portion, especially of the boys, just …. They go to some store, and people are wandering around and hanging out with their friends, and so …. They seem to use very little …. They spend as little of their time as possible when they have free time or breaks at school.R5Pupil, school 1: If you’re a friend [with someone] who goes to [stores] and buys food every recess, you don’t always have time to grab some fruit. At least not now, [with fruit] being in the canteen. Yes, because often you have to catch buses, so you run out. But if [fruit] were at [corridor’s name], you could have taken it in your hand while running out.R6Pupil, school 1: If you look at availability…. When you entered the classroom and delivered, it was super-accessible. Not only did you get a break in the teaching that was taking place there. You also got fruit and it works on a part-time basis. You can sit where you are. Then you were very accessible. Then you go to the canteen, which might be the least accessible.R7School personnel, school 3: … [pupils], in cooperation with the class teacher, took responsibility for the basket. Everyone also had to discard the fruits that they didn't want to eat themselves.Pupil, school 3: … No one had the heart to throw it away either. Because it’s not us who usually threw it away. I don’t know who threw it away, but it was just left in the classroom, so it started to smell.Schools made changes to the implementation to deal with waste and costB2School personnel, school 4: We order so that it meets the needs of the pupils. And then we try to …. … since we have a school subscription. So if there is a lot left over, we try to adjust it down. If we see that there is too little fruit, we adjust it up a bit. And those who cut up the fruit can see this.B3School personnel, school 3: It [offering fruit in the class] lasted for a little while. Then everyone came down at half-past eleven with the basket and [took fruit] to the classroom. They had to eat the fruit in the classroom, [but] they got littering, they had struggled with it at school before.B4 & B5School personnel, school 2: … as the school has previously had experience with a school fruit programme. It was a less extensive arrangement than this one. … but you still had issues like food waste, food fights, and lots of garbage and all that. And in the memo [guideline], there was actually a suggestion on how we could manage it so that pupils wouldn’t have to use their recess time for this, to ensure that the box would be returned and all this.B6School personnel, school 4: We have leftover fruit. But there is also a positive side, which is that the teachers who stay after the pupils have left can eat some of that fruit. And the after-school programme can get help.B7School personnel, school 3: The boys [pupils] keep nagging that maybe we should get a bit more [fruit] variety. So we make sure to speak up.Facilitator: Who do you speak up to?.School personnel, school 3: It’s the one who delivers [fruit].B8School personnel, school 4: We don’t get whole fruit at the school here. If it were a whole apple, it would be fine, but when the apple is cut into pieces, pupils eat more fruit than if given a whole apple. They often eat many apple pieces but won’t take a whole apple.Schools’ goals clashed with and took precedence over changes to implementation that could have addressed pupils’ needsB9School personnel, school 2: It was important to have a separate employee handling it [NPPP], instead of it being a teacher. This is because, for the teachers’ part, it is something that we view as taking our time away from other things that we consider more important during the school day. Things like pupil contact, communication with parents, planning lessons. So, it comes down to the roles that we have at the school and how we can best use our time.R8Pupil, school 2: We were thinking about having better organization, because in my class, for example, no one ever informs us when we need to pick up the school fruit. It would have been nice if the teachers told [pupils], because no one in our class does anything unless they are told to do it.R9School personnel, school 1: Previously, I used to go into every classroom and hand out fruit. Not all teachers thought that it was a good idea. When you finally get the class to settle down, and then I come in to hand out fruit, and then everyone needs to wash the fruit, and then it’s time to eat … it’s actually quite disruptive to the lesson. And I understand that. So we had to stop doing that. That’s why it was moved to the canteen.B10School personnel, school 3: Because I think that if you just leave them [fruit] available in the hallways, it will result in a lot of fruit mess everywhere.Pupil, school 3: Then there has to be an adult stationed at each of those fruit stations if you go with the first option.B11School personnel, school 1: For here, the administration has a positive attitude to begin with. But I can imagine that if you have an administration that …. There are many administrations around. If you spend more time on organization, it will be at the expense of your time usage and logistics.The degree to which the design of the NPPP could be changedB7School personnel, school 4: But what kind of fruit they get means a lot for how much fruit is eaten. Some types of fruit are very popular, and others almost always get left behind. We try to manage this from [NPPP’s website] …. So people place orders, but maybe you don’t always have the option to choose exactly what you want.Letter B denotes, balancing feedback loops, that resist changes, while letter R denotes reinforcing feedback loops, that amplify changes. Feedback loops in this table are grouped by themes, and are organized by order of their appearance in the paper The merged causal loop diagram synthesized from group model building workshops in four Norwegian schools investigating the No Parental Payment Programme (NPPP) implementation structure and dynamics Note. Positive signs indicate that variables change in the same direction, whereas negative signs show that they change in the opposite direction, assuming that other factors are constant. Causal loop diagrams feature two feedback loops: balancing loops (B) that resist changes and reinforcing loops (R) that amplify changes. Each colour in the diagram represents one theme: dark blue (schools see the benefits of the NPPP, which facilitates more fruit orders), red (pupils not eating enough fruit is a barrier to implementation), green (schools made changes to the implementation to deal with waste and cost and the degree to which the design of the NPPP could be changed) and light blue (schools’ goals clashed with and took precedence over changes to implementation that could have addressed pupils’ needs) Table 2Main themes, corresponding feedback loops and examples of quotations from the group model building workshopsThemeLoopQuotationSchools see the benefits of the No Parental Payment Programme (NPPP), which facilitates more fruit ordersR1School personnel, school 1: Fruit contributes to a bit of leveling … evening out social inequalities. … Yes, right? We need to offer something for everyone regardless.B1School personnel, school 3: We see that not all pupils have a healthy packed lunch. Some do not eat breakfast. They do not bring a packed lunch. And we can see throughout the day that their energy decreases. Their mood and blood sugar levels and such. So, there is simply a need for some children to have access to something else. We do have a bit of a breakfast arrangement as such, but, still, it is not enough.Pupils eating insufficient fruit is a barrier to implementationR2Pupil, school 2: … the fact that you feel that you’re eating healthier, the fact that you get more energy …. And things like that. That you gain energy, you have enough stamina. You get in a better mood. … It is much better for the pupils to relate to each other when we are in a good mood.R3Pupil, school 4: If someone else is eating and says the apples were really good today, you might feel like having one. But what if they say, ‘Stay away from the clementines, they weren’t so good’.R4Pupil, school 2: And then, in our class specifically, a large portion, especially of the boys, just …. They go to some store, and people are wandering around and hanging out with their friends, and so …. They seem to use very little …. They spend as little of their time as possible when they have free time or breaks at school.R5Pupil, school 1: If you’re a friend [with someone] who goes to [stores] and buys food every recess, you don’t always have time to grab some fruit. At least not now, [with fruit] being in the canteen. Yes, because often you have to catch buses, so you run out. But if [fruit] were at [corridor’s name], you could have taken it in your hand while running out.R6Pupil, school 1: If you look at availability…. When you entered the classroom and delivered, it was super-accessible. Not only did you get a break in the teaching that was taking place there. You also got fruit and it works on a part-time basis. You can sit where you are. Then you were very accessible. Then you go to the canteen, which might be the least accessible.R7School personnel, school 3: … [pupils], in cooperation with the class teacher, took responsibility for the basket. Everyone also had to discard the fruits that they didn't want to eat themselves.Pupil, school 3: … No one had the heart to throw it away either. Because it’s not us who usually threw it away. I don’t know who threw it away, but it was just left in the classroom, so it started to smell.Schools made changes to the implementation to deal with waste and costB2School personnel, school 4: We order so that it meets the needs of the pupils. And then we try to …. … since we have a school subscription. So if there is a lot left over, we try to adjust it down. If we see that there is too little fruit, we adjust it up a bit. And those who cut up the fruit can see this.B3School personnel, school 3: It [offering fruit in the class] lasted for a little while. Then everyone came down at half-past eleven with the basket and [took fruit] to the classroom. They had to eat the fruit in the classroom, [but] they got littering, they had struggled with it at school before.B4 & B5School personnel, school 2: … as the school has previously had experience with a school fruit programme. It was a less extensive arrangement than this one. … but you still had issues like food waste, food fights, and lots of garbage and all that. And in the memo [guideline], there was actually a suggestion on how we could manage it so that pupils wouldn’t have to use their recess time for this, to ensure that the box would be returned and all this.B6School personnel, school 4: We have leftover fruit. But there is also a positive side, which is that the teachers who stay after the pupils have left can eat some of that fruit. And the after-school programme can get help.B7School personnel, school 3: The boys [pupils] keep nagging that maybe we should get a bit more [fruit] variety. So we make sure to speak up.Facilitator: Who do you speak up to?.School personnel, school 3: It’s the one who delivers [fruit].B8School personnel, school 4: We don’t get whole fruit at the school here. If it were a whole apple, it would be fine, but when the apple is cut into pieces, pupils eat more fruit than if given a whole apple. They often eat many apple pieces but won’t take a whole apple.Schools’ goals clashed with and took precedence over changes to implementation that could have addressed pupils’ needsB9School personnel, school 2: It was important to have a separate employee handling it [NPPP], instead of it being a teacher. This is because, for the teachers’ part, it is something that we view as taking our time away from other things that we consider more important during the school day. Things like pupil contact, communication with parents, planning lessons. So, it comes down to the roles that we have at the school and how we can best use our time.R8Pupil, school 2: We were thinking about having better organization, because in my class, for example, no one ever informs us when we need to pick up the school fruit. It would have been nice if the teachers told [pupils], because no one in our class does anything unless they are told to do it.R9School personnel, school 1: Previously, I used to go into every classroom and hand out fruit. Not all teachers thought that it was a good idea. When you finally get the class to settle down, and then I come in to hand out fruit, and then everyone needs to wash the fruit, and then it’s time to eat … it’s actually quite disruptive to the lesson. And I understand that. So we had to stop doing that. That’s why it was moved to the canteen.B10School personnel, school 3: Because I think that if you just leave them [fruit] available in the hallways, it will result in a lot of fruit mess everywhere.Pupil, school 3: Then there has to be an adult stationed at each of those fruit stations if you go with the first option.B11School personnel, school 1: For here, the administration has a positive attitude to begin with. But I can imagine that if you have an administration that …. There are many administrations around. If you spend more time on organization, it will be at the expense of your time usage and logistics.The degree to which the design of the NPPP could be changedB7School personnel, school 4: But what kind of fruit they get means a lot for how much fruit is eaten. Some types of fruit are very popular, and others almost always get left behind. We try to manage this from [NPPP’s website] …. So people place orders, but maybe you don’t always have the option to choose exactly what you want.Letter B denotes, balancing feedback loops, that resist changes, while letter R denotes reinforcing feedback loops, that amplify changes. Feedback loops in this table are grouped by themes, and are organized by order of their appearance in the paper Main themes, corresponding feedback loops and examples of quotations from the group model building workshops School personnel, school 3: … [pupils], in cooperation with the class teacher, took responsibility for the basket. Everyone also had to discard the fruits that they didn't want to eat themselves. Pupil, school 3: … No one had the heart to throw it away either. Because it’s not us who usually threw it away. I don’t know who threw it away, but it was just left in the classroom, so it started to smell. School personnel, school 3: The boys [pupils] keep nagging that maybe we should get a bit more [fruit] variety. So we make sure to speak up. Facilitator: Who do you speak up to?. School personnel, school 3: It’s the one who delivers [fruit]. School personnel, school 3: Because I think that if you just leave them [fruit] available in the hallways, it will result in a lot of fruit mess everywhere. Pupil, school 3: Then there has to be an adult stationed at each of those fruit stations if you go with the first option. Letter B denotes, balancing feedback loops, that resist changes, while letter R denotes reinforcing feedback loops, that amplify changes. Feedback loops in this table are grouped by themes, and are organized by order of their appearance in the paper Schools see the benefits of the NPPP, which facilitates more fruit orders The reinforcing feedback loop R1 and balancing feedback loop B1 (Fig. 3) highlight NPPP’s advantages for the pupils (innovation: relative advantages) and principals’ support (individual: high-level leaders – motivation), reinforced personnel’s belief in the need for NPPP (inner setting: tension for change; culture: recipient centeredness). This perceived need influenced schools’ decisions to order fruit (implementation outcome: implementation). Personnel perceived the NPPP offering various advantages, such as providing healthy snacks to pupils who skipped breakfast, forgot lunch or lacked fruit at home. They believed it could improve pupils’ mood stability and allow personnel to give pupils positive attention. Moreover, fruit provision without parental payment (innovation: relative advantages) was perceived to address social inequality in fruit access, bring pupils together and prevent some from feeling excluded for not visiting stores during recess to buy food. The reinforcing feedback loop R1 thus led to an increase in the fruit order over time, as long as the fruit was not wasted. Fruit waste, along with the handling of leftovers and litter, could be considered as a cost of NPPP, as it leads to the loss of time and resources, especially nutritious food that could have been utilized otherwise. The balancing feedback loop B2, which regulates the fruit waste (innovation: cost), stopped R1 from increasing the fruit order any further.Fig. 3Causal loop diagram showing the mechanisms that led schools to recognize the benefits of the No Parental Payment Programme (NPPP)Note. R1: Perceived NPPP relative advantages → Staff perceived need for NPPP → Ordering fruit from NPPP → Perceived NPPP relative advantages. B1: Discrepancy in pupils eating fruit → Staff perceived need for NPPP (reducing social inequality) → Ordering Fruit from NPPP → Perceived NPPP relative advantages → Pupils eating fruit provided by NPPP → Discrepancy in pupils eating fruit. B2: Leftover fruit → Amount of fruit waste → Inadequacy of fruit waste → Ordering fruit from NPPP → Leftover fruit. Positive signs indicate that variables change in the same direction, whereas negative signs show that they change in the opposite direction, assuming that other factors are constant. Causal loop diagrams feature two feedback loops: balancing loops (B) that resist changes and reinforcing loops (R) that amplify changes. Constructs of the Consolidated Framework for Implementation Research (CFIR) are mapped to the diagram in the blue boxes in the following order [Domain: Determinant] Causal loop diagram showing the mechanisms that led schools to recognize the benefits of the No Parental Payment Programme (NPPP) Note. R1: Perceived NPPP relative advantages → Staff perceived need for NPPP → Ordering fruit from NPPP → Perceived NPPP relative advantages. B1: Discrepancy in pupils eating fruit → Staff perceived need for NPPP (reducing social inequality) → Ordering Fruit from NPPP → Perceived NPPP relative advantages → Pupils eating fruit provided by NPPP → Discrepancy in pupils eating fruit. B2: Leftover fruit → Amount of fruit waste → Inadequacy of fruit waste → Ordering fruit from NPPP → Leftover fruit. Positive signs indicate that variables change in the same direction, whereas negative signs show that they change in the opposite direction, assuming that other factors are constant. Causal loop diagrams feature two feedback loops: balancing loops (B) that resist changes and reinforcing loops (R) that amplify changes. Constructs of the Consolidated Framework for Implementation Research (CFIR) are mapped to the diagram in the blue boxes in the following order [Domain: Determinant] Pupils not eating enough fruit is a barrier to implementation The reinforcing feedback loop R2 (Fig. 4) shows how pupils perceiving NPPP’s advantages (innovation: relative advantage) led to eating the provided fruit (innovation outcome: reach). Eating fruit, in turn, reinforced their initial belief about the advantages of NPPP, which led to eating more fruit. Pupils noted advantages such as maintaining a healthy, diverse diet, gaining energy to concentrate on their tasks (especially if they skipped breakfast or forgot launch) and having a break from studies. Moreover, pupils discussed how peers picking up fruit could encourage them to do the same and influence their perception of its taste and quality (reinforcing feedback loop R3 in Fig. 4).Fig. 4Causal loop diagram showing the mechanisms driving pupils’ eating of the fruit within the No Parental Payment Programme (NPPP)Note. R2: Pupils eating fruit provided by NPPP → Perceived NPPP relative advantages → Pupils eating fruit provided by NPPP. R3: Socializing in NPPP → Perceived NPPP relative advantages → Pupils’ motivation to pick up fruit for themselves → Pupils picking up fruit → Socializing in NPPP. R4: Perceived store advantages → Pupils visiting stores → Socializing in stores → Perceived store advantages. R5: Pupils visiting stores → Socializing in stores → Perceived store advantages → Pupils’ motivation to pick up fruit for themselves → Pupils visiting stores. R6: Offering fruit in accessible places → Perceived NPPP relative advantages → Pupils eating fruit provided by NPPP → Leftover fruit → Amount of fruit waste → Inadequacy of fruit waste → Offering fruit in accessible places. R7: Following up the guidelines → Taking the fruit basket back → Perceived NPPP relative advantages → Pupils eating fruit provided by NPPP → Leftover fruit → Littering → Teachers perceived value of NPPP → Teachers’ motivation → Following up the guidelines. Positive signs indicate that variables change in the same direction, whereas negative signs show that they change in the opposite direction, assuming that other factors are constant. Causal loop diagrams feature two feedback loops: balancing loops (B) that resist changes and reinforcing loops (R) that amplify changes. Constructs of the Consolidated Framework for Implementation Research (CFIR) are mapped to the diagram in the blue boxes in the following order [Domain: Determinant] Causal loop diagram showing the mechanisms driving pupils’ eating of the fruit within the No Parental Payment Programme (NPPP) Note. R2: Pupils eating fruit provided by NPPP → Perceived NPPP relative advantages → Pupils eating fruit provided by NPPP. R3: Socializing in NPPP → Perceived NPPP relative advantages → Pupils’ motivation to pick up fruit for themselves → Pupils picking up fruit → Socializing in NPPP. R4: Perceived store advantages → Pupils visiting stores → Socializing in stores → Perceived store advantages. R5: Pupils visiting stores → Socializing in stores → Perceived store advantages → Pupils’ motivation to pick up fruit for themselves → Pupils visiting stores. R6: Offering fruit in accessible places → Perceived NPPP relative advantages → Pupils eating fruit provided by NPPP → Leftover fruit → Amount of fruit waste → Inadequacy of fruit waste → Offering fruit in accessible places. R7: Following up the guidelines → Taking the fruit basket back → Perceived NPPP relative advantages → Pupils eating fruit provided by NPPP → Leftover fruit → Littering → Teachers perceived value of NPPP → Teachers’ motivation → Following up the guidelines. Positive signs indicate that variables change in the same direction, whereas negative signs show that they change in the opposite direction, assuming that other factors are constant. Causal loop diagrams feature two feedback loops: balancing loops (B) that resist changes and reinforcing loops (R) that amplify changes. Constructs of the Consolidated Framework for Implementation Research (CFIR) are mapped to the diagram in the blue boxes in the following order [Domain: Determinant] However, a lack of information about the NPPP, perceived stores advantages, and discrepancies between their preferences and the NPPP’s design or implementation weakened feedback loops R2 and R3, which influenced pupils’ perceived NPPP advantages. Pupils described lacking information about the NPPP, including where it was available in their schools, the fruit of the day and the Programme’s origin (inner setting: communication). They believed that having this information would highlight the NPPP’s advantages better. Many pupils in schools 1, 2 and 3 mentioned that they were more inclined to visit food stores (outer setting: local condition) than to pick up the fruit at school. Visiting food stores increased peer interactions, further encouraging store visits (reinforcing loop R4 in Fig. 4). Consequently, their motivation to pick up fruit waned, especially if the fruit was not nearby, leading to a decline in picking fruit up (feedback loop R5 in Fig. 4). Pupils across all schools noted that the characteristics of fruit provided by the NPPP, including variety, quality, presentation and accessibility (innovation: design), differed from their preferences (individuals: innovation recipients – need). For example, with regard to the type and quality of fruit, they mentioned that they expected strawberries but instead got mushy bananas. Some of these differences were inherent in the NPPP’s design, whereas others were the result of implementation processes (Feedback loop R6 and R7 in Fig. 4), which we discuss in the next sections. In summary, various factors lowered pupils’ perception of NPPP’s advantages and their eating of fruit. Limited amounts of fruit eaten led to more leftovers, occasionally causing littering. This, in turn, increased fruit waste and added to the operating costs of the NPPP for schools. Schools made changes to the implementation to deal with waste and cost Schools, as complex adaptive systems, can self-organize, learn and create new structures [7]. These new structures in NPPP’s implementation, illustrated in Fig. 5, comprised a series of balancing feedback loops B2–B8. Balancing feedback loops typically aims to maintain the system’s condition near its desired condition, and loops B2–B8 aimed to minimize fruit waste to the lowest acceptable level (innovation: cost), namely, after a period of implementation, the innovation cost led to a reflective and evaluative process (process: reflecting and evaluating). During this, schools operationalized either one or a combination of these feedback loops as implementation processes to reduce the cost of the NPPP. For example, as shown in feedback loop B2 (Fig. 5), all schools ordered less fruit because fewer pupils were eating it than anticipated (process: tailoring strategies). In line with this process, Fig. 1 shows a recurring pattern across all schools over the years, indicating that a large fruit order is often followed by a smaller one.Fig. 5Causal loop diagram showing the mechanisms driving the changes that schools made within the No Parental Payment Programme (NPPP) implementation to address waste and costNote. B2: Inadequacy of fruit waste → Ordering fruit from NPPP → Leftover fruit → Amount of fruit waste → Inadequacy of fruit waste. B3: Littering → Amount of fruit waste → Inadequacy of fruit waste → Offering fruit in accessible places → Littering. B4: Inadequacy of fruit waste → Providing guidelines → Following up the guidelines → Picking up fruit for classroom → Offering fruit in accessible places → Perceived NPPP relative advantages → Pupils eating fruit provided by NPPP → Leftover fruit → Amount of fruit waste → Inadequacy of fruit waste B5: Inadequacy of fruit waste → Providing guidelines → Following up the guidelines → Taking the fruit basket back → Perceived NPPP relative advantages → Pupils eating fruit provided by NPPP → Leftover fruit → Amount of fruit waste → Inadequacy of fruit waste. B6: Inadequacy of fruit waste → Redistributing → Leftover fruit → Amount of fruit waste → Inadequacy of fruit waste. B7: Inadequacy of fruit waste → Asking different fruit types → Perceived NPPP relative advantages → Pupils eating fruit provided by NPPP → Leftover fruit → Amount of fruit waste → Inadequacy of fruit waste. B8: Inadequacy of fruit waste → Cutting → Perceived NPPP relative advantages → Pupils eating fruit provided by NPPP → Leftover fruit → Amount of fruit waste → Inadequacy of fruit waste. Positive signs indicate that variables change in the same direction, whereas negative signs show that they change in the opposite direction, assuming that other factors are constant. Causal loop diagrams feature two feedback loops: balancing loops (B) that resist changes and reinforcing loops (R) that amplify changes. Constructs of the Consolidated Framework for Implementation Research (CFIR) are mapped to the diagram in the blue boxes in the following order [Domain: Determinant] Causal loop diagram showing the mechanisms driving the changes that schools made within the No Parental Payment Programme (NPPP) implementation to address waste and cost Note. B2: Inadequacy of fruit waste → Ordering fruit from NPPP → Leftover fruit → Amount of fruit waste → Inadequacy of fruit waste. B3: Littering → Amount of fruit waste → Inadequacy of fruit waste → Offering fruit in accessible places → Littering. B4: Inadequacy of fruit waste → Providing guidelines → Following up the guidelines → Picking up fruit for classroom → Offering fruit in accessible places → Perceived NPPP relative advantages → Pupils eating fruit provided by NPPP → Leftover fruit → Amount of fruit waste → Inadequacy of fruit waste B5: Inadequacy of fruit waste → Providing guidelines → Following up the guidelines → Taking the fruit basket back → Perceived NPPP relative advantages → Pupils eating fruit provided by NPPP → Leftover fruit → Amount of fruit waste → Inadequacy of fruit waste. B6: Inadequacy of fruit waste → Redistributing → Leftover fruit → Amount of fruit waste → Inadequacy of fruit waste. B7: Inadequacy of fruit waste → Asking different fruit types → Perceived NPPP relative advantages → Pupils eating fruit provided by NPPP → Leftover fruit → Amount of fruit waste → Inadequacy of fruit waste. B8: Inadequacy of fruit waste → Cutting → Perceived NPPP relative advantages → Pupils eating fruit provided by NPPP → Leftover fruit → Amount of fruit waste → Inadequacy of fruit waste. Positive signs indicate that variables change in the same direction, whereas negative signs show that they change in the opposite direction, assuming that other factors are constant. Causal loop diagrams feature two feedback loops: balancing loops (B) that resist changes and reinforcing loops (R) that amplify changes. Constructs of the Consolidated Framework for Implementation Research (CFIR) are mapped to the diagram in the blue boxes in the following order [Domain: Determinant] Furthermore, as indicated by the balancing feedback loop B3 (Fig. 5), schools moved the offer of fruit to supervised locations, such as classrooms and later canteens, to prevent littering (process: tailoring strategies). In all schools, teachers were asked to engage in implementing the NPPP and, with the help of pupils (process: engaging), offer the fruit in classrooms (balancing feedback loop B4) and return the fruit baskets to their designated spot (balancing feedback loop B5) (process: tailoring strategies). Moreover, schools 3 and 4 redistributed leftover fruit to teachers, after-school programmes or primary schools (balancing feedback loop B6) (process: tailoring strategies). Schools 3 and 4 asked the delivery person or organization responsible for the NPPP to provide different types of fruit, based on their pupils’ preferences (balancing feedback loop B7 in Fig. 5) (process: adapting). School 4 was the only school that cut the fruit to encourage consumption (balancing feedback loop B8 in Fig. 5) (process: adapting). It is important to note that a system’s self-organization can sometimes produce counterintuitive outcomes [53]. For instance, feedback loop B3 caused a conflict in fruit placement preferences: pupils preferred easily accessible locations, whereas personnel preferred areas that allowed for supervision to mitigate littering. This reduced pupils’ perceived advantages of the NPPP, decreased eating of fruit and unintentionally increased leftovers and waste (feedback loop R6 in Fig. 4). Moreover, feedback loops B4 and B5, which aimed to offer fruit in classrooms for better accessibility, did not work as intended. In two schools, fruit baskets were often not returned, exacerbating littering or causing fruit to become warm or rot, thus diminishing the NPPP’s advantages (feedback loop R7 in Fig. 4). Schools’ goals clashed with and took precedence over changes to implementation that could have addressed pupils’ needs At schools 1, 2 and 3, the administration developed a simple guideline to inform teachers about the NPPP and how to implement it in their classrooms, as described above. However, teachers and school personnel in schools 1 and 2 stated that classroom activities were prioritized, with education as the primary purpose of schools rather than fruit distribution (inner setting: relative priority). Following the guidelines necessitated pupils picking up and/or washing the fruit; consequently, processes B4 and B5 were perceived as interrupting teaching and contributing to the inadequacy of teaching time (inner setting: compatibility). This reduced teachers’ motivation and adherence to the guidelines, as indicated in feedback loop B9 in Fig. 6A (implementation outcome: adoption; individual: innovation deliverers, motivation).Fig. 6Causal loop diagrams showing the mechanisms driving the changes that made within the No Parental Payment Programme (NPPP) implementation to address waste and cost, and their interaction with schools’ goals and priorities are presented in panels A and BNote. A B9: Perceived inadequacy of teaching time → Teachers’ motivation → Following up the guideline → Teaching time → Perceived inadequacy of teaching time. R8: Teachers’ motivation → Pupils motivation to pick up fruit for class → Pupils asking teachers about fruit → Teachers’ motivation. R9: Teachers’ motivation → following up the guidelines → Taking the fruit basket back → Littering → Teachers perceived value of NPPP → Teachers’ motivation. B B10: Littering → Allocating human resources for NPPP → Overseeing fruit distribution → Littering. B11: Perceived principal support → Allocating human resources for NPPP → Human resources for school activities → Perceived quality of school activities → Perceived principal support. B3: Littering → Amount of fruit waste → Inadequacy of fruit waste → Offering fruit in accessible places → Littering. R6: Offering fruit in accessible places → Perceived NPPP relative advantages → Pupils eating fruit provided by NPPP → Leftover fruit → Amount of fruit waste → Inadequacy of fruit waste → Offering fruit in accessible places. Positive signs indicate that variables change in the same direction, whereas negative signs show that they change in the opposite direction, assuming that other factors are constant. Causal loop diagrams feature two feedback loops: balancing loops (B) that resist changes and reinforcing loops (R) that amplify changes. Constructs of the Consolidated Framework for Implementation Research (CFIR) are mapped to the diagram in the blue boxes in the following order [Domain: Determinant] Causal loop diagrams showing the mechanisms driving the changes that made within the No Parental Payment Programme (NPPP) implementation to address waste and cost, and their interaction with schools’ goals and priorities are presented in panels A and B Note. A B9: Perceived inadequacy of teaching time → Teachers’ motivation → Following up the guideline → Teaching time → Perceived inadequacy of teaching time. R8: Teachers’ motivation → Pupils motivation to pick up fruit for class → Pupils asking teachers about fruit → Teachers’ motivation. R9: Teachers’ motivation → following up the guidelines → Taking the fruit basket back → Littering → Teachers perceived value of NPPP → Teachers’ motivation. B B10: Littering → Allocating human resources for NPPP → Overseeing fruit distribution → Littering. B11: Perceived principal support → Allocating human resources for NPPP → Human resources for school activities → Perceived quality of school activities → Perceived principal support. B3: Littering → Amount of fruit waste → Inadequacy of fruit waste → Offering fruit in accessible places → Littering. R6: Offering fruit in accessible places → Perceived NPPP relative advantages → Pupils eating fruit provided by NPPP → Leftover fruit → Amount of fruit waste → Inadequacy of fruit waste → Offering fruit in accessible places. Positive signs indicate that variables change in the same direction, whereas negative signs show that they change in the opposite direction, assuming that other factors are constant. Causal loop diagrams feature two feedback loops: balancing loops (B) that resist changes and reinforcing loops (R) that amplify changes. Constructs of the Consolidated Framework for Implementation Research (CFIR) are mapped to the diagram in the blue boxes in the following order [Domain: Determinant] Furthermore, at school 2, teachers lost motivation and disregarded guidelines because they felt that the administrators did not follow up on the NPPP’s implementation (inner setting: incentive system). Reminders and follow-up with teachers were important, particularly with the high teacher turnover after the pandemic or whenever there was a substitute teacher, as mentioned in schools 2 and 3. The teachers’ lack of motivation also influenced pupils’ engagement in NPPP’s implementation (individual: innovation recipients – motivation), resulting in reinforcing cycles of disengagement among both teachers and pupils (reinforcing feedback loops R8 and R9 in Fig. 6A). Furthermore, the low priority of the NPPP caused schools to overlook processes like overseeing fruit distribution (balancing feedback loop B10 in Fig. 6B), which pupils from schools 1, 2 and 3 suggested could reduce littering and improve fruit accessibility (process: tailoring strategies). School 3 reported limited personnel, whereas schools 1 and 2 cited the NPPP’s low priority in their busy schedules as reasons for not implementing this process (inner setting: available resources, relative priority). School 1 acknowledged reallocating resources and that approval of the principal was needed (individual: high-level leaders – motivation). However, as shown in balancing feedback loop B11 (Fig. 6B), they believed that, if this reallocation influenced the quality of ‘main’ school activities (like planning the new school year), there would be less support from principals, given NPPP’s low priority compared with the school’s goals (inner setting: compatibility). School personnel acknowledged that improved teams and planning could effectively manage leftover fruit and minimize waste (process: teaming and planning), an aspect previously overlooked (inner setting: relational connections) (Fig. 6B). For example, personnel in school 3 indicated that a process flow chart would help handle excess fruit more efficiently, reducing pressure and frustration. Faced with disengaged teachers and pupils, schools 1, 2 and 3 reduced fruit orders and placed fruit in supervised areas like canteens rather than distributing it in classrooms. Most schools experimented with different implementation processes, which could explain the variability in fruit orders (Fig. 1). The degree to which NPPP’s design could be changed As previously described, the NPPP’s design already allowed for some adaptability (innovation: adaptability). Nevertheless, school personnel believed additional changes could improve its implementation, although they acknowledged the high costs of these changes made them a ‘dream scenario’. For example, they mentioned the NPPP allowed ordering various fruit selection, but the current variety did not fully match pupils’ preferences. Figure 5, feedback loop B7, shows how the NPPP’s adaptability limits schools’ requests for different types of fruit. Suggested changes to the NPPP’s design by personnel included conducting surveys to identify preferred fruit, providing pre-cut fruit, and dedicating extra personnel to help with tasks such as overseeing fruit distribution and cutting. Summary of the results through the lens of CFIR determinants The synergetic relative advantages of innovation, along with motivation by high-level leaders and an emphasis on the needs of recipients, initiated reinforcing feedback loop R1 (Fig. 7), driving initial implementation. Over time, the continued recognition of innovation benefits by implementers sustained implementation, provided that it did not increase operating costs for the schools as settings.Fig. 7Causal loop diagram showing interrelationships between the determinants of the Consolidated Framework for Implementation Research that shaped the implementation dynamics of the No Parental Payment Programme (NPPP)Note. R1: Perceived need for the innovation → Implementation outcome: outcome → Innovation perceived relative advantages → Perceived need for the innovation. R2: Innovation perceived relative advantages → Innovation outcome: reach → Innovation perceived relative advantages. R3: Innovation: cost → Implementation process: reflecting & evaluating: implementation → Implementation process: engaging → Demand on inner setting: available resources → Individual: innovation implementers motivation → Implementation outcome: adoption → Implementation process: tailoring strategies: adapting → Innovation: cost. R4: Individual: innovation implementers motivation → Individual: innovation recipients: motivation regarding innovation → Individual: innovation implementers motivation. B1: Innovation: perceived inadequacies → Innovation outcome: reach → Innovation: perceived inadequacies. B2: Innovation: cost → Implementation process: reflecting & evaluating: implementation → Implementation process: tailoring strategies: adapting → Innovation: cost. B3: Demand on inner setting: available resources → Individual: high-level leaders: motivation → Demand on inner setting: available resources. B4: Innovation: cost → Implementation outcome: implementation → Innovation: cost. Positive signs indicate that variables change in the same direction, whereas negative signs show that they change in the opposite direction, assuming that other factors are constant. Causal loop diagrams feature two feedback loops: balancing loops (B) that resist changes and reinforcing loops (R) that amplify changes. Colours represent different domains of the Consolidated Framework for Implementation Research (CFIR): black (individual), red (inner setting), green (innovation), purple (innovation outcome), pink (outer setting), and blue (implementation process) Causal loop diagram showing interrelationships between the determinants of the Consolidated Framework for Implementation Research that shaped the implementation dynamics of the No Parental Payment Programme (NPPP) Note. R1: Perceived need for the innovation → Implementation outcome: outcome → Innovation perceived relative advantages → Perceived need for the innovation. R2: Innovation perceived relative advantages → Innovation outcome: reach → Innovation perceived relative advantages. R3: Innovation: cost → Implementation process: reflecting & evaluating: implementation → Implementation process: engaging → Demand on inner setting: available resources → Individual: innovation implementers motivation → Implementation outcome: adoption → Implementation process: tailoring strategies: adapting → Innovation: cost. R4: Individual: innovation implementers motivation → Individual: innovation recipients: motivation regarding innovation → Individual: innovation implementers motivation. B1: Innovation: perceived inadequacies → Innovation outcome: reach → Innovation: perceived inadequacies. B2: Innovation: cost → Implementation process: reflecting & evaluating: implementation → Implementation process: tailoring strategies: adapting → Innovation: cost. B3: Demand on inner setting: available resources → Individual: high-level leaders: motivation → Demand on inner setting: available resources. B4: Innovation: cost → Implementation outcome: implementation → Innovation: cost. Positive signs indicate that variables change in the same direction, whereas negative signs show that they change in the opposite direction, assuming that other factors are constant. Causal loop diagrams feature two feedback loops: balancing loops (B) that resist changes and reinforcing loops (R) that amplify changes. Colours represent different domains of the Consolidated Framework for Implementation Research (CFIR): black (individual), red (inner setting), green (innovation), purple (innovation outcome), pink (outer setting), and blue (implementation process) The perceived advantages of the innovation by pupils initially reinforced its reach through feedback loop R2. However, R2 was weakened by factors like poor communication, the relative advantages of local conditions (food stores outside schools) and discrepancies between recipients’ needs and the innovation’s design or implementation (balancing feedback loop B1, Fig. 7). The lack of innovation’s reach increased its operating cost because it necessitated those who implemented having to deal with waste and littering issues. Observing increased operating costs of implementation, implementers organized new implementation processes to keep the cost manageable (balancing feedback loop B2). However, some of these implementation processes were constrained by the setting’s goals and priorities or the innovation itself. As a result, they did not work as intended (reinforcing feedback loops R3 and R4) and even reduced the innovation’s perceived advantages, leading to less reach and higher costs. The persistent cost of innovation, combined with its low priority within the setting, led to a reliance on ineffective implementation processes (balancing feedback loop B4), failing to address NPPP’s insufficient reach. Once the symptoms of this issue, such as littering and waste, had been addressed, the need for reflection and evaluation decreased. Implementers did not recognize this problem because their main goal was minimizing cost and waste, overlooking recipients’ needs, which were crucial for improving innovation’s reach. Discussion The present study aimed to investigate how NPPP is being implemented in schools and how it interacts with the school system structure and dynamics, as perceived by school personnel and pupils. Furthermore, it aimed to understand the dynamic interplay between the determinants of the CFIR in the specific context of the NPPP. The initial implementation of the NPPP faced challenges owing to pupils’ low levels of eating fruit, leading to increased operating costs for schools in managing leftover fruit and litter. Attempts to address this through various implementation processes conflicted with school goals or the NPPP’s design. Counterintuitively, some processes diminished the NPPP’s perceived relative advantages from the pupils’ perspective, causing further decline in pupils eating of fruit and a reappearance of litter and waste. Schools then prioritized implementation of waste-eliminating processes, thereby neglecting those that would improve the NPPP’s reach to pupils. The system’s structure determines its observed behaviour [8]. For example, issues in NPPP’s implementation, such as pupils not eating enough fruit and the inefficiencies of the implementation processes, resulted from the interactions of multiple feedback loops, the smaller components of the system’s structure. This discussion investigates the link between the dynamics presented in the results and the underlying system structure, focusing on its implications for NPPP’s implementation. The competition between food stores and innovation The interaction between reinforcing loops R3 and R5 illustrates the competition that exists between the NPPP and the food stores outside the schools (e.g. local grocery stores). Food purchased outside school is more appealing due to its better perceived taste and satiety, diminishing pupils’ interest in the NPPP. This is a common systems failure, known as the ‘success to the successful’ archetype [54], which emerges from two activities: going to food stores versus picking up fruit competes for pupils’ limited time during recess. The more attractive option of socializing at food stores gains more participation, causing erosion of the appeal and social engagement in picking up fruit. Recent systematic reviews of implementation evaluations of school nutritional policies similarly highlighted the presence of local food stores as a barrier [55, 56]. Moreover, our study, in line with that of Shaw et al. [57], demonstrates that the reinforcing mechanism of R4, socializing with peers, plays an important role in making food stores more attractive. To balance these competing mechanisms, we can consider either reducing the attractiveness of the food stores outside school or enhancing the perceived attractiveness of the NPPP. Highlighting its no-cost nature for pupils and creating a social environment around it are potential pathways for achieving this. Classic dilemma: innovation goal versus inner setting goal Schools have a goal to increase children’s knowledge and skills, whereas the NPPP’s goal is to change their dietary habits. Goals have powerful leverage in influencing the behaviour of a system because they set the direction for the activities within it [53]. Given the limited time and resources available, schools often find it challenging to satisfy both goals simultaneously. As a result, they tend to prioritize their educational goals over those of the NPPP. This situation is a common systemic failure known as ‘competing goals’, an archetype that arises when two balancing feedback loops with different goals interact [54]. These interactions were highlighted in the results between feedback loop B9 and loops B5 and B4, as well as between loops B10 and B11. This finding aligns with previous research that identified schools’ priorities and goals as the primary barriers to the implementation and sustainment of school-based nutritional interventions [12, 55, 56]. To address this systemic issue, it is suggested that we either redefine a broader goal that aligns all efforts or find a collaborative approach to achieve both goals [54]. Innovation’s goal eroded over time Schools implemented the NPPP and placed their fruit order based on reflection and evaluation of their own performance. For example, if they had a lot of waste, they would reduce their fruit order or, if there was no leftover fruit, they believed that more fruit was needed. From a system perspective, adjusting goals based on internal performance can lead to a systemic failure known as ‘drift to low performance’ [58]. Prescriptive actions include identifying and reconciling the competing goals, as discussed earlier. Furthermore, it entails maintaining the NPPP’s original vision [54], which is to improve pupils’ eating of fruit and anchoring it to an external reference [59], as described below. An external reference is needed to prevent drifting towards low performance standards. We can compare this to the ‘performance-measurement pressure’ factor in CFIR [16]. This factor underscores the significance of establishing a goal for innovation implementation from the perspective of an outer setting. The influence of an outer setting on implementation of the NPPP, as our results showed, was absent. The absence of monitoring and enforcement at the macro-level has also been recognized as a barrier to implementing the Norwegian school meal guidelines [60]. Holding to the original vision for the NPPP as a prescriptive action of the ‘drift to low performance’ archetype [58] supports Hawe et al.’s theory that focusing on the main goal or ‘function’ of an intervention, rather than its ‘form’, is required when undertaking implementation processes [9]. This is also consistent with the dynamic sustainability framework, which stresses use of progress measures aligned with recipients’ desired outcomes and context to prevent ‘programme drift’ [61]. For example, when considering the NPPP’s ‘recipients’ desired outcomes’ or ‘function’ is to increase eating of fruit by pupils, the focus should be on monitoring this rather than details of fruit ordering, because the latter pertains more to implementation ‘form’. Ideally, a successful implementation should recognize the value for both pupils and school personnel, aiming to carry out implementation processes in a way that optimizes value across different levels and among various stakeholders [62], rather than solely focusing on reducing cost. This perspective shift allows schools to prioritize context-compatible implementation processes that enhance the NPPP’s function. Our findings confirmed that schools that prioritized monitoring of the fruit order over the NPPP’s primary function experienced more waste and littering issues, which resulted in ‘program drift’. Reflection on using CFIR In the present study, we aimed to understand the dynamic interplay between CFIR determinants within the NPPP’s context. However, we acknowledge that the interrelationships found in this study are specific to the NPPP context and may differ in other contexts. An added benefit of using the CFIR framework was that it facilitated the alignment of terminology and concepts from implementation science with those of systems science, thus enhancing interdisciplinary communication. To analyse the interaction between outcomes and determinants, we required an outcome framework, for which the CFIR addendum was suitable [17]. CFIR’s ability to articulate determinants for both implementers and recipients was also crucial for the present study. This framework holds significant potential for future use in similar studies. Reflection on using GMB We observed that GMB workshops fostered rich discussions among school personnel and pupils. Particularly, they described a broader perspective on NPPP implementation processes, by reflecting on and building upon each other’s thoughts. The use of CLD as a visual tool to anchor discussions was instrumental in this regard, contributing to the advantages of GMB over conventional methods like focus groups. While conducting and analysing GMB sessions were time-intensive and required specialized skills, we are confident that the resulting fruitful discussions justified the effort. Overall, we agree with Black (2013) that understanding complex challenges, like implementation issues, is inherently a social process, and GMB offers a valuable approach to facilitate this [63]. Strengths and weaknesses A key strength of the present study was its focused investigation of the dynamic interplay between implementation determinants and the complex aspects of NPPP implementation within school system, such as feedback loops, self-organization and counterintuitive processes. This nuanced focus provided a deep understanding of the mechanisms at play. Furthermore, the study’s inclusion of both school personnel and pupils in the GMB workshops enriched the discussions, providing a complete view of the context. However, the substantial time commitment required for workshops might lead to participation of more invested schools. To mitigate this, we aimed to include schools with varying levels of implementation based on their fruit orders. Moreover, the relative novelty of the system dynamics methodology which is still developing standardized procedures, present a challenge when applying it to the equally emerging field of implementation science. The authors addressed this by staying updated with recent developments in both fields through active engagement with publications and dissemination events. Overall, we have taken steps to mitigate these challenges, but the extent of our success is uncertain, and the results should be viewed within the context of these potential limitations. Furthermore, this study reflects solely on the implementation dynamics in retrospect. Ideally, we should have collected data over multiple phases and applied a phase-based framework to enhance our analysis. However, as outlined in the introduction, the NPPP’s implementation, in its current format, began in 2019 in Oslo, three years prior to the inception of this project. This timeline made it impossible to investigate implementation using a phase-based approach. Lastly, this study used the CFIR framework to understand the dynamic interplay between implementation determinants; however, given the wide range of available implementation frameworks [64], it is possible that alternative frameworks such as Multi-level Quality Implementation Framework [65] may have offered a better contextual fit or yielded different insights. Future research could explore other frameworks to validate or expand our findings. Conclusions This research advances existing literature on implementation of evidence-based primary prevention in schools by employing qualitative system dynamics to analyse the dynamic complexity of NPPP implementation. Key systemic characteristics identified include feedback loops, self-organization, and counterintuitive mechanisms. Systemic archetypes, such as success to the successful, competing goals and drift to low performance, were also identified. Future studies may build on our CLD to develop quantitative system dynamics models for further analysis. The main conclusion of the present study is that lack of communication with pupils, the food stores outside schools, and the NPPP’s design reduced its reach and inflated its costs. Moreover, schools’ goals and priorities hampered the success of processes implemented to mitigate these issues. These findings have two main implications for implementation of the NPPP and similar programmes. First, programme developers should consider the attractiveness of the food stores surrounding schools during implementation of the NPPP. To enhance the reach of the programme, they should explore ways to increase its attractiveness and consider changes to reduce the appeal of off-campus food stores. Second, our findings revealed instances where the NPPP’s design or implementation diverged from pupils’ preferences, thereby impeding its success in improving eating of fruit. Therefore, it is important to establish clear implementation progress measures to track the recipients’ outcomes, specifically the changes in pupils’ eating of fruit. These measures should be aligned with the schools’ main goals and can be monitored by both inner setting, at leadership level, and outer settings, such as municipalities, which oversee primary and lower secondary schools in Norway. When deviations from these progress measures occur, feedback should be sought from both pupils and school personnel. Workshops, such as GMB workshops, offer a promising approach to bring these groups together and facilitate dialogue. Future research should also investigate the stakeholders’ perspectives in outer settings like municipalities to gain a better understanding of their role in monitoring the implementation of the NPPP and how it can be further enhanced. Supplementary Information Additional file 1. COREQ checklist.Additional file 2. Seed structure.Additional file 3. Workshop activities. Additional file 1. COREQ checklist. Additional file 2. Seed structure. Additional file 3. Workshop activities.